For what it’s worth, I’ll post a chunk of the paper I just gave at the APA meeting in New York, at a panel on the challenges of teaching Chinese philosophy. Though not a paper that I’m going to pursue much further I am, as always, still interested in any comments or questions that you may have.
Is a Little Bit of Chinese Better or Worse than None?
APA Eastern Division Meeting – Dec. 28 2009
Teaching Chinese Philosophy – Challenges and Promises
(Arranged by the APA Committee on the Status of Asian and Asian-American Philosophers and Philosophies)
In thinking about the challenges of teaching Chinese philosophy that I’ve faced, I’ve chosen to limit myself to a few that surround language and style of early Chinese discourse. In part this is because they were the first to come to mind, but also in part I’m interested in raising some questions for discussion that are both philosophical as well as pedagogical.
There are a few challenges I’ve been able to overcome in my experience—some through a bit of effort, others by blind luck. Take for example my students’ reticence to try to pronounce Chinese terms and names, or (slightly better) to pronounce them fearlessly though in incorrect or unrecognizable ways. These pose, each in their own way, significant hindrances to in-class discussion, something on which I absolutely rely. Of course, pronunciation problems multiply when some of the assigned reading pieces use one Romanization system and others another. I’ve taken to spending nearly two entire class periods going over an introductory lesson on Romanization systems with an accompanying pronunciation guide that I have developed. But, by and large, that seems to work.
Something I’ve overcome largely through blind luck, on the other hand, is the relative lack of systematic discussion in the Chinese philosophical literature that I assign. That is something I always found frustrating as a philosophy undergraduate myself. I think it lends itself to the belief among students that the Chinese didn’t really produce rigorous philosophical thought—which may be to some extent and in some sense, true. But I’ve been “fortunate” that many of my students tend actually to be less interested in systematic discussions as in pithy statements, so they find many of the Chinese texts much more interesting than, say, the Nicomachean Ethics, or even the Republic. Not that they necessarily understand them much better—just more interesting (perhaps it seems more familiar to them, at least in length, like Facebook scrawl or texted messages). So, quite by accident, I have a foot in the door when I try to get the students to expand the textual material into some semblance, when possible, of the systematic assumptions, background beliefs, and even arguments that might generate pronouncements such as, “If the mat is not straight, [the gentleman] does not sit” (Analects 10.9 – 席不正，不坐).
More difficult for me, however, is an issue—not unrelated to the two I’ve just mentioned—that seems to me to raise questions requiring philosophical resolution both in the classroom and possibly beyond. Is it helpful or harmful to insert classical Chinese terms for important concepts, untranslated, into philosophical discussion conducted otherwise completely in another language—in our case, English? Should we, for example, discuss the dao of the Confucian junzi and its relationship to li or, instead, the way of the Confucian gentleman and its relationship to ritual? (Hence the title of my presentation.) Is this an issue that transcends the particular background or preparation of the audience in question? I want to suggest so by raising some possible ways in which it may be more problematic than it seems; and I think the problems raised apply both to the pedagogical context as well as more generally for doing history of Chinese philosophy or more comparative endeavors.
Potential Problems with Untranslated Terms
To scratch this itch, let me suggest that it would be reasonable to think using the English term we’ve chosen as the translation of a classical Chinese term is sufficient for the task of philosophical instruction and discussion. For example, it should be enough to use the term “way” or “ways” to discuss dao 道 and the passages that include that term; or if one is partial to “guidance” over “way” then use “guidance,” and so forth, for whatever translation or multi-term gloss one prefers or thinks appropriate for the context. After all, why have translations for terms at all if one is not going to stick to them and use them? (To avoid confusion and potential inconsistency, I’m going to strive exclusively to mention dao and other Chinese conceptual terms throughout this presentation, rather than use them.) We might ask, “What might be the point of saying ‘dao’ rather than ‘way’?”
One reply to this might be, as already hinted, that the Chinese term in question is polysemic—that it has a large semantic field, a cluster of possible meanings that vary according to context; so, better to leave the term untranslated than bias or narrow the discussion with one’s own preferred meaning. But in doing so, I think, there are important opportunities that are passed up—or, to put it more bluntly, this is a cop out, both in the classroom and more generally. Take the pedagogical context; one might think that the student should be allowed some leeway to interpret the various instances of dao; this will help them to become better interpreters of the text. But I think the latter outcome is best served by laying out the possible interpretive possibilities, using the various translations, and discussing their strengths and weaknesses. Otherwise, the student doesn’t receive any actual guidance (what are we doing in the classroom then, quoting the Daodejing and then posing meaningfully?). Worse, the student may quite possibly end up with the very problematic impression that Chinese philosophical concepts are impenetrable in some mystical way, that they somehow elude not only translation but also finite human understanding (more on this later). Similar remarks apply outside of the pedagogical context. Mere invocation of a term’s polysemic plenitude and refusal to translate it block any actual understanding of it and encourages the illusion that once we are textual experts, we all know what it means. But in fact the plausibility of any given interpretation requires at least some discussion of the (perhaps stereotypically “analytic”) question: “What does that mean, exactly?” But to do that, some level of translation commitment seems required, whether it is term-for-term translation, or conceptual elaboration through a multi-term gloss.
I think there are some interesting potential rejoinders to my suggestions so far, so let me trot out two for consideration. As I’ve indicated at the start, the issues here are not entirely settled in my own mind, hence remain unresolved pedagogically as well as philosophically. So trotting out these rejoinders to my initial suggestion about translation are meant to stimulate and channel discussion, not to indicate the ways in which I have already anticipated and utterly defeated objections to my view prior to discussion. (That isn’t for lack of trying—actually, it probably is for some lack of trying.) Furthermore, some of these rejoinders will collapse together views that we might wish to separate.
The “Native Conceptual Grasp” or ”Going Conceptually Native” Response
One might say, “Well, how did the early Chinese thinkers conceive of these terms? Presumably, it was without translation; and we don’t actually see a lot of conceptual analysis or glossing on their part. Shouldn’t we be trying to recreate their understanding, as much as possible, in our own understanding of those terms? Isn’t the ultimate goal at least—maybe for students and more certainly for scholars—to hear, read and think dao (for example) and have an untranslated, unglossed conceptual grasp of it?”
That is different, of course, from Analects 4.8, in which Confucius states that having heard dao in the morning, he could die peacefully in the evening (朝聞道，夕死可矣). I make the contrast partially in jest, but also because there are two things we probably should distinguish as the goal of “going conceptually native.” What we could take seriously, at least in contemporary pedagogical and scholarly contexts, is the goal of understanding the term as if we were competent users of the language in which the term is embedded. This would involve understanding what its range of meaning includes and some set of uncontroversial inferential relationships into which sentences that include it would enter. All of this would have to be understood in the target language of course—classical Chinese—otherwise we would just be doing the work of translating.
Somewhat different are the possible goals that Confucius may be referring to in Analects 4.8 and elsewhere, the ones that are the product of “attaining the way,” as he puts it. This involves a different kind of understanding that goes much further than what the term means to understanding putative, substantive truths — possibly contentious — about the thing to which the term refers and in some instances, also having some kind of normative response to those truths. Surely that isn’t what anyone would be after in attaining native conceptual grasp. That should only require modest competency; otherwise, no one who understood the term dao in early China could have disagreed with Confucius about it.
But even if it is this more pedestrian, merely competent usage that we seek, there may be some obstacles to characterizing it as native competence. For one thing, it’s not clear how one could plausibly claim a sort of native linguistic competence, for oneself or for another, unless the language in question were (a) a living language, at least at one time or other, and relatedly, (b) there were actual speakers of the language around to judge the matter. Some obstacles exist to satisfying those conditions. First, it’s not so clear that very much of the literary and stylized written language of the texts we have at our disposal was ever a living language, at least construed as a spoken, natural language. Second, even if the written works could give us clues as to the native linguistic environment within which they were composed, there are still no speakers of the language. There are in fact very competent composers of classical Chinese poetry and prose in a variety of genres, but whether they should be attributed the authority of a native-speaker seems problematic. So, recreating native understanding of classical Chinese terms might be misguided. In any case, that goal seems too lofty as a piece of pedagogy.
However, short of native competence, isn’t it possible to understand a term well enough so that translating or glossing it is unnecessary? Perhaps, but is it desirable? The answer to this depends on what one thinks is valuable for the practice of philosophy. Take the English terms with which we have competence. Is it desirable to carry on philosophical discussion without further glossing some of them so that their meanings are sharpened or disambiguated? I don’t think so; the inferential relationships among statements that include some targeted terms may be the very relationships that should undergo philosophical scrutiny….
Polysemy is a slippery slope (not an informal fallacy, but a distinction with potentially catastrophic consequences). If you allow for too many definiens, a translation doesn’t appear to offer any effort to render the writings of the text self-relevant (cutting any real argument short by distancing premises via an over-broad semantic field), and if you demand too few, you may falsely criticize inconsistency of the work.
From a pedagogical perspective on any topic, I would be personally disenchanted to simply hear, “We don’t translate t because t means lots of things,” since, as you put it, it strikes me as a cop-out against investigation of those contexts in which the term arises. Certainly there is some sort of aggressive lexicography that you propose or employ when rendering the Chinese text in a way that actually makes it meaningful for argued discourse. Is it enough to simply recommend extreme parsimony, and then to provide a smaller group of definiens from which students would be tasked with qualifying one definiens over another for a given passage?
I only know this via formal proofs on the Daodejing: If you devise a definiens for 道 that is immensely broad, then the whole metaphysics of Daoism becomes a series of rather obvious truisms, while its ethics begin to ring of deontology. However, if you reduce the definiens to something narrower (to something that is less modally rich), then the Daoist ontology is foreign, but arguable, and the ethics becomes far more cryptic. I’m far more in favor of the former definition, namely because it has captured all of the text that I’ve investigated thus far consistently, but that preference comes at a cost of reducing the overall interest in Laozi, since a bulk of his work then produces rather vacuous and often tautological propositions.
Hi Joshua; thanks for the comments. I’m not for extreme parsimony in principle. But it can be a good pedagogical heuristic for getting students to sharpen their interpretive skills.
Actually, now that you bring it up, with the Daodejing specifically, I think the text itself invites at times broad and at times narrow interpretations of dao. But I think what the Daodejing does with the term shouldn’t determine how we understand it as used in other texts, if we think it is a text that is meant to subvert more conventional normative responses to ideas associated with the term. (I know I’m responding tangentially to your comments, but it’s what your comments made me think.)
As I think Manyul suggests, we might want to use untranslations as a way of trying to introduce into English some terms we think English should have. Of course, the terms would shift in meaning within English, which fact might in turn spread misunderstanding of Chinese texts. Using the untranslated terms widely in translations would tend to reduce the divergence.
I think salient marks of disambiguation such as Hansen’s superscripts can impede the learning of the Chinese concept (though they can also be very convenient, especially in interpretive discussion). In the original Chinese, disambiguation (when appropriate) is done mainly by context, and a translator can usually preserve that situation in English, thus forcing a reader to understand ‘dao’ more quickly and intuitively and accurately than if she habitually divided ‘dao’ into several distinct terms such as ‘dao: path’ and ‘dao: speak’.
Joshua, you write, “From a pedagogical perspective on any topic, I would be personally disenchanted to simply hear, “We don’t translate t because t means lots of things,” since, as you put it, it strikes me as a cop-out against investigation of those contexts in which the term arises.” I agree that if one heard only that, it would be a cop-out. But more could be said in preface, notes, and running commentary. And shouldn’t a translation’s direct representation of the text be to some extent a cop-out, allowing for various old and new interpretations?
Joshua, I agree with your criticism of using too many alternative terms to translate a single original term, but I’m not sure I grasp your argument against the opposite extreme. Perhaps this paraphrase is fair: “If a translation encourages readers to think that a certain word used in two places is used in the same sense, when it’s really used in two distinct senses, the result may be to encourage readers to find the text inconsistent, because it would be inconsistent if it were using the word in the same sense in both places.” If that’s the argument, then I think it’s not a very strong argument against e.g. leaving ‘dao’ untranslated, so long as readers are warned about the range of its possible meaning. We don’t think we’ve encountered a contradiction when someone shouts, “This meeting has one chair, and one only! So you two sit back down in your chairs and I’ll sit back down in mine.” An advantage of leaving e.g. ‘dao’ untranslated is that there may well be room to debate whether what appears to be ambiguity is in fact ambiguity. If we think a “path” is a course of action and “guidance” is a kind of speech, then the two will seem quite different; but if we remember that literally a path is not the walking but is a kind of guidance (cf. Mencius 7B21), and right patterns of action also amount to guidance for our neighbors and subordinates, we may decide that the term ‘dao’ is not so ambiguous as we had originally thought, or anyway not neatly divisible into two senses.
The second extreme assumes that readers will associate a single referent to a single term.
The issue is perhaps best explained via an analogy of learning new words of any language through its context. When we explore the contexts of uses of words, we often try to keep alternatives to a minimum so that we can isolate a single definiens more regularly. Unless we encounter sentences that appear openly contradictory (as we do in your example), a reader would not have a basis from which he could assert that a common order of graphemes was actually two distinct terms, abbreviated. It would be uncharitable of a reader to just dismiss a term as contradictory because his semantics was too stubbornly narrow, but it remains a real risk when presenting a text.
This is worsened by pervasive myths about the Chinese language, one being that for every concept in their discourse, the Chinese invented a character to match it.
What I’ve found is that one could interpret 道 in the Daodejing with two definiens and arrive at conclusions radically different from those wherein there is only one definiens, even though the text remains coherent and (so far) internally consistent with either option.
Manyul made a point that doing this by text would be rather narrow, but as a sort of emergentist approach, I would promote coming to a clear understanding of a term via some narrower collection of sentences, and then gradually introduce new texts to see how much muddier the semantics are against an already constructed definiens (or two).
I have to agree with you that my ‘chair’ example was too special to support my general point. But context can inform us in many other ways (beyond potential overt contradiction) about the likely meaning of a word, and of likely polysemy. For example, context can tell us that a word is a noun here and a verb there. Context can also suggest that the noun and verb forms are closely related, as in the opening lines of the Daodejing. Context can tell us that F here is a feature of cats and there is a feature of numbers, or here is something wished-for and there is something feared, leading us to be more or less friendly to the hypothesis of ambiguity. Nor do we need to resolve such questions in order to understand what’s being said: there’s debate about how ambiguous ‘good’ is in ‘good person’, ‘good food’, ‘good painting’, ‘good at math’, etc.
Just as importantly, the translator’s introductory notes on main terms can and should make some general suggestions about the likely range of meaning of key terms.
One reason people may like to hypothesize ambiguity is that it’s easier to think of a meaning that fits one context than a meaning that fits two. So for two different contexts we often start out with two different hypotheses, and people like their first impressions.
Granted, as semantic contexts go, the Daodejing is pretty far on the uninformative side: it’s abstract, it’s brief and nondiscursive, and it revels in surface contradiction.
Context can be informative, but a la Austin, I am attracted to a sort of “circle” that then encompasses some general idea. From that, I opt for the term that is most likely to be judged as hypernymous of that synonymous group. It’s one reason why I prefer translating 道 nominally as “way” over “path” or something else. While anecdotal only, I think that people agree more commonly that paths are ways, but would need more convincing and clever metaphor to be convinced that ways are paths.
Some of the context is more syntactic than semantic, as you say. For some interpretations of the sentence structure of ancient Chinese, 道 doesn’t seem to be working as a noun in the sentence, and since we English speakers don’t have any word like “to way”, we have a motivation to find a fitting verb, perhaps even at a semantic loss. That in itself is condemning to all language translation, though, but something that could be overcome to a point just by allowing awkward-sounding sentences to serve as an accurate, though unattractive translation (we could use copulas and end up with something like “to be a way”, as I think some interpreters [Watts?] have done).
Now, opting for “speak”/”utter”/”say”/”tell” verbal translation must be a semantic consideration, built on some mass of texts where 道 is more apparently being used in a way that appropriates that shift in definition. What is the philological or lexicological basis for such a translation, especially when words like 曰 and 言 have common semantic roles? I’m just not swayed by the idea that a (Swoon!) analytic language in its infancy had really matured to a point of introducing multiple characters to account for greater definitional subtleties like the ones various translators offer. This bodes even worse when the term in question appears to be cogent when just the single definiens that is then extended to different syntactic functions appears to offer legitimate translation (as far as my rather slim coherentist criterion is concerned). I’m openly appealing to “desert landscape” translation, but actually more is afforded by this than one might think if he were to demand the sort of exact POS matching that translations seem to seek, which is more like phantom hunting exercise than a real translational pursuit. Commentary on translation and suggestion for alteration could be reasonably proposed, as well, but so often in that gig there is an impetus not to build from another’s rendition, but to start practically tabula rasa (I blame copyright law for that one, and that’s an argument that really needs its own thread [and a forum in which to put it…hint, hint].).
Now, my fascination and experience with various syntaxes may point somewhere relevant. Modern Chinese has a few regular ways to reverse verbal/predicative (that is V>O) forms to nominal terms by reversal (O>V). Korean apparently has this sort of transitional allowance in both directions for a great chunk of it’s nominal lexicon via a ‘…하다’ (and perhaps also ‘…고 있다/없다’) introduction/removal. English and Chinese, however, have a trend of dual POS categorizations by brute force of use (e.g. “(to) murder”:”(a) murder”, “謀殺”:”謀殺”). As I said before, though, these are syntactic considerations. The semantics of the issue really belong to the proposed interrelations of terms, and that has to appeal to all sorts of expectations from speakers and demands for reasonable semantic framing from those who are more critical of the relations that a text is proposing, particularly when a work is donning robes of truth.
I have all sorts of motivations, some pragmatic and some idealistic, for opting for inter-translatability of all terms, which I think that I could illustrate a sort of nightmarish scene for any lay reader of translated texts. If we consider a great expert in a foreign language (the Dworkian “Hercules” of translation from Lx to Ly), his knowledge of both languages would be quite thorough, enough even to work around the syntactic hurdles and most of the semantic entanglements of the language. If such a translator were, for external reasons, compelled not to translate a given term because there was simply no adequate way to do it without, say, rendering the term into a half-page dictionary entry (this being one reason external to the accuracy of the translation, itself), it seems like such a case could be made against a great many of the terms that are just “so much deeper in significance” than the other language can accommodate. Disaster seems to follow from this, since our Hercules could dismiss many, many more terms than just those central to a certain philosophical debate, historical record, etc. How could such a person have confidence, given his tremendously high standards for accuracy in usage, to translate much of anything? I think the cruelest translation in the world would be one wherein the translator stated in his preface or introduction, “Well, t is kind of captured by P, but its qualifications are pretty long, so hopefully you’ll acquire my Herculean comprehension as you read and pick up the meaning from the context of this work, alone” (remembering, of course, that the Herculean translator will have a much bigger well from which he poured his knowledge of the immense difficulty in t’s concise translation).
It would be far preferable for a lay reader (or maybe just for me and other enthused readers of a text) to have even a supplemental, expert translation dictionary, so insofar as the reader engaged the text, he could always investigate further into the Herculean Lx-to-Ly dictionary and, even if the translator used the untranslated grapheme (e.g. 道) for it (or, hell, even some other wacky symbol to invoke its sentential use), it would at least still qualify as a translation, and the term would still have a clear and finite definiens.
I really hope this helps clarify exactly what I would seek from a translation. I don’t really care about the length or scope of a single definiens, or even too much of multiple definiens (which I would address on a case-by-case basis), but I think that only through maximal parsimony and maximal coherence does a translation give fair reading to a text’s truthful parts and to its nonsense. There’s a lot of special pleading that aims to avoid nonsense that doesn’t appear to fairly address the actual ratio of t’s invocations to t’s invocations with a given unique definiens.
There’s an awesome special pleading trick that happens here involving people just importing definitions to preserve the text’s coherency or cogency against observation, but really just sneakily revising what the text states in order to rescue it from refutation. On not-so-rare occasions, they can then hope that the special exception will be forgotten so that they can simply recommit the same conflation and sustain and propagate the same refuted conviction. But that’s a separate nightmare…that we unfortunately live.
Joshua, thanks for your reply. You point to some profound issues for any project of interpreting or accurately translating a text. You are right that I glossed over some things, such as difficulties about the noun/verb distinction in Chinese. I meant to be talking about how readers (real ones, casual ones, rushed students) actually interpret, toward a narrow practical question for translators; not about the principles on the basis of which we should interpret. But the way I expressed myself invited misunderstanding on that point.
I take it we’re not in any apparent disagreement now on the value of leaving the occasional key term untranslated.
Actually, that over-self-critical Hercules scenario was supposed to point at the problem of leaving terms untranslated. My position is that I prefer a bloated definition in translation over none at all. In this way, rushed readers could gloss over the more exacting meaning for its term, while interested ones could inspect, evaluate, criticize, and revise them at some given times. In that scenario, a text is translated, but the translated term is not necessarily reprinted on every page in which the term occurs.
I agree with the value of context in establishing alternative definitions, and you’re totally right that context alone can give us enough knowledge of the likely definitional predications that lead us to assert ambiguity of a term (so long as it is a regular occurrence), but I would accuse special pleading against a translation that opts for more definiens when fewer appear to keep a text consistent.
Thanks Joshua! Sorry I misunderstood. I’m still not sure I understand. I think maybe you’re saying this: “Better than to leave a term untranslated throughout is to find one or a small set of rough equivalents in English and provide more accurate definitions in the book’s vocabulary list, even if some readers will neglect these definitions.” Leaving a term untranslated would indeed make the book more daunting at first glance, if the term is left in italics so that a browser would notice. And I think your argument is that it would make the book seriously more daunting in fact. Maybe that’s right.
You write, “I think the cruelest translation in the world would be one wherein the translator stated in his preface or introduction, “Well, t is kind of captured by P, but its qualifications are pretty long, so hopefully you’ll acquire my Herculean comprehension as you read and pick up the meaning from the context of this work, alone” (remembering, of course, that the Herculean translator will have a much bigger well from which he poured his knowledge of the immense difficulty in t’s concise translation).”
Yes, a translator who leaves a term untranslated risks having the reader think she’s being lazy and abusive. But the task she gives the reader is different from the task she is shirking herself. Learning to use a term well is easier than figuring out how to define it accurately, and figuring out how to define it is different from figuring out how to translate it. In the preface the translator might just as well talk of her own limitations and her confidence in the abilities of the reader, as well as the value (and fun) of learning to use new ideas, new mental tools. She might remind the reader that we learn new terms each day in the same way, and or even that small children do it all the time without much thought – while at the same time warning that the text presents certain interpretive difficulties so deep that the experts disagree and a translator should leave matters open for the reader’s judgment. Her vocabulary list need not offer a discrete number of definitions for the key terms; it can give other kinds of help that might be more engaging, such as a discussion of some of the core metaphors and how they might relate to one another.
The term ‘li’ 禮 reflects an assumption that some kinds of activity that strike modern Westerners as formal and arbitrary are in fact sorta paradgmatic of good behavior generally. That assumption, and a sense of its ordinariness, is integral to the use of the term. I think it’s easier to grok that idea, and not lose sight of it, if we get the new term to go with it. And if the idea is a valuable one, then using ‘li’ directly in translations might perform the important service of bringing the idea further into currency, which could be worth doing even if the books themselves are a little more daunting at first.
I do think you’re right that the risk that a reader will wrongly attribute an inconsistency to a text is increased by using fewer terms, and that is one point against the minimal extreme. I just don’t see that as a powerful enough consideration to show that the minimal extreme (for a few terms) isn’t best. Beyond the reasons I’ve given above, I’d add especially your own argument against using too many terms: it blocks the reader’s perception of connections among different parts of the text.
Hmm. More than the myth you describe, maybe the source of the risk you mention is in readers’ general attitudes toward technical terms. When a new and alien term is introduced, especially in classroom text material, we adopt toward it an attitude different from the general attitude we have as children toward the words we hear. We think of it as a term of art, an artificial term, and we may even bring into play a sort of Theory of Meaning: that the meaning of a term is captured in a Definition, which in the case of a term of art would be a unique definition.
That strikes me as a real problem for the teacher or translator. A teacher can address it by what she says in class, and perhaps by using the untranslated term casually in class to encourage students to hear it casually and use it casually themselves. A translator can address it by introductory discussion and by underplaying the foreignness of the term: not using italics, and using endings (daos, daoed, daoing).