«Language and Ontology in Early Chinese Thought∗ Chris Fraser Department of Philosophy Chinese University of Hong Kong July 2005 (Revised January ...»
Forthcoming in Philosophy East & West 57:4 (2007)
Language and Ontology in Early Chinese Thought∗
Department of Philosophy
Chinese University of Hong Kong
(Revised January 2007)
Chris Fraser (方克濤) (Assistant Professor)
Department of Philosophy
Rm. 430, Fung King Hey Bldg.
Chinese University of Hong Kong
Shatin, N.T., Hong Kong
E-mail: email@example.com Copyright © 2005, 2007 Brief Summary for Table of Contents The paper critiques Hansen’s “mass noun hypothesis,” arguing that though most Classical Chinese nouns do function as mass nouns, this fact does not support the claim that pre-Qín thinkers treat the extensions of common nouns as mereological wholes, nor does it explain why they adopt nominalist semantic theories. The paper shows that early texts explain the use of common nouns by appeal to similarity relations, not mereological relations. However, it further argues that some early texts do characterize the relation between individuals and collections as a mereological relation.
2 Abstract According to Chad Hansen’s “mass noun hypothesis,” the semantics of Classical Chinese nouns is similar to that of English mass nouns, and this helps to explain why ancient Chinese theories of language are nominalist. Mass nouns, Hansen suggests, are most naturally interpreted as denoting mereological wholes, a construal that intuitively tends to motivate a nominalist theory of language. The paper argues that most instances of Classical Chinese nouns indeed function as mass nouns.
However, this fact does not support the claim that ancient Chinese thinkers regarded the extensions of common nouns as mereological wholes, nor does it explain why Chinese semantic theories take nominalism for granted. Hence Hansen’s interpretive argument from mass nouns is unsound: His observations about the Chinese language do not support his interpretive conclusions about ancient Chinese semantic and ontological views. The paper shows that in fact pre-Qín philosophers of language explain the use of common nouns by appeal to similarity relations, not mereological relations. The paper then reviews the textual evidence for Hansen’s thesis that early Chinese thinkers employ a mereological ontology, concluding that some early texts do characterize the relation between individual things and the collections to which they belong as a part-whole relation, rather than a member-set relation.
In Language and Logic in Ancient China (1983), Chad Hansen proposes that the semantics of Classical Chinese nouns is similar to that of English mass nouns, a view he calls the “mass noun hypothesis.” He contends that this hypothesis helps to explain why ancient Chinese theories of language are nominalist. Mass nouns, he suggests, are most naturally interpreted as denoting mereological wholes, a construal that intuitively tends to motivate a nominalist theory of language. Hansen’s views have incited much controversy among scholars of early Chinese thought. They raise important questions concerning interpretive methodology, because as he expounds them, they represent an attempt to justify interpretive hypotheses by appeal not to the content of the texts under study, but to grammatical features of the language used by their writers.
In this paper, I review critical responses to Hansen and present a new critique of my own. Following Robins (2000), I contend that, given crucial qualifications, most instances of Classical Chinese nouns indeed function as mass nouns. Thus one premise in Hansen’s argument from mass nouns is justified. However, contrary to his claims, this point provides no reason to accept his interpretive hypotheses. It does not support the claim that Chinese thinkers regarded the extensions of general terms as mereological wholes, nor does it explain why Chinese semantic theories take nominalism for granted. Hence Hansen’s interpretive argument from the pervasiveness of mass nouns in Classical Chinese is unsound: Even if correct, his observations about the Chinese language do not support his interpretive conclusions about early Chinese semantic and ontological theories. Moreover, as I will show, pre
the use of general terms. Thus as an instance of a language-to-thought argument—one from features of a thinker’s language to interpretive conclusions about the content of his thought—Hansen’s argument fails. This failure raises doubts about the justificatory power of such arguments in general, which I explore in my conclusion.
After presenting this critique of Hansen’s views, I consider textual evidence for his hypothesis that early Chinese thinkers employ a mereological ontology. I conclude that the hypothesis is highly plausible. The grounds for it are limited and mostly confined to Mohist and Daoist texts, and the exact extent and nature of Chinese mereological views are unclear. But at least some pre-Qín philosophers do seem to have conceived of the relation between individual things and the collections to which they belong as a part-whole relation, rather than a member-set relation.
Section 2 summarizes the two core interpretive hypotheses Hansen presents in chapter 2 of Language and Logic and distinguishes them from two distinct, less plausible claims with which they are run together. The next pair of sections lay the groundwork for my subsequent arguments by explaining the distinction between word class and word function (section 3) and clarifying key features of mass nouns and their semantics (section 4). Section 5 presents and evaluates Hansen’s argument from mass nouns, concluding that it is unsound. Sections 6 and 7 review critical responses to Hansen, attempting to clear up several misunderstandings and clarify where his critics were on the mark and where not. Section 7 also briefly rebuts two competing accounts of the nature of Classical Chinese nouns. Section 8 summarizes the Mohists’ and Xúnzǐ’s accounts of the relation between things and the kinds to which they belong, showing that the considerations Hansen cites in fact play no role in their theories. Section 9 reviews textual evidence for attributing mereological views to
In chapter 2 of Language and Logic in Ancient China, Hansen proposes two interpretive hypotheses about the early Chinese view of language and the world (30The mereological worldview. Early Chinese thinkers implicitly accepted a mereological1 ontology, on which collections of things are regarded as wholes of which the things that constitute the collection are parts. For instance, instead of thinking of all the horses in the world as elements of the set or class of horses, ancient Chinese philosophers thought of them as spatially scattered parts of the concrete whole that is the sum of all horses.
2. Behavioral nominalism. Early Chinese philosophy of language is nominalistic, in that it is not committed to recognizing any entities other than words, or “names” (míng 名), and the things that form their extensions. It does not appeal to universals, essences, concepts, meanings, Lockean ideas, or Platonic forms to explain the semantics of general terms or the relation between a particular thing and its kind. Early Chinese views of the mind are “behavioral” in that they explain thought and understanding by appeal to the ability to discriminate things and act in appropriate ways. Understanding a word (such as “horse”) is not a matter of having a certain
object in one’s mind, but of having the practical ability to distinguish the things
Both of these hypotheses are highly plausible. The proposal that pre-Qín philosophers applied a mereological ontology is an important, credible interpretive hypothesis for which, as I will argue, there is solid, though limited textual evidence.
interpretation of ancient Chinese philosophy of language, as I have argued elsewhere (1999, 2002, 2005). I will review some of the grounds for this contention in section 8.
Along the way, I will explain why the statement of it given above is incomplete, requiring an account of the features by which things count as “similar” and thus as part of the extension of the same general term.2 In presenting these hypotheses, Hansen (1983) combines them with the following claims, from which they are in fact conceptually distinct.
1′. The stuff ontology. On the mereological worldview, the world is regarded as an aggregate of overlapping and interpenetrating “mass stuffs” (32).
2′. The singular term claim. Ancient Chinese behavioral nominalism regards terms for kinds of things, such as “water” and “horse,” roughly as singular terms (35).3 A term such as “horse” is regarded as the name of the discontinuous totality of the stuff “horse” (35–37). The word “horse” is
correctly used of individual horses because in a sense it is their proper name:
they are parts of the whole named “horse.” Both of these claims are mistaken. The first wrongly assumes that only things that are unstructured masses, such as liquids, can stand in part-whole relations.
Hansen describes the mereological worldview in this way because he attributes it to Chinese thinkers partly on the basis of the prevalence of mass nouns in Classical Chinese. Some paradigm mass nouns, such as “water,” “metal,” and “mud,” do refer to “mass stuffs.” But in general, using a mass noun to refer to something does not entail conceiving of it as an unstructured mass or stuff. (“Luggage,” for instance, does not refer to unstructured stuff. I will clarify this point further in section 4.) Nor does treating collections of things as mereological wholes entail regarding them as masses.
are not “overlapping and interpenetrating mass stuffs.” The second point is inspired by the grammar of mass nouns, by the fact that the Classical Chinese word for “word,” míng 名, literally means “name,” and by the mereological worldview. Hansen suggests that “mass nouns…play the same role in sentences that proper nouns do” (35). This is false, of course, as he himself points out (35). Unlike proper nouns, mass nouns can be modified, can refer to more than one thing, and can be used to form predicates. Presumably, what Hansen has in mind is that, like proper nouns, English mass nouns can stand alone as the subject of a sentence, without requiring an article or pluralization. Moreover, in Classical Chinese, both proper nouns and common nouns are referred to as “names” (míng 名). Hence, Hansen contends, if we ask what mǎ 馬 “horse” is the name of, in ancient Chinese thought the natural answer was that it is the name of “horse-stuff,” a concrete mereological whole scattered in space-time (35). Accordingly, for Chinese thinkers the one-many problem does not arise. They do not need to explain the relation between a thing and the kind it belongs to, by which one and the same word denotes the many instances of a kind. For on this view, words stand in a one-to-one relation with wholes, of which particular bits, such as individual horses, are parts (36). There are no questions to raise about the relation between particular things and the kinds to which they belong, because there are no kinds. There are only vast, spatiotemporally discontinuous individuals, such as “horse-stuff,” all of which bears the name “horse.” Learning to use a word is not a matter of learning how to identify the kind of thing denoted by that word, but of learning how to reidentify (parts of) the bearer of a proper name. There is thus no motivation for Platonic realism or Lockean conceptualism, both of which were developed as responses to the one-many problem.
implausible—so much so that no reasonably competent thinker is likely to have held it. To do so, a thinker would need to overlook the fundamental functional differences between proper nouns and general terms. He would have no concept of a kind, as distinct from an individual, and would be unable to explain how we are able to say that the black horse before us is a different horse from the white one we saw yesterday.
In any case, we have compelling reasons to reject this view as an interpretation of early Chinese thought. As section 8 will explain, the later Mohists explicitly distinguish general terms for lèi 類 (kinds), such as mǎ 馬 (horse), from what they call sī míng 私名 (personal names, proper nouns), such as “Jack,” which are singular terms referring to a single individual. The proposal that early thinkers regarded kind terms such as mǎ 馬 (horse) as akin to singular terms cannot explain why the Mohists would draw this distinction. The only reasonable explanation is that they recognize that mǎ 馬 (horse) typically functions as a general term.4 Hypotheses 1 and 2 are seminal contributions to the study of ancient Chinese thought, but they must be distinguished from hypotheses 1′ and 2′, which are false.
Unfortunately, the arguments Hansen gives in his initial presentation of hypotheses 1 and 2 are unsound (1983: 30–54). He has other evidence, collected mainly in chapters 4 and 5 of his (1983), that is more convincing. In section 5, I will review and critique his initial arguments, before moving on in sections 8 and 9 to explore some of the other, more compelling evidence for the hypotheses.
Much confusion has surrounded Hansen’s argument from mass nouns because of a widespread failure to mark the crucial distinction between word class and word function. As a result, scholars have devoted much energy to discussing whether
question is ill formed.