Saturday, January 13, 2018


By Liam Kofi Bright and Aaron Novick.
(Joint work with equal contributions.)

Kongzi, on being asked the first thing to do in administering government, gave a surprising answer:

Zilu asked, “If the Duke of Wei were to employ you to serve in the government of his state, what would be your first priority?” 
The Master answered, “It would, of course, be the rectification of names.” (Analects 13.3, tr. Slingerland)

Rectifying names (正名 zhengming), Kongzi says, is the basis of social flourishing. If names are out of order, speech will not match reality, plans will be impossible to put into action, culture will decline, and punishment will be ineffective. The central task of the gentleman, then, is put names in order. As Kongzi puts it, “The gentleman simply guards against arbitrariness in speech” (13.3).

We suggest that Carnap (though we expect that similar things could be said for others of the logical empiricists) had an interestingly similar conception of the philosopher’s intellectual task. Our aim here is to draw out these connections, focusing especially on Xunzi’s constructivist Confucianism and Carnap’s work on logical analysis. For Xunzi, we are highly indebted to Kurtis Hagen’s interpretation in The Philosophy of Xunzi: A Reconstruction (2007, Open Court).

Xunzi attempted to provide a theoretical basis for Kongzi’s emphasis on the importance of rectifying names. This basis he substantially co-opted from Zhuangzi, who had argued for the conventionality of both language and social customs. Zhuangzi saw these insights as posing a serious challenge to Confucianism; Xunzi sought to show how they could support it. He granted that “names have no predetermined appropriateness” (“Correct Naming”, tr. Hutton), by which he meant not just that the application of a particular sound to a particular kind of object is arbitrary, but also that the the boundaries between kinds, (henceforth “kind-boundaries”) are themselves not to be evaluated against a standard set by nature or the metaphysical structure of the world. For instance, we may use the word “cow” to pick out certain animals. The animals picked out by our usage of this term will in some ways be similar to other critters that we do not label with “cow”. Likewise the critters picked out by “cows” will have certain dissimilarities with each other. The privileging of certain (dis)similarities as more important than others is a pragmatic matter: it requires the judgment that those (dis)similarities are relevant to the tasks for which it is important to distinguish cows from non-cows. (Throughout this paragraph especially we are following Hagen; see his book for a full defense of this interpretation.)

The question, then, is how names and their associated kind-boundaries are established. Here, Xunzi offers an empiricist, pragmatist theory. Kind-boundaries are established on the basis of perceivable similarities and differences. Each of the senses has a proper realm of differentiation (e.g. “form, color, and pattern” for the eyes), while the heart/mind (心 xin) “has the power to judge its own awareness,” i.e. to recognize what the senses detect and to form judgments on that basis.

Because kind-boundaries are not predetermined, the criterion of good judgment cannot be correspondence to reality. Rather, the criterion is pragmatic. Xunzi thinks that language is open to social design, and that it should be judged based on its effectiveness in facilitating social order and human flourishing. Consider the following passage, in which Xunzi criticizes claims advanced by earlier philosophers:
Claims such as “To be insulted is not disgraceful,” “The sage does not love himself,” and “To kill a robber is not to kill a man” are cases of confusion about the use of names leading to disordering names. If one tests them against the reason why there are names, and observes what happens when they are carried out thoroughly, then one will be able to reject them.
Appropriate naming is naming that can be put into practice with beneficial social consequences. As Hagen argues, this implies that there may be multiple acceptable systems of naming, each of which facilitates social ordering. The point is just to pick and adopt one, on the basis of its ability to be taken up and applied to good end by both the government and the people.

A central task of the intelligentsia, for Xunzi, is to rectify names. Often, the form that this takes is bringing into alignment the descriptive and normative aspects of thick concepts. This can be done by either insisting on stricter application of a term’s present meaning, or clarifying the sense of a term to resolve ambiguities. One of Kongzi’s central concerns, for instance, was to modify the sense of the word junzi (君子), which roughly means “gentleman.” Kongzi recognized that the term picked out both those who were noble by birth and also those whose behaviour was noble in the sense of being morally admirable, and, crucially, was often misleadingly used to imply that those who were noble by birth were therefore morally admirable. He thus claimed that the term was more properly reserved for those whose behaviour qualified them as moral exemplars (regardless of birth), and modified his use of junzi accordingly.

For Xunzi ideally a sage king or ideal ruler would judiciously design a language system and propagate it throughout the empire. But short of direct instruction from a sage king, the ruist (i.e. Confucian) intelligentsia, which a good ruler would employ as ministers, would also be engaged in zhengming.

Further, as part of the task of rectifying names, advisors to the ruler were expected to remonstrate: in effect, to call out the ruler for failing to live up to his obligations. Rectifying names involved clarifying social roles (e.g. “parent” signifies both a biological fact and an associated set of obligations), and this clarification was not merely theoretical. For instance, a ruler who governed oppressively must be corrected. At the extreme, a ruler who failed in his obligations could lose all claim to the title (and to life), as in this case, from the Mengzi (tr. van Norden), concerning the tyrant Zhou:
The King said, “Is it acceptable for subjects to kill their rulers?” 
Mengzi said, “One who violates benevolence should be called a ‘thief.’ One who violates righteousness is called a ‘mutilator.’ A mutilator and thief is called a mere ‘fellow.’ I have heard of the execution of a mere fellow ‘Zhou,’ but I have not heard of the killing of one’s ruler.”
Moving on to Carnap, we begin with the well known fact that he was a verificationist. The exact form that verificationism took changed over his life - for a reasonably mature and influential statement of his view see here. But the broad idea of the more mature position is that any claim that is a candidate for truth or falsity must either be analytic or stand in some kind of (dis)confirmation relation to empirical evidence. Simplifying somewhat, this is to say that if the “claim” or its negation is not made true by our logico-mathematical framework, and there is no empirical evidence an ideal (team of) scientist(s) could gather that should leave you any the wiser as to whether my “claim” is true or false, then I have failed to make a cognitively meaningful claim. The hedging of “cognitively” before “meaningful” is to accommodate Carnap's recognition that there are other things one may wish to do with language besides make descriptive claims, and he was fine with that; but he thought that it was improper to try and evaluate as true or false such linguistic acts as commands, questions, poetical expressions of our yearnings, or exasperated sighs.

Carnap was also a conventionalist about kind-boundaries. For detailed discussion of the origins and development of this element of his thought, see here. For our purposes suffice it to say that according to Carnap there are no natural kinds, joints in nature, or Platonic forms, which our linguistic practices must or will inevitably line up with or pick out and attach to. Rather, we may decide upon linguistic practices on the basis of their pragmatic usefulness in achieving certain practical or theoretical goals that various language forms stand to assist us in attaining.

Most famously this view is elaborated upon by Carnap in Empiricism, Semantics, and Ontology - but the same idea can be found (perhaps in more restrictive contexts of application) in other works. For instance in The Continuum of Inductive Methods Carnap discusses how one should pick an inductive method as such:
The adoption of an inductive method is neither an expression of belief nor an act of faith, though either or both may come in as motivating factors. An inductive method is rather an instrument for the task of constructing a picture of the world on the basis of observational data and especially of forming expectations of future events as a guidance for practical conduct. X may change this instrument just as he changes a saw or an automobile, and for similar reasons.

Finally, we draw attention to the fact that Carnap also thought that an important task for a philosopher was bringing people into line with the austere verificationist standards he advocated. This was the basis for the (in)famous attack on Heidegger in Carnap’s The Elimination of Metaphysics. But such language policing was also advocated as a proper intellectual activity elsewhere. For instance in the Vienna Circle manifesto, which Carnap helped edit, it is said that in the glorious philosophy of the future “[n]o special 'philosophic assertions' are established, assertions are merely clarified”. Implicitly in the former but explicitly in the latter, it is clear that the authors of the manifesto believe that rendering language empirically tractable will assist the progressive segments of humanity hold to account the representatives of the failed ancien regime. Strikingly, it is even claimed that “in many countries the masses now reject [metaphysical] doctrines much more consciously than ever before, and along with their socialist attitudes tend to lean towards a down-to-earth empiricist view”. Carnapian linguistic policing is thus meant to help ensure our theoretical or scientific projects are fruitfully and efficiently carried out, and our shared social life is free of superstition and the obscurantist propaganda of tyrants.

To review, we have now seen that according to both Xunzi and Carnap the following are true. Linguistic categories do not and need not reflect some objective true or accurate mode of dividing up the world. Rather, we have a kind of epistemic free choice in deciding upon our preferred kind-boundaries. That is not to say, however, that there are no standards of better or worse for linguistic conventions: it is just that the appropriate way of evaluating proposed linguistic conventions is how well they help us advance our practical goals. Both Carnap and Xunzi think that if this is done properly we will end up with a system wherein our utterances are answerable to empirically discernible features of the world.  They then think an important task for intellectuals is to ensure that people are in fact engaging in this kind of empirically responsive and responsible speech.

As an aside, we note that there is even a kind of stylistic similarity between them in their more condemnatory modes. Here is Carnap on where Heidegger and those like him go wrong:
Thus, the words of the foolish person are hurried and rough. They are agitated and have no proper categories. They are profuse and jumbled. He is one who makes his words seductive, muddies his terms, and has no deep concern for his intentions and thoughts. Thus he exhaustively sets out his words yet has no central standard. He works laboriously and has no accomplishments. He is greedy but has no fame.
Compare this with Xunzi’s talking about those who lack enough culture to express themselves clearly:
The harmonious feeling or attitude, which the metaphysician tries to express in a monistic system, is more clearly expressed in the music of Mozart. And when a metaphysician gives verbal expression to his dualistic-heroic attitude towards life in a dualistic system, is it not perhaps because he lacks the ability of a Beethoven to express this attitude in an adequate medium? Metaphysicians are musicians without musical ability. Instead they have a strong inclination to work within the medium of the theoretical, to connect concepts and thoughts. Now, instead of activating, on the one hand, this inclination in the domain of science, and satisfying, on the other hand, the need for expression in art, the metaphysician confuses the two and produces a structure which achieves nothing for knowledge and something inadequate for the expression of attitude.

This degree of similarity between these thinkers so divided by time, geography, and culture, is, we think, enough to merit a blog post! However,  before concluding we note some pertinent contrasts. The first thing we acknowledge, just to satisfy the increasingly agitated scholars in the back, concerns the details of their metasemantic theories. Xunzi claimed that each word functions as a name for some feature of the world we can discern with our sensory apparatus, and that a sentence consists in stringing together names in order to ever more precisely narrow down the class of things one is concerned with. Desirable tractability of a sentence is thus achieved when each name is itself properly tractable. Carnap, on the other hand, had a more sophisticated syntactic theory, and eventually allowed that it is (logically interlinked networks of) propositions or sentences which must be answerable to empirical evidence, rather than individual terms. This is indeed a difference in the letter of their theories, but nonetheless we think the spirit is the same. These differences seem to us largely due merely to the fact that (writing thousands of years earlier) Xunzi had available a much less sophisticated theory of language and logic than Carnap.

The second contrast is the particular kinds of terminology to which Carnap and Xunzi applied their respective zhengming. Xunzi tends to be in the business of clarifying thick ethical concepts and policing usage of them to ensure that those so described live up to the attached normative requirements. Carnap, on the other hand, is largely concerned to clarify concepts for use in the mathematical or empirical sciences. This is not to deny that he would police language that was of socio-political significance - one of us has a published discussion of Carnap’s linguistic reformism regarding human racial taxonomy. But the direct analysis of thick ethical terms is absent from the vast majority of his corpus (though it is gestured to in the Aufbau).  Here it seems somewhat arbitrary features of what caught Carnap’s interest limited his philosophical purview, and Xunzi’s more expansive project of zhengming strikes us as liable to be more philosophically fruitful for anyone who wanted to revive this project.

Finally,  there is the question of why one ought be a linguistic empiricist of any sort - and especially how this aspect of their thought related to their conventionalism. Xunzi, as we read him, is clearer that the linguistic empiricism of his position in Correcting Names is itself a convention. It is a semantic stance one adopts or guiding principle to be used when engaged in zhengming, because shared and empirically tractable terminology is an enabling condition for various of the social goods Xunzi hopes to secure through clarification of terms. Carnap’s position on exactly why we ought be verificationists seems on the other hand to have been unclear, and at times in his intellectual development to have left him open to charges of vicious circularity or self-refutation. Perhaps in the end Carnap had a position close to that which we see in Xunzi. But, in any case, Xunzi has at least outlined an attractive rational for any project of empiricist zhengming, which it seems Carnap or those sympathetic to him may themselves wish to explore and adopt.

Both authors of this blog post are sympathetic to something like empiricist zhengming as a fruitful project for contemporary philosophers. We thus hope that, this connection being noticed, the modern heirs to the logical positivists and modern ruists may seek greater community and dialogue. However, we think there is something to be appreciated here even if one does not wish to take up the task of rectifying names through the logical analysis of language. Philosophy, at its best, embodies a kind of cosmopolitan ideal. The superficial distinctions between people are erased, and what remains are opportunities for peaceful collaborative effort in a transnational and transtemporal republic of letters. That thinkers as different as Xunzi and Carnap gesture towards a common project is, we think, an example of such cross cultural discourse that might inspire even those who do not share their peculiar concerns.


  1. I might be wrong, but it looks like your last two Carnap and Xunzi quotes are the wrong way around.

  2. Ha, yeah, that was a joke! I thought it was funnier not to acknowledge the joke, and I am not sure if this comment will ruin things, but enough people have messaged me about it that I feel like having this attached to the post!

  3. I really enjoyed this. I just finished Karl Sigmund’s *Exact Thinking in Demented Times.*. Interesting account of the Vienna Circle. I like your mind bending creative comparison very much.

    1. Thanks! That book is on my To Read list :)

  4. I second Owen's comment! This post is fun to read and draws some illuminating connections. To fill out the picture more, it'd be worth considering what Xunzi thinks language is mainly used to do — which is significantly different from the use of language the empirical sciences, for example — along with the political side of Xunzi's views on language. (I don't know much about Carnap's politics.)

    By the way, in the original _Analects_ passage, names being incorrect hasn't much to do with speech not matching reality. The issue is more likely that teachings and instructions will not be followed (言不順).

    1. Interesting point about the Analects, and thanks for the kind words. Cheers!

      To reply to one aspect of this off the cuff, since we maybe should have stressed it more in the OP - the political ends they wish to put this too are pretty far apart! Carnap was an egalitarian socialist, and even has a passage (in a letter to Popper now published in After The Open Society) where he specifically objects to monarchy as a form of social organisation. Xunzi on the other hand (for any reading less familiar with him) places great importance on the establishment of a social hierarchy with some kind of monarch figure at the top. Carnap seems to have seen some connection between his zhengming and egalitarian social reform, whereas Xunzi argues that his preferred form of zhengming will help a good ruler govern more effectively from the top.

      So there's something pretty interesting about how these authors apparently saw their (as we have argued) morally similar proposals as conducive to near opposite social forms!

    2. Quick note: In mentioning that speech will not accord with reality, I was following Slingerland's translation:

      "If names are not rectified, speech will not accord with reality; when speech does not accord with reality, things will not be successfully accomplished..."

      Do you think this is a misleading translation?

  5. I'd say the translation is misleading, although Slingerland may simply be following an earlier translation, such as Legge's.

    The words are 名不正則言不順,言不順則事不成... "Names not being correct, speech/statements are not shùn (obeyed/complied/followed/flowing); speech/statements not being shùn, affairs are not completed/accomplished." I doubt there is any context in early texts where shùn refers to names being in accord with things. But it's regularly used to refer to obeying, complying, or following along with leaders or with circumstances.

    In view of the political context, the text is probably referring to people obeying instructions or teachings, or perhaps to being able to act on what others say. That would explain why an inability to shùn what's said leads to failure to accomplish things and why the passage later says the gentleman's speech can surely be carried out — because he uses names properly, listeners can comply with what he says and put it into practice.