Analogical Reasoning — not just ordinary inference?

Some commentators seem to regard “analogical reasoning” to be a distinct form of reasoning in the early Chinese philosophical context, attributable mostly to the later Mohists and to Mencius. So far as I’ve been able to understand it, the distinctiveness of such reasoning seems suspect; but I’m not 100 percent sure I’ve understood it correctly.

Part of the issue seems to be that we get an account of the method or technique of reasoning in the later Mohist canons, but we get what some people consider the most explicit, or at least self-conscious, use (or misuse) of the method in the Mencius. “Argument” in Mencius, according to most commentators, seems primarily to work on a model of analogy.

There is the statement from 3B9 of the Mencius that Mencius sees himself as engaged in bian 辯, “disputation,” in response to the Mohists. There are technical discussions in the later Mohist writings about the various modes of bian. Among them is the technique called tui 推, “to push” or “to extend.” As A. C. Graham translates the definition of the technique (A. C. Graham, Later Mohist Logic, Ethics and Science, 483), to tui is “…using what is the same in that which he refuses to accept and that which he does accept in order to propose the former” (tui ye zhe yi qi suo bu qu zhi tong yu qi suo qu zhe yu zhi ye 推也者以其所不取之同於其所取者予之也). So, as the later Mohists understand it, in this technique, one “pushes” the opponent’s judgment from a case he already accepts to another case, using something that both cases have in common. Hence this has often been called the method of “analogical inference.”

But now, consider the following reconstruction of so-called “analogical inference,” using the example from Mencius 1A7 that both Graham and Nivison examine as an example of such inference. The example is Xuan’s compassionate response to the ox being taken away to slaughter and the analogous case of his people, the people of Qi, who also seem to require compassionate response. Graham, Nivison, and most others think there’s an inference, then, that Mencius encourages Xuan to make.

Here’s what it seems to me to take, to go from the one case to the other:
1. Xuan ought to feel compassion for O, because O possesses F.
2. Q also possesses F.
3. Therefore, Xuan ought also to feel compassion for Q.

But it seems like we can further analyze 1 as relying on a suppressed premise, one which says something more general about feeling compassion for objects possessing F. If we push the analysis further, the entire inference involved in the case of this “analogical” inference seems most plausibly to be the following:
A. Xuan ought to feel compassion for objects that possess F.
B. O possesses F.
C. Xuan ought to feel compassion for O. (from A, B)
D. Q (also) possesses F.
E. Xuan (also) ought to feel compassion for Q. (from A, D)
A-E seems to spell out more explicitly how the Mohist principle of extending, which is the principle of applying rational pressure to treat similar cases similarly, would actually apply the rational pressure of maintaining consistency among one’s judgments. But then the technique of tui 推 is not so distinct or different from regular, boring inference. This raises two questions for me:

Q1) Have I misunderstood or misconstrued the technique?

Q2) Was I looking for something unreasonable in looking for something distinct or different in a reasoning technique in the first place? In other words, should we think, “Of course it turns out that analogical reasoning is just a familiar form of reasoning!”?

90 replies on “Analogical Reasoning — not just ordinary inference?”

  1. Yet another great post!

    A worry about argument A-E is that steps B and C do no work toward supporting E.

    If we read 1A7 as arguing for the conclusion that Xuan ought to feel compassion for the people of Qi on grounds F from the premise that he ought to feel compassion for an ox on the same grounds, then we might build the argument this way:

    1a. Xuan ought to feel compassion for O because it possesses F.
    2a. Probably Xuan ought to feel compassion for whatever possesses F. (1a)
    3a. Q possesses F.
    4a. Probably Xuan ought to feel compassion for Q. (2a,3a)

    I have purposely omitted the comma from 1a. Hence the intentional action described in 1a is more complex than the action described in Manyul’s 1, and 1a offers no justification. 2a proposes the simplest principle that can explain 1a.

    If we read the argument as a bit of Mohist disputatio (sic) involving the move tui 推, then I guess we should cast it in one of these two ways:

    1b. Xuan grants that he ought to feel compassion for O.
    2b. Q is similar to O in possessing F.
    3b. Xuan should grant that he ought to feel compassion for Q. (1b, 2b)

    1c. Xuan grants that he ought to feel compassion for O because it possesses F.
    2c. Q is similar to O in the respect that Xuan cites.
    3c. Xuan should grant that he ought to feel compassion for Q. (1c, 2c)

    Here’s a very different way to look at the conversation in 1A7, based on a hypothesis. Hypothesize for a moment that Mencius sees compassion as being or involving a perception of the undesirability of someone’s actual situation (those groaning and shuddering under Xuan’s yoke) or potential situation (at the bottom of a well). The heart of compassion would then be a cognitive power. Mencius might use the case of the ox to show Xuan that he has this power, and then suggest that he apply it in other cases. Mencius’ project would then include showing Xuan that he ought to help the people, without presenting any argument for that at all. And since the claims on our help can be overridden by other and stronger claims, nobody need be committed to the idea that the ox should have been saved. As for whether a properly developed faculty of compassion would feel significant compassion for the ox, the hypothesis leaves that question open.

  2. My thought in the last paragraph in #1 was that compassion might be a kind of sympathy that could be aroused by the imaginative exercise of thinking about someone’s condition. Also I was assuming that recognizing the undesirability of someone’s condition amounts to seeing a prima facie reason to help.

  3. Maybe we can also throw PJ’s interpretation in from his “Confucian Self-Cultivation and Mengzi’s Notion of Extension” in Essays on the Moral Philosophy of Mengzi .

    If I’m reading him correctly, analogical “projection” is not about identifying some common feature F, but a common spontaneous reaction (‘compassion’ in case of 1A7) which can be easily drowned out without learning to recognize and cultivate it.

    Using a music metaphor he explains that Mencius “gets the king to experience the chord (i.e., the feeling of compassion), understand how it differs from other chords (e.g., other emotions that might have motivated such action and similar emotions that have moved him in the past), and develop the ability to identify and attend to it (i.e., focus upon it, get a feel for it, and appreciate the satisfaction associated with it)” (232). In new circumstances one reconstructs the feel of the old experience. “In the same way, Mengzi hopes the king will pick out his compassionate feelings from among the other strong emotions, i.e., his desires for conquest, sex, wealth, and power, which prevent him from caring for his people” (232).

    Unfortunately I have to run and so don’t have time to evaluate his position, but I thought I’d toss in now to see other’s thoughts.

  4. This isn’t really a good test case for the idea that there’s a distinctive use of argument by analogy in some early Chinese texts. Mencius is portrayed as thinking that the king is wrong to feel compassion for the ox, so any argument by analogy is going to reach the wrong conclusion.

    I’d also resist equating tui in the Mencius with tui in the Later Mohist Xiao Qu. The Later Mohists describe tui as “proposing on the basis that what they do not accept is the same as what they do accept.” It’s an argument based on a similarity between views (namely a view the opponent does accept and one s/he doesn’t accept), but acceptance of a view is never what’s at issue in Mencian references to tui. If these are arguments of any sort (I don’t think they are), then they’d be better associated with what the Mohists call pi, or comparison, “mentioning another thing and using it to clarify this.”

    All four kinds of arguments that the Later Mohists discuss in the Xiao Qu turn on (different kinds of) similarities, so it’s reasonable to think of them as arguments by analogy. Tui (pushing) and yuan (pulling) are two familiar kinds of ad hominem argument; and pi (comparison) looks like the familiar kind of analogical argument. Mou (parallelism) is the only one that really seems unfamiliar, and it’s the one the Later Mohists seem to understand least (though they do devote a large part of the Xiao Qu to showing that it doesn’t always work).

    The place to look for a distinctive use of analogy is, I think, in the frequent appeal to paradigm cases in early Chinese argument. (I learned this from Chris F, who I hope will be around to comment.) An example in MC 1A/7 is the use of the cases of lifting a feather and carrying Mount Tai across the northern sea to illustrate the distinction between not-doing and not-being-able.

  5. Oh, and a corollary to Bill’s opening comment. In an argument by analogy, if you can spell out what shared features of the cases being considered ground the argument, this can put you in a position to reformulate the argument as a deductive argument. You (Manyul) have in effect done that in your reconstruction, and that’s why what you end up with isn’t an argument by analogy.

    (It is interesting that early Chinese texts don’t try this reformulation even when the grounds of their arguments by analogy are explicit. A nice case in point is the anti-war argument in Book 17 of the Mozi, where they consider a series of cases in which an action harms others and is therefore wrong, in order to set up the claim that aggressive war is wrong because it harms others; it would be easy to reformulate this as a deductive argument, perhaps with inductive support for the major premise, but the Mohists never do that.)

  6. Dan says: “I’d also resist equating tui in the Mencius with tui in the Later Mohist Xiao Qu.” I agree with this. If Mencius is using tui in the Later Mohist sense, then it will look like Bill’s 1b~3b or 1c~3c in Comment #1.

    Dan, you also say: “Mencius is portrayed as thinking that the king is wrong to feel compassion for the ox….” I’m inclined to disagree. What evidence is there in the text? I’m supposing that you are perhaps referring to “a gentleman keeps his distance from the kitchen”. My impression is that that might be a popular saying, or a rule of ritual like the ones listed in Analects Bk.10, meant to illustrate Mencius’s point that gentlemen cannot bear to see the suffering of animals.

    Manyul writes: “the principle of applying rational pressure to treat similar cases similarly, would actually apply the rational pressure of maintaining consistency among one’s judgments.” I’ve been wondering about this a bit, and I don’t think tui is just a matter of being consistent. If it were, Mencian extension would be symmetric in the following sense: one might extend one’s concern for the seen ox to the unseen sheep, or one might extend one’s lack of concern for the unseen sheep to the seen ox. But that is not how Mencian extension seems to work: it is asymmetric, in the sense that concern for nearby objects are extended to concern for distant but relevantly similar objects. (In short, the deliberative rules involved in Mencian extension are not just “Be consistent!”, but also “Extend concern!”–Mencian extension I think involves primarily the latter rule, and involves the former rule only in a subsidiary manner, in the sense that consistency is implicit in any form of reasoning.)

    I’d like to insist that the transfer of concern from the proximal object to a relevantly similar distal object is not reflexive and automatic; instead, it is a reflective and deliberative process. I’d like to see it as a distinctive form of practical reasoning, distinct, that is, from instrumental reasoning. Instrumental reasoning is concerned only with discovering means to given ends, but Mencian extension expands the range of our concerns (ends) via a deliberative route.

    p.s. Bill et. al, I regret my intermittent posting, I really should be posting 17 minutes a day or so according to my policy, but Han Fei is right about people being lazy slobs.

  7. Hi Boram,

    “A gentleman keeps his distance from the kitchen (君子遠庖廚)” also occurs in the Liji, fwiw. Certainly whoever wrote 1A/7 is taking this to be a bit of wisdom that the audience will recognise.

    But—the author also portrays Mencius as endorsing that piece of wisdom, and providing it with a rationale. And the rationale is precisely that the gentleman will feel inappropriate compassion if he enters the kitchen.

    (Maybe we need a distinction. Of course it’s supposed to be appropriate for the gentleman to have the sort of character that will cause him to feel compassion if he witnesses an animal’s suffering; it’s the activated compassion and not the disposition towards it that’s being rejected as inappropriate; that’s why the gentleman is to stay out of situations that will trigger it, and not to reform his character.)

  8. I’m actually on Dan’s side about 1A7; I don’t think the idea of extending compassion (tui en 推恩) is an example of “tui-ing” in the Mohist sense. As Dan says, other examples do appear in 1A7 that are more accurately and *competently* cases of Mohist “tui-ing.” Here’s a snippet from my 2002 paper, discussing the feather-lifting analogy that Mencius brings up in 1A7:

    Cases A and B:
    A. If one can lift a heavy thing, then one certainly can lift a light thing.
    B. If one can see a tiny thing, then one certainly can see a large thing.

    Principle illustrated by A and B:
    P. If one can do a difficult task with a capacity, then one certainly can do an easier thing with that same capacity.

    Case C:
    C. If one can feel compassion for a mere ox, then one certainly can feel compassion for one’s fellow humans.

    The analogy:
    Case C is like A and B in that it is also a case which falls under P: One has demonstrated the ability to do a difficult thing with a capacity—feel compassion for a mere ox—and so, has demonstrated that one is capable of doing an easier thing with that same capacity—feel compassion for one’s fellow humans.

    The inference:
    If one affirms A and B in virtue of being cases which fall under the principle P, then, since C has in common with A and B that it falls under P, one must also affirm C.

    So let me offer, two clarifications, I guess.

    The ox and people of Qi case isn’t really the right example. However, other, more influential people–principally I’m thinking of Nivison and Graham–fix on that case and try to squeeze it into the Mohist formula. That’s a mistake, and I tried to argue that in the 2002 paper. Ivanhoe’s position (see Agui’s comment #3) on the case seems better in not trying to attempt that squeeze.

    But I’m still interested in the issue, following Dan’s (comment #5) formulation, about why the shared feature of the analogous cases is not foregrounded and then formulated into some sort of principle to apply to both, and to other, cases. If you look at the Mohist formulation of what tui-ing involves, the natural thing to do would seem to be to “bring up the thing” (予之也) that the two cases have in common. Isn’t it puzzling that the step of formulating a broad principle isn’t taken, given the momentum of the tui-ing method toward isolating the important common element in different cases? Am I being too Western here? (And what does that mean here?)

  9. I share Manyul’s puzzlement about the absence of an account of the common feature. In 6A1f, I wonder whether the analogies Gaozi proposes are meant to make a literal point, or are instead simply proposed as proto-conceptual schemes.

    Regarding tui 推 and other such Mohist rules of disputation, Chris Fraser gives a very nice backgrounder here:

    I’m puzzled about the relation between tui 推 and yuan 援. Here’s the account of each, with Chris’s translation:

    Tui: “On the grounds that what they don’t accept is the same as what they do accept, propose it.”
    Yuan: “Saying, ‘You are so, how is it that I alone cannot be so?’”

    It looks as though while tui is about neutrality or universality over different kinds of case, yuan is about objectivity or neutrality over persons.

    But maybe that’s not right. The above interpretation suggests that yuan is a kind of application of tui, but the terms tui 推 (pushing) and yuan 援 (pulling) suggest that the two moves are supposed to be mirror images of each other in some way. Another possible reading is that tui is proposing something new that fits what the opponent has accepted, and that yuan is challenging the opponent’s rejection of something that is similar to what the opponent has accepted. Or vice versa. As inference rules these would be identical, but as debating moves they would be distinct. (I think the term “ad hominem argument” normally means an argument of the form “P is false because the person who said it is bad” – a form of inference Confucius opposes in 15.23.)
    Here’s a discussion of formalized debating or disputation in the medieval West:

    Dan, is it certain that the Mohist disputation rules are about views/beliefs? Couldn’t they be about sayings (and then anything equivalent to sayings)? An imperative saying might be equivalent to an attitude, or only trivially nonequivalent to a practice? Either could be shi-ed or fei-ed?

    Dan, I’m not sure an argument by analogy has to begin with something that is still accepted at the end of the argument, and I think the ox argument might be a good example of what I mean. Here maybe I’m using “argument” in a broad sense; I don’t know. I’m thinking of this kind of argument that might be addressed to a young person:

    1d. You think all your desires should be satisfied (except where they conflict of course).
    2d. You do not base this on any difference between you and other people.
    3d. Therefore, to be consistent, you should grant that all desires should be satisfied (except where they conflict of course).
    4d. Therefore the view reported in 1d is wrong (since some of your desires conflict with other people’s desires).

    The idea is that the person has always rightly felt the claim of desires, but through inattention to other people or to their personhood has neglected to notice the scope of the point.

    Manyul argues in his 1999 paper that Mencius seems to signal that it is virtuous to feel compassion for the ox if you see it (such a disposition is ren shu 仁術), but right to set things up so that you won’t see it. Hence Dan would be right to say in #7 that Mencius is signalling that it’s inappropriate, at least in general, to feel compassion for properly doomed oxen, but perhaps Dan is not right to infer that Mencius is signalling that in the case where Xuan has in fact seen the ox he was wrong to feel compassion. Dan, I imagine you mean to be challenging Manyul’s virtue-theory assumption that the expresssion of a virtuous state is necessarily itself virtuous. But I wonder whether there are good grounds for such a challenge in this passage.

    Dan, offhand I’d guess that it’s at least as likely that the kitchen line is in the Liji because it’s in the Mencius as the other way around. Do we know anything about the origin of that book of the Liji (Yu Zao)? Legge has nothing to say about that in his introduction.

    Dan, I don’t understand why you say a deductive argument (even one with merely inductively reasonable steps) can’t be an argument by analogy. How about this:

    1e. P is true.
    2e. Q is analogous to P.
    3e. Q is true. (1e, 2e)

    Is the worry that that’s an invalid deductive argument? Or is it something else? And if an argument by analogy is essentially different from a deductive argument, then why exactly can’t an argument by analogy have a false premise?

    Boram, I really like your important point about symmetry. Maybe the solution is that an argument by analogy should address doubtful cases on grounds of more certain cases. I don’t think that means that the starting-point has to be preserved exactly, but it suggests that the starting point should be preserved in broad outline. I wonder whether it requires that.

    Boram, I had all my hopes pinned on your being too lazy to stop each day after just 17 minutes. Come on, prove that you’re not.

  10. I think I misreported what the 1999 paper says about the goodness of Xuan’s compassion for the ox. I should simply have said this:

    It’s not clear to me that Mencius is saying that once Xuan has seen the ox he was wrong to feel compassion for it. He holds that a good person would feel compassion in such a case, so good people should avoid such sights, that’s all.

  11. Interesting discussion. I’ve learned a lot from it.

    I don’t think we should be surprised that analogical reasoning turns out to be familiar or that it can be reformulated in a (quasi)deductive form. Informal logicians cash out arguments by analogy much as Manyul cashed out Mencius’ argument.

    What’s surprising or different about this, as compared to Western approaches, is just that, F, the similarity between O and Q, is never “foregrounded” or explicitly spelled out. Plato would have spent an entire dialogue trying to figure out what it was, but Mencius leaves it vague. One possible reason for this is that stating a principle like ‘All F things are G’ invites counterexamples (which is what Socrates would have given us). Leaving F unstated allows for an intuitive understanding of the nuances of the particular things being compared, so that we can make progress on this particular case, even if we don’t have a Socratic definition of F-ness. Whether this is, on the whole, better or worse argumentative practice is a hard question.

  12. David, you’re right that general principles are likely to invite counterexamples. But I think use of an analogous case, even without a general principle, still invites things that function sort of like counterexamples–namely, the counter-analogy (bringing in a third case) or, what is perhaps more damning, arguing for the stronger dis-analogy between the two cases already on the table. I think the disputes with Gaozi in Mencius 6A 1-3, and maybe the disputes in 4-5, display this sort of thing. So, the absence of movement toward generalization still seems curious to me.

  13. (Bill, this relatively long post is dedicated to you. ;))

    I’ve finally downloaded Manyul’s two papers, which I will certainly enjoy reading.

    Dan, thank you for tracing that saying to Liji: “A gentleman keeps his distance from the kitchen (君子遠庖廚)”. That’s very helpful to know.

    Now, though, I still disagree with both Dan and Manyul on the import of that saying, as Mencius uses it. As I read it, the import is NOT: the feeling of compassion (on that particular occasion or in situations of that type) was inappropriate.

    I’m still looking where the inappropriateness is suggested in the flow and gist of the surrounding passage. I can see the implicit hypocrisy (the kind of hypocrisy I used to accept when I used to eat fish and meat, though I felt sorry for the animals, and still accept by eating dairy products). I see the suggestion that, if King Xuan felt sorry for the ox for the reason that it was innocently being led to death, then by the same token he should have felt sorry for the sheep as well. But I don’t see Mencius suggesting that feeling or acting on the compassion (either on that occasion or for animals slaughtered for ritual purposes or whatnot) is inappropriate. Perhaps that can be read between the lines, and if you press Mencius on the hyprocrisy of the gentleman he might say it was inappropriate to act on the compassion for animals because the ritual, or serving your aged parents meat, is weightier than our concern for animals (using quan). But let me insist that all that’s besides the point in the surrounding passage.

    Here’s how I read the immediate surrounding passage. King Xuan is wondering why he felt sorry for the ox but ordered the sheep to be slaughtered instead. Mencius explains it, by noting that he saw the ox but didn’t see the sheep, and that the feeling of compassion is activated by what is immediately before one’s eyes (Hume would agree with all this, and also with the further unmentioned point that the feeling can be transferred to less contiguous, more distant objects by reflection, though weakening as it extends outward). Then, to support the point that the feeling of compassion is activated by what’s immediately presented to the senses, he notes:

    (1) 見其生,不忍見其死
    (2) 聞其聲,不忍食其肉
    And to further support (1) and (2), Mencius invokes:

    Staying away from the kitchen may be the appropriate thing to do, but I hope you see why I claim that’s besides the point in the surrounding passage.

    So, to go on to other matters. As Dan and Manyul pointed out earlier, the lifting-feather/seeing-a-wagonload-of-firewood analogy is more apposite example of Mohist tui. Mencian extension is quite different, and as I noted it is asymmetric. It’s never the case that one’s unfeelingness to distant concerns is to be transferred to nearer concerns:



    The direction in which one transfers one’s concerns here is outward. (The last passage of Analects Bk.6 talks about “仁之方”, which may be read, not just as the method of ren, but also the direction of ren, as Irene Bloom translates it. The direction is from self to others, from near and dear to distant, from in-group to out-group. I would go so far as to claim that 近思 in 19.6 should be read as “reflecting from the near and dear [to the more distant]”. I wonder if that’s an intemperate reading, not warranted by grammar and context, but I suspect the Neo-Confucians might agree with my interpretation. In the Neo-Confucian compilation Jinsi Lu, Zhang Zai (if I remember correctly) pushed his feeling of compassion outward to such an extent that he did not cut the grass growing outside his window, because he felt towards it as he did towards himself. And Zhang Zai or another of these hyper-sensitive Neo-Confucians heard the cry of a donkey being whipped and his heart ached with it. I wonder what Zhang Zai ate.)

    Anyway, let me make a tentative go at addressing Manyul’s puzzle as to why the commonly shared feature is never brought to the foreground, specifically in reference to what I take to be Mencian extension. My suggestion is that in extending one’s concerns beyond the normally accepted boundary, one’s conception of what is of the same kind also shifts, and becomes broader. I do think Mencian tui naturally extends up to but not beyond whatever’s capable of feeling pain and pleasure, but perhaps I’m wrong, thinking back to Zhang Zai’s case.

  14. Boram, thanks for your dedication and your comment!


    I read a bunch of Manyul’s pieces a few months ago and loved them, and have recently been discovering that my memory of their contents is minuscule and distorted. I don’t know how many of my views are really his.

    I’m not tempted to think Mencius had the (Later) Mohist technical term in mind in 1A7. My main reason is that he has left no comment about disputation or its rules. Another reason is that I can’t see the rhetorical value of alluding to disputation rules in talking to Xuan (in a real or fictional model dialogue), especially if Mencius doesn’t discuss these with his students. Manyul also points out in his paper that in English “extend compassion” means “exercise compassion”, and that Mencius says in 1A7 that having compassion even for one’s closest kin requires tui-ing it.

    My reasons don’t include the idea that the move in the Mencius doesn’t fit the Mohist term. I think it kinda fits. The question whether or how closely it fits is interesting in connection with one of our topics in the thread, viz.: what kind of animal can analogical reasoning be? (Though as Dan points out in #4, tui hardly exhausts even Mohist conceptions of argument by analogy, and the ox-tui incident is in a way not even salient as a Mencian use of analogy.)


    The Mohist move seems to be about consistency, but not exactly about avoiding contradiction. I imagine it’s to be applied in cases where nobody is committed to any clear general principles as being the relevant ones. Smith has committed herself to “A1 B” and perhaps has given some loose account of what it is about A1 that supports “A1 B”. Jones then proposes that Smith ought then to grant “A2 B,” either because A2 seems to fit that loose account or because A2 just seems offhand quite similar to A1.

    If that’s tui, then, as Manyul says on p. 241, tui is primarily Jones’ debating move rather than Smith’s changing her mind in acceptance of that move. I don’t know whether either party’s move is a kind of “inference,” mainly because the Mozi’s and my accounts don’t pin down precisely enough what the move is, but also maybe because of the fuzzy borders of the term ‘inference’ (about which I’m finding my linguistic intuitions are fluid).

    Insofar as we extend the notion of tui from disputation to the inner conversation of a thinking individual, we would lump together the two kinds of move – by Smith and Jones – as a single way of proceeding reasonably from one thought (or intention or policy, etc?) to another.


    One likely kind of nonkinship between “tui” and valid deductive inference is suggested by a comparison to what looks like a partner move, yuan 援. I think yuan 援 is probably about consistency or neutrality among people (though I raised a doubt about that in #9 above). Early in the Xiao Qu 小取 (Book 11, ch. 44), in a paragraph summarizing the main methods toward being right, we read this:

    What one has in oneself one must not reject in another; what one lacks in oneself one must not demand from another. (my translation)

    What does this saying mean? Maybe it is only the trivial logical principle that it is wrong to say “p, and you are mistaken in saying p.” But it can very easily be read more broadly: as an injunction to avoid covertly indexical language, or to be fair and polite in discussion, or to apply the golden rule when considering aggressive skepticism, or simply to apply Kant’s Formula of Universal Law to check all policies verbal and otherwise. The saying looks like good epistemic (and moral) advice even if it’s more than just logic. It reminds me of Hume’s idea that in order to have mutually intelligible language about the objects of our sentiments as such, we have to make our terms neutral as among bearers of those sentiments: hence natural language tends to have words that mean e.g. “such as to tend to be pleasing to people” rather than “pleasing to me.” Hume’s point is not about logical consistency. The saying also reminds me of Youzi’s couplet at Analects 1.13: 信近於義,言可復也 “Trustworthiness is close to being just: it enables one’s words to be repeated.”

    Maybe similarly the “tui” move reflects not so much a concern for logic as the sense that our thinking is more likely to be right insofar as we are thinking in terms that are neutral among exactly or roughly similar cases. Trying to think in general terms is good epistemic policy. That is, it’s not so much that we should draw the correct inferences from the principles we have, as that we should try to have principles.

    In this context I think we should understand David’s #11 as pointing out that toward the end of having clear general principles, or having something correct before we get there, having vague inchoate principles that we gradually sharpen may be more helpful than thinking only in terms of clear candidate principles. For the latter will too often be too easily disprovable. If we insist on clarity, we will too often simply erase the whole blackboard.

    And then the difference is obscure between (a) a proto-principle, and (b) a paradigm case or a metaphor or simile.

    What I’ve said here in C more or less fits Manyul’s 2002 characterization of yuan and tui: “In one form of analogizing, one ‘pulls’ (yuan2) into the argument an aspect or principle that one’s opponent accepts, and asks why it may not also be used by oneself. In the other form, one ‘pushes’ (tui1) the opponent’s judgment from the case he or she already accepts to another relevantly similar case.” But my guess is that when these moves are made it’s not necessarily given in advance what principles Smith has been working with. And one point Manyul’s description might not stress enough is that part of the point of yuan is to rule out having principles with one’s own name written in, such as “Smith may do/say …”


    As I’ve described tui, it is in some sense directional. It starts from particular things the person already accepts, and heats them up (under the magifying glass of attention?) to say something further.

    Boram, I wonder whether Mencian extension is always from the near to the far. That baby at the well is no relation to me. Is its lack of kinship-nearness outweighed by its literal nearness? If the progression “Mom, donkey, grass” is a progression from the near to the far, is that because “near” means similar to me? Similar to Mom? When I extend my courtesy, am I beginning with courtesy toward myself? And when I extend my shame at my big mistakes to shame at my small mistakes, is that a move from the near to the far? (2A6 doesn’t use ‘tui’, but it uses ‘kuo er chong擴而充, expand to completion’.)

    Mencius allows at least that people wrongly extend callousness from the far to the near (7B1), though he doesn’t use the word “tui” to say it. I sometimes extend my not mowing your lawn to not mowing my own.

    Boram, I take you to be saying this: “If Smith accepts p and rejects q, but p and q are similar, then Jones can ask her to accept q or ask her to reject p. Both moves are tui, or easily translatable into tui. That’s symmetry.”

    But in either case Jones is starting from something Smith has already adopted, and asking her to go farther.


    Maybe the kinds or degrees of generality that make for the best epistemic/moral policy depend on what capacities we have developed. In that case we should not only work toward principles, we may always have to work toward even more principly principles, and (a parallel point) work toward extending our moral circle ever farther, as Boram suggests at the end of #13. I think Manyul’s 1999 paper could be interpreted as arguing that Mencius doesn’t accept the latter point. But I’ll have to check.

    The point that Mencius’ purpose is to show Xuan that he has abilities does seem to be a point of disanalogy between the Mencius move and the disputational move of tui. That’s a point in favor of my speculation in #1-2.

  15. This discussion is a little out-of-date, since (with the exception of Agui’s citation of PJ’s work) it doesn’t take into account any of the scholarship on this topic since 1999. Instead of re-inventing the wheel, why not benefit from the work that has already been done, so that you are in a position to make a genuinely original contribution?

    Hutton, Eric. “Moral Connoisseurship in Mengzi.” Pp. 163-86 in Liu and Ivanhoe, Essays on the Moral Philosophy of Mengzi (Hackett, 2002).

    Wong, David. “Reasons and Analogical Reasoning in Mengzi.” Pp. 187-220 in Liu and Ivanhoe, Essays on the Moral Philosophy of Mengzi (Hackett, 2002).

    Van Norden, Bryan W. Section IV.B.2, pp. 234-46 in Virtue Ethics and Consequentialism in Early Chinese Philosophy (Cambridge, 2007).

    And I assume everyone commenting here has read the classic paper that started this whole debate?

    Nivison, David S. “Motivation and Moral Action in Mencius.” Pp. 91-119 in The Ways of Confucianism (Open Court, 1996).

  16. Bryan; thanks for returning with more references!

    Lots of things are out of date; much of Confucianism is out of date. But I do like to discuss and re-discuss things from time to time to see what else might come out of it than what has been already been pumped out, including what I might have tried to pump out in earlier work. It helps me at least to rethink what I’ve thought in the past and I find it useful to engage in critical discussion with people who are willing, especially if published work since my own hasn’t really addressed it (i.e. my work) in any detail. The detailed criticisms from people like Bill, Boram, and others really help me.

    But I take your point that there certainly has been quality work, and not merely mine or Chris Fraser’s, done on analogical reasoning since Nivison’s.

    Actually, the Nivison work you cite is somewhat different from the piece that I thought of as *starting* the whole debate, which is the earlier, 1979 piece: “Mencius and Motivation,” Journal of the American Academy of Religion, Thematic Issue, Vol.47, no.3. The 1996 piece seems to represent Nivison’s own reaction, in retrospect, to his earlier piece. Nivison 1996 is much clearer about at least this aspect:

    “…Mencius is led from the over-simple “you can do it” position of 1A7 to a more sophisticated conception that one will need to “cultivate” (yang3) oneself carefully. The most memorable discussion of this is in 2A2” – 1996, 108

    Of 2A2, Nivison says

    “…the question asked [in 2A2 by Gongsun Chou] was, ‘tell me about your, and Gao’s, “unmoved mind”.’ On its face, it is a question about methods and goals of self-cultivation, and Mencius’s later cultivation analogy confirms this. Gaos’s maxim therefore must be understood by Mencius as indicating in some way how Gao would have people cultivate themselves.” – 1996, 125

    I don’t have Nivison 1979 in front of me right now, but my recollection is that these views are not explicit in that seminal piece.

    I think, actually, that you might have some good insight into Nivison’s thinking–especially whether he modified his stance between the two papers–since you edited the 1996 volume, working closely with Nivison. It would be wonderful if you could shed some light based on those interactions.

    And really, it would have been nice for you to say something more about the Ivanhoe view as well, since Agui wanted to introduce it for the sake of discussion.

  17. I would also have to add (although coming from an earlier date), sections 2.6-3.2 of Cua’s Ethical Argumentation (primarily because it’s neglected in the Nivison et al. debate).

    Having read all the articles/books Bryan listed (some more recently and more closely than others), I would say that this discussion does have some originality in two aspects: 1) In asking the question of whether or not tui is different from more traditional notions of inference. 2) In contrasting the Mencian notion of tui with a Moist notion of tui. To my memory, none of the articles go into much depth on these issues.

    More importantly, I think Bryan’s comment raises the question of the purpose of this kind of community (i.e., this blog), in particular for the specialist. I imagine we would all agree that reading these articles are not prerequisites for participation. In fact I would go so far as to say that those who have read the pieces have an obligation to bring them into the discussion so we don’t reinvent the wheel, so to speak. At the same time I’m not sure how helpful it is, given our time restraints, and the nature of this forum (being somewhere between an informal conversation and a response/rebuttal in a journal–although much closer to the former), to list references as a form of participation without summarizing or evaluating them. It comes across as if one is making the claim that others would have to have read the material in order to make the conversation worthy of one’s participation (which I’m sure was not intended). That said, I think the question remains, how can blogging become a productive use of an academic’s time? (Assuming it can be productive at all.) Sorry to thread-jack. We can all ignore this comment if it’s irrelevant.

  18. Bill, thanks for your Comment 14, D, those are some excellent, incisive observations!

    I’ve claimed that Mencian extension is asymmetric, in the sense that it transfers our concerns from near to far, but I think you are pressing me on my use of “nearness”, that my use of it may very well be loose and metaphorical to the point of being otiose.

    You may be right, and I’m beginning to feel I might be mistaken. I need to mull over this more, and will get back to you.

    As for 7B1, that’s a great find! Here it is, courtesy of the Chinese Text Project (my comment on it follows after):

    Mencius said, ‘The opposite indeed of benevolent was the king Hui of Liang! The benevolent, beginning with what they care for, proceed to what they do not care for. Those who are the opposite of benevolent, beginning with what they do not care for, proceed to what they care for.’

    Gong Sun Chou said, ‘What do you mean?’

    Mencius answered, ‘The king Hui of Liang, for the matter of territory, tore and destroyed his people, leading them to battle. Sustaining a great defeat, he would engage again, and afraid lest they should not be able to secure the victory, urged his son whom he loved till he sacrificed him with them. This is what I call “beginning with what they do not care for, and proceeding to what they care for.”‘

    My comment on the passage:

    To me it seems there are two modes of reasoning involved here. One is the Mencian extension that begins from what one cares for, and proceeds to what one does not care for. That’s the direction in which benevolence grows.

    The other mode is profit-driven instrumental reasoning, which Mencius always puts in a negative light, and even views as self-defeating. I read 7B1 as claiming that profit-driven instrumental reasoning, when acted upon, will even work in the opposite direction as that of benevolence.

    So, I do not read 7B1 as suggesting that Mencian extension can work its way from what one does not care for to what one does care for. A different mode of reasoning is to be blamed for that.

  19. Thanks Boram, that’s very interesting. I’ll think about it.

    (The next comment that will appear from me addresses the near/far stuff slightly, and was written before this one, after Agui’s #17. It has a link in it, which means it waits for Manyul to certify that it’s not spam.)

  20. Thanks Agui. I wouldn’t have known to look to that Cua on this topic.
    Here’s a more general reading list from Bryan:

    This blog is immensely helpful to me in many ways that reading print publications can’t be. Print publications don’t speak up to correct my false assumptions, e.g. about the ancient use of ‘hao’. They don’t answer questions about what they meant (as Plato relevantly pointed out), or listen when I try to work something out (cf. Lu Xun’s preface to Na Han), or feed my natural need to exhange help or to fight. They don’t throw random topics before me without my initiative. Most importantly, they don’t help me meet people. (Find me at HKU Philosophy.) The blog would be more valuable if we had more and diverser participants, and I hope Bryan’s interest is a harbinger of that.

    Agui, I appreciated the brief reminder of Ivanhoe, and I find his point very plausible and very interesting, especially because it notices Mencius’ concern with distinguishing one feeling (or whatever) from another. I’m inclined to think the four duan or four hearts are not supposed to be feelings exactly (I forget the details of the literature’s debate on that, but I’ll be going back to it this month). That seems relevant to 1A7. But maybe that’s like saying “they’re not notes,” when Ivanhoe never said they were. Anyway I don’t think their not being feelings would ruin Ivanhoe’s point about learning to distinguish.

    One important question Ivanhoe’s paragraph recalls for me is the question whether the four whatevers (and other things like that in the Mencius) are cognitive in some way beyond showing us the kinds of flavors other people can taste. That’s related to this question for Mencius: What’s good about being good? Is virtue good just in the sense that it’s nice for the bearer?

    Maybe learning to distinguish and recognize the feelings/motivations/whatever behind our actions is like learning to recognize/have general maxims behind our actions, which is like learning to recognize/have general concepts and principles in our thinking.

    Boram, my questions about near/far were real, not rhetorical. Another relevant passage is 7B31, where Mencius could (but needn’t) be interpreted to say that pretty much any starting-point is OK. I’ve tended to assume that that’s one of the subtexts of 1A7 and 1B1 (on liking pop music).

  21. Agui, quite right on all counts. Blogs are a relatively new phenomenon and it does make sense to ask wherein their value lies. I think “somewhere between an informal conversation and a response/rebuttal in a journal” is accurate. A blog can be useful for anyone who is interested in casual perusal, but let me say something about why this forum–this particular blog that I run–can also be very useful for academic types. It’s good to have a “virtual place” that functions as a kind of intellectual salon or, by way of more contemporary analogy, a philosophy department lounge or coffee room, for less than formal exchange of ideas. And I take the “exchange of ideas” part quite seriously. It’s good to know, of course, whether there is a published piece that addresses the concerns pretty directly but we can’t really exchange ideas with a publication. Another person, engaged sincerely (not to sound too Confucian) in a dynamic conversation can be invaluable to developing an opinion or for testing an idea. I think that is sometimes best done live, but sometimes it helps to have the more “textual” mode of something like a discussion board or blog.

    I feel like all of this was already implicit in what was going on here. But maybe it’s good to put it out into the open.

    One more point on the helpfulness of existing publications: Sometimes it’s good to embark on what may seem like “re-inventing the wheel” just because there is nothing likely to be so definitive as to be considered “the wheel” here. What one makes of Mencius on tui, in the (cold and dark) vacuum of not having read every piece that touches on it, might for that very reason turn out different in interesting respects from what each those published pieces has to say about it. There should be place for that in a reasonable model of intellectual inquiry and I think a blog, similar to a serendipitous conversation with a colleague, can provide the forum for such an endeavor, among other things.

  22. Above I write, “Boram, my questions about near/far were real, not rhetorical.” That’s not a complaint. It’s just trying remove an earlier false appearance of complaint. … sigh.

  23. Manyul, you ask for feedback now, but when you showed me a draft of your article years ago, I gave you extensive written comments, including citations of passages from the Mengzi that I thought your work had trouble accounting for. I wasn’t expecting agreement, but I was honestly looking forward to seeing how you would carry the dialogue further by saying something about the passages that I saw as problematic for you. When you published the piece, you simply ignored everything I had said.

    More recently, you attended a panel I was on at the APA, asked a question, I answered it, and later David Wong spoke up in support and amplification of my answer. Then you raised the same question in your blog, without any indication that you had heard or given any thought to the replies you got. When I posted a comment asking for what you thought of the answers given at the panel, your reply showed that you did not remember what was said.

    The reason all this bothers me so much is that people often pay lip-service to notions like “open-mindedness” and “dialogue.” But it is actually extremely challenging to live up to these ideals. You need to make the effort to find out and understand what others are saying. I don’t find that you do that. Frankly, it frustrates me a lot.

    As for the difference between Nivison (1980) and Nivison (1996), I explain their relationship in the editorial note to the latter in Nivison, The Ways of Confucianism, pp. 293-294n1. Basically, Nivison wrote a paper for a public presentation in 1975, modified it in 1976, and wrote a supplement to it, before publishing a much-edited version in 1980. The 1996 version is simply the material from all three versions spliced into one paper. Nivison checked it to make sure I had combined the papers correctly. If anyone wants to see the enedited papers, send me your surface mail address and I’ll send you copies.

  24. The present comment accepts the terms ‘near’ and ‘far’ uncritically until maybe the last line.

    One kind of question about Mencius is whether he distinguishes between natural-or-normal improvement and other kinds, holding that the former proceeds from the near to the far and the latter need not. If he does so distinguish, then we can ask: Are the two kinds of improvement otherwise analogous? (And would he call them both “tui”?)

    Here’s a separate set of questions. For one or another kind K of Mencian improvement, we can ask various things about its relation to a debating move or form of inference T:
    a. Is K analogous to T?
    b. Is K analogous to T as applied to a certain material (specifically, applied to the near and carrying us to the far)?
    c. Is K disanalogous to T for reasons bound up with near/far directionality?

    I think (c) is a very interesting question.

    Boram, you write in #6: “Instrumental reasoning is concerned only with discovering means to given ends, but Mencian extension expands the range of our concerns (ends) via a deliberative route.”

    That’s consistent with Yes to (b), I think. But your comments in #13 and #18 make me wonder whether your thought is also that instrumental reasoning is the transfer of concern from the far to the near because it starts from relatively distant ends and traces a path from them back to what we might immediately do.

  25. In speaking of “improvement” just above, I was confusing changes in character with deliberating toward a particular action. Maybe Mencian “tui” is mainly the latter.

  26. Bryan, speaking of interrogating a text: I’d like to ask you about something in your new book, regarding the lines in 2A2 that introduce the story of the man from Song. The received text has been challenged. We might accept the phrase zheng xin 正心 from the received text as e.g. Legge and you do, or amend them to wang 忘 as e.g. Lau does.

    After saying 告子未嘗知義,以其外之也 “Gaozi has never understood rightness, because he makes it something external” (Legge), Mencius says

    In the received text:

    必有事焉而勿正心, 勿忘, 勿助長也。

    “There must be the constant practice of this righteousness, but without the object of thereby nourishing the passion-nature. Let not the mind forget its work, but let there be no assisting the growth of that nature.” (Legge)

    “[One] must work at it, but do not aim at it directly. Let the heart not forget, but do not help it grow.” (Van Norden, p. 245)

    Changing 正心 to 忘 yields:

    必有事焉而勿忘. 勿忘, 勿助長也。

    “You must work at it and never let it out of your mind. At the same time, while you must never let it out of your mind, you must not forcibly help it grow either.” (Lau)

    This seems to me a very important matter, and if you’ve already addressed my questions in print maybe you can just tell us where to go for that.

    1. There’s a lot to be said for your reading. But I can’t convince myself that it’s definitely right. Are there are other occurrences of zheng 正 or zheng xin 正心 in early texts that support your reading here? (I’ve done a quick search and my impression is that zheng xin 正心 seems to be a term with a strong positive connotation, referring in a broad sense to having one’s heart-mind in true or proper order. There’s one case in the Lienü Zhuan — 所以正心壹意,自斂制也 — where the phrase is used to mean something like “be well and properly focused on”, which is pretty close; but the phrase seems to take an object that comes after, which we don’t find in the Mencius passage; and it is followed by another phrase which could also mean “be fully focused,” so it’s hard to know what the instance demonstrates.)

    2. Is it your view that what we are not supposed to aim at directly is yi 義?

  27. More to Agui’s #17 (or thereabouts): There’s something really problematic about academic blogs, though. Nobody has a moral responsibility to slog through our blog. And yet …

    When the academic conversation takes place in a limited number of books and refereed journals (that publish not long after submission), the overall conversation can be clearer and more definitive in various ways than it can when more scattered. There’s more of a divide between public and private. A scholar can more easily preserve a kind of sense of responsibility to the whole, including a sense of responsibility to give credit where it is due. In short, academic blogs might be a moral burden on the community, or at least felt as such.

  28. Bill, on Bryan’s behalf maybe, what do you make of the use of “正” in Xunzi’s “正利而為謂之事,正義而為謂之行”? (It’s in the list of explications towards the beginning of “Correct Names.”)

    (There’s other stuff in the thread I want to take up, but a couple of things need to get done first.)

  29. Bill, to your Comment #28,

    No, I don’t think instrumental reasoning necessarily involves a transfer of concern from far to near. Mencius seems to believe, or at least seems to claim, that at least for some external good X (music, wealth, and women [apologize for the sexism there]), sharing one’s fondness for X with the people is the best means of getting and enjoying X. Mencius seems to offer these to King Xuan of Qi as pieces of instrumental reasoning. (Even though Mencius himself doesn’t condone profit-driven instrumental mode of reasoning, he may be offering these pieces of instrumental reasoning as–to use a Buddhist term here–upaya, or a ladder to kick away once one has climbed on top of it to a different sympathetic mode of reasoning that would enable the king to see that sharing his joys with the people is an intrinsic good, not just instrumental to his getting what he wants.)

    In general, instrumental reasoning transfers concern or motive force from an end to the choice of action most conducive to that end.

    Here’s how I see Mencian extension. I define it narrowly (but quite imprecisely for now) as pushing one’s concerns outward from the near and dear, as far as it can go, by using a deliberative process. Here the items one deliberates with are not judgments about what one ought to feel, but the feelings themselves. I take these feelings as emotions that have intentional content (emotions are directed towards, or are about, this or that object), and deliberation operates on the intentional contents of emotions. So I dissent from Hume in Treatise 2.3.3 who could be read as arguing that there’s no such thing as practical reasoning involving the passions, because passions have no representational content. To cite a piece of literature that’s in keeping with my interpretation, there’s Wong’s “Is There a Distinction between Reason and Emotion in Mencius?” (unfortunately I re-invented this wheel while piecing together what I read of Mencius, Hume and a bit of recent literature on emotions, before reading David Wong last year).

    My current view is that the role of Mencian self-cultivation should be kept distinct from Mencian extension. The deliberative process itself, involving the transfer of concern, is immediate, but is still subject to a type of weakness of will that Bryan points out in his introduction to Nivison’s WAYS OF CONFUCIANISM (i.e., acedia, as opposed to akrasia). The goal of self-cultivation, as I read Mencius 2A2, is to overcome acedia through accumulation of right conduct. Bill, let me know if this has any connection to what you’ve been exploring in your Comment #28.

    p.s., where does the concept of acedia come from? If I had read all the literature that Bryan cites, I would probably know, but for now all I can remember is that it has some connection with the Catholic tradition (?), and the closest word my Greek lexicon (“the middle Liddell”) lists is akEdia, meaning “indifference”.

  30. Boram, that’s interesting that you want to keep self-cultivation distinct from Mencian extension. I think I would agree, though maybe “agree” is too strong since we seem to disagree on what Mencian extension is. However, I’m confused by this: “The deliberative process itself, involving the transfer of concern, is immediate, but is still subject to a type of weakness of will that Bryan points out in his introduction to Nivison’s WAYS OF CONFUCIANISM (i.e., acedia, as opposed to akrasia).” If the deliberative process actually *involves* a transfer of concern, then it can’t suffer from acedia, if I understand the latter correctly. Nivison’s gloss of it is, “Sometimes it seems that I judge I should do something yet just do not, or perhaps cannot, care enough about it to act” (1996, 92). I suppose if “concern” does not carry with it “caring enough,” then what you say would make sense. But that doesn’t seem to follow ordinary or philosophical usage for “concern.”

    The answer to where acedia comes from probably lies in what Bryan writes in footnote 3 on Nivison 1996, 283: “For philosophical discussions of acedia, also called “spiritual apathy,” see Thomas Aquinas, Summa Theologiae II-II.35…” He also cites two Aquinas scholars, Heath and Fairlie, for further amplification. So, I guess it is Aquinas who is likely to be the locus classicus, or whoever influenced him.

    My reaction to the phenomenon of acedia when I first read Nivison (1979) on it and when I re-read it now, is that it seems not very different from what Darwall (Impartial Reason) would likely call the denial of “judgment internalism.” Judgment internalism is the view that it is a necessary condition on a judgment’s genuinely being of a certain type—say, the judgment that one ought to φ or the judgment that one has a reason to φ—that the individual who makes the judgment be disposed to act in accordance with it. That makes it sound to me like acedia, understood as a denial of judgment internalism, would be more like a philosophical view, not neutrally describable as a “problem” or a “weakness” of some practical reasoning sort.

  31. Dan, thanks! That’s certainly to the point and promising (and impressive). Worries are: If we take zheng 正 by itself to mean “aim directly at” in the context of the Mencius line, then what we’re not supposed to aim directly at is xin 心, which could be right but sounds odd. I’d want to do more digging into zheng 正 (mainly as not followed by 而為), or better yet hear that the digging’s been done; but if I proceed indirectly toward that digging by further developing my facility with the language first, the digging will be quicker and I may not even need to dig.

    Manyul, I think someone who is acedic about X wouldn’t deliberate about X. I see how judgment internalism might have to be false in order for there to be acedia, but I don’t see why the acedic has to have a view on such things.

    Boram, I didn’t think you were thinking that instrumental and extensional thinking are exactly opposite, but I thought you might have in mind some remote tension between the two that sprang from their heading in some faint sense in opposite directions. I think that might be right …

    I agree with you about the prima (and maybe secunda) facie appearance about what Mencius is doing (I’ll look into upaya, thanks), but I’m hoping there’s a nicer face to be put on it. Also I agree that passions are normally representational. (I think they represent the way pictures do, but that’s a long story.) I think that fully having moral view X is a matter both of feelings and words. But sorta having moral view X can be a matter of only feelings or only words. (I’m simplifying a little!)

    I also agree with you that Mencian extension isn’t instrumental reasoning. I like very much and am inclined to accept your idea that Mencius’ opposition to aiming at li 利 is specifically an opposition to casting one’s big deliberations in a basically instrumental form, not a list of oppositions (to, say, egoism and utiltiarianism).

    I associate your point with my idea (not about Mencius) that even though anything is good insofar it means net pleasure for the universe, there are no fundamental ends (in the sense of well-articulated goals, things to be deliberately brought about). Here’s how I argued for that denial in the March 21 family thread, #46: “I think the idea of an “ultimate end” is a false idea. The ends one articulately grasps in one’s life tend to be for the sake of farther goods one less articulately conceives and less determinately aims at. The more “ultimate,” the less like a “goal.” Playing chess is for the sake of more nebulous things like harmless pleasure, company, maybe mental exercise, maybe … I don’t know what else. Plato and Aristotle held that we all aim ultimately at the Good or at Eudaimonia, and that virtually everyone lacks even a remotely adequate concept of what that is.”

    I don’t see (yet) a sharp distinction in Mencius or in reality between deliberation toward particular actions and thinking toward new concerns that will stay with me.

    Boram, for at least an indirect connection with my question about Bryan’s reading in #28, we might start with this passage from p. 304f of Bryan’s book (my quoting Bryan does not signal agreement with his sense of the ground-rules for blogs):

    “… Mencius’ argument in 1A1 and 6B4 is that consequentialism (of the Mohist or any other variety) is self-effacing. [footnote on Chad’s version of this point] This is similar to one major line of critique of contemporary forms of consequentialism. One way on which consequentialists have replied is to develop *indirect* versions of consequentialism, according to which one develops, for consequentialist reasons, motivations and dispositions to follow rules that are not consciously guided by consequentialist considerations, but that will, in general, produce the correct consequences. … This is slightly different from … ‘rule consequentialism’ .… An ‘indirect consequentialist’ is an act consequentialist who holds that the right action is always the action that maximizes good consequences, but she also acknowledges that consciously aiming at good consequences may not always be the best course of action.”

    (Thus an indirect consequentialist is going to end up holding that it can be wrong to aim at the right action.)

    Bryan’s reading/translation of the Mencius line (“do not aim at it directly”) suggests that Mencius’ advice there is itself analogous to indirect consequentialism, where the consequence indirectly aimed at is the full development of one’s xin or qi or yi or something.

    I like to think Mencius thinks even one’s own xin is ultimately, indirectly, for the sake of something further. But if it isn’t, then maybe it’s easier to make sense of passages like 7A35, which we’ve discussed, in which Hypothetical Shun casts off the empire to live by the sea with Dad.

    (I agree with Dan that Mencius’ discussions of Shun aren’t the best evidence of Mencius’ views. I think Bryan disagrees.)

  32. I should have said: “I like to think Mencius thinks even one’s own VIRTUE is for the sake of something further.”

  33. Hi Bill, and everyone else: the right parenthesis occasionally codes for “winking face” icon. I don’t know why! Does anyone here speak mark-up well enough to know why?

    Bill said: “Manyul, I think someone who is acedic about X wouldn’t deliberate about X. I see how judgment internalism might have to be false in order for there to be acedia, but I don’t see why the acedic has to have a view on such things.”

    I should be clearer; what I meant was that Nivison’s discussion of the phenomenon he calls “acedia” seems to regard that phenomenon as prima facie irrational or at least some sort of practical pathology. I just meant that regarding the phenomenon in such a way implies *Nivison’s* acceptance of judgment internalism. If someone did not accept judgment internalism, she might not find the phenomenon problematic and in need of special explanation or therapy. So in that sense, assuming that there is something irrational or pathological about judging something to be what one ought to do but not being moved to do it, seems too quick.

  34. Oh – then I agree with you, Manyul. But then anyone concerned with promoting virtue (in herself or others) might regard acedia as a problem to be attacked, rather as poverty might be a problem to be attacked, even if it isn’t irrational or pathological. My recollection – and I don’t have time to check – is that this rather than irrationality or pathology is Nivison’s main concern.

  35. Since I’m commenting from a hotel and a bit tired, I will be quick. To Manyul’s Comment 34, actually (since 4 years to be precise) I’ve believed that acedia is compatible with judgmental internalism. Let’s say that a reason for action can motivate, but cannot motivate you all the way to action, even if nothing intervenes. I don’t see why that would be impossible at all. I see and can appreciate the reason why I should participate in the Boston Marathon to help with cancer, but it doesn’t motivate me enough to get me off my armchair.

    Unless judgmental internalism can allow for the possibility of weakness of will, I can’t see how the position can remain true to fact.

    Thanks for the other helpful info, and also to Bill; I will respond more fully once I return.

  36. To briefly explain further: The more distant the object of concern from one’s usual targets of concern, the motive force conveyed to it by the deliberative process will be the weaker. Hence acedia. (I’m borrowing a lot of this from Hume’s Treatise, I think.)

  37. Boram, that’s helpful to me (I’m not Manyul but I’m the one who’s awake now). I think I’d been misconceiving the ak/ac distinction. Anyway how about this. Here’s Nivison’s distinction:

    AK: my judgment is thwarted by contrary motives
    AC: my judgment isn’t helped by enough motives

    If one accepts a version of judgement internalism such that the motives on the right are also judgments, then these states are not only motivational self-conflicts, they’re judgmental self-contradictions, hence irrational in a strong sense.

    But one might accept, say, a Platonic view on which a motivation counts as a judgment only if it’s in one of the higher parts of the soul. And then the motives on the right don’t have to be judgments. That looks like one version of the view you set out in #42.

  38. Hmmm. Interesting, Boram. I’ll have to think about this, but it seems like you’ve reduced the problem of acedia now to something that is nearly indistinguishable from akrasia.

  39. Maybe learning to distinguish and recognize the feelings/motivations/whatever behind our actions is like learning to recognize/have general maxims behind our actions, which is like learning to recognize/have general concepts and principles in our thinking.

    “Impulses” perhaps?

    The general interpretation (in the Mencian context) seems to be that underlying 推 is not rules, maxims, or principles. Tui instead is skill-driven, and not meant to lead to or discover abstract concepts. Since we’ve been talking about Bryan’s book, let me quote from a few relevant portions:

    “So Mengzian extension has, at least conceptually, two aspects, which we might label ‘cognitive extension’ and ‘affective extension’. Cognitive extension is coming to see the ethically relevant similarities (or differences) between two actions, individuals, or situations. Affective extension is coming to have the motivations and emotions that are appropriate in response to two actions, individuals, or situations” (236).

    “I think we must acknowledge that, for Mengzi, cognitive extension can guide affective extension” (237).

    “In summary, cognitive extension is not a process of coming to generalize to an inviolable rule, which we then apply to further cases. Rather, it is developing a skill at thinking about the similarities and differences between ethical cases” (241).

    It’s not clear to me that Mencius makes such a distinction between ‘extensions’, but it is an attempt to make sense of the differing ways it’s used.

    I’ll look into upaya, thanks

    Pye, Michael. Skilful Means: A Concept in Mahayana Buddhism. London: Duckworth; Dallas: Southwest Book Services, 1978.

    Schroeder, John W. Skillful Means: The Heart of Buddhist Compassion. Honolulu: University of Hawaii Press, 2001.

  40. Thanks Agui! Some of Mencius’ rare references to filling out categories (lei 類) suggest that he’s not all the way particularist (3B3, 10), but I agree too that he’s not at the opposite extreme. Noncommittal of me, eh?

    Aside from that point, I think there may not be much conflict between what you’re proposing and my extremely vague speculations. I didn’t mean to suggest that Mencius thinks of what’s extended as actually being a rule, maxim, or principle. Indeed he doesn’t seem to think what’s extended need be something verbally articulated at all. I think that leaves open the possibility that it is “like” a maxim in some sense relevant to analogies with disputational moves.

    I hope to have more interesting things to say about Mencius in a few weeks. Right now I just don’t! Meanwhile, your comment helpfully corrects some of my drift, and I appreciate the references.

  41. Above I gave a putative example of how an argument by analogy, or disputational move by analogy, might end up rejecting its own starting-point. But I didn’t say anything about how it can make sense for an argument or d-move to reject its own starting point. Here’s a brief account.

    Let’s say argument by analogy works this way:

    The starting point is that I accept X. You point out that X and Y are analogous, so I should accept Y. That move makes sense if we take the claim that X and Y are relevantly analogous as an abbreviation of the claim that there is some G such that what leads me to accept X is my articulate or inarticulate allegiance to G, which is more general, and G supports Y in the same way that it supports X. There may also be a gesture in the direction of spelling out G.

    Since in a sense I have already granted G, and G supports Y in the same way it supports X, I should accept Y, other things being equal.

    Let’s say that’s argument by analogy.

    One way for other things not to be equal is that Y and X may conflict. That means my own views at the end of the above process are conflicted. To simplify, let’s say we’ve uncovered a conflict between G and X. How do I deal with that conflict?

    It doesn’t look especially rational for me to reject or adjust G to keep X, since my acceptance of X is based on G. There is no parallel objection to my rejecting or adjusting X to keep G. So it would seem to make more sense to reject or adjust X to accommodate G than vice versa.

    Other things being equal.

  42. How about this?

    G = killing of the innocent should be minimized
    X = killing of this innocent ox should not be done

    These conflict because part of keeping social violence to a minimum is keeping up with ritual.

  43. More generally, my thought is that the same G may support many X, Y, Z, etc., other things equal; but that these latter may not be all compossible, so that X or some of X may be outweighed and have to be sacrificed in service to G.

  44. To Manyul’s #44, which was to Boram’s #40 and #41:

    For clarity I’ll use the terms ‘acedia’ and ‘akrasia’ only for conditions of a person, not also as names for problems or theories about those conditions.

    Here are Nivison’s definitions (1996, 91f):

    AK: “I judge I should do something, yet cannot or do not control the temptations which move me not to.”
    AC: “I judge I should do something yet just do not, or perhaps cannot, care enough about it to act.”

    But there’s an aspect of the distinction that isn’t reflected in these words, and that commonly isn’t reflected in definitions of akrasia in the literature. It’s that on AK, the judgment does motivate me some.

    Hence the following is a clearer presentation of Nivison’s distinction, making it more clearly not a distinction between acedia and akrasia as the latter is traditionally defined:

    BK: I have a moral judgment that motivates, but contrary motives win.
    BC: I have a moral judgment that just doesn’t motivate (or not much).

    For I think one of Nivison’s purposes in distinguishing akrasia from acedia and drawing our attention the latter is that he wants to draw our attention to early Chinese concern about the problem of having any moral motivation at all (cf. 1996, pp. 80, 96). This purpose would be served by distinguishing BK from BC, but is not served by distinguishing AK from AC.

    The standard definition of ‘akrasia’ in the philosophical literature is AK rather than BK. But the standard definition may reflect a little bit of false consciousness about the term. Consider WW:

    WW: I have an intention (a motivating concern I identify with), but contrary motives win (these might include simple lassitude).

    WW is “weakness of will” at least as understood outside the philosophical literature, and sometimes even akrasia as more or less explicitly understood inside the philosophical literature. See the initial pages of:

    WW is close kin to BK. Nivison uses the terms ‘weakness of will’ and ‘akrasia’ synonymously, so far as I can tell and so far as Bryan’s index suggests.

    I think Boram and I were operating with a conception of the distinction that follows Nivison’s words in not seeing the distinction as having anything to do with the judgment side of it, the left-hand side of the accounts, hence not having to do with whether or not the judgment involves a defeasible general motivation; but rather as having to do with the right-hand side, hence the particular motivations attached to particular actions: applications of the rule accepted by the judgment, or means to the end approved by the judgment. Akrasia is the presence of little enemy motives, acedia is the absence of little friend motives.

    On a kind of strong internalism that regards apathy about X as a judgment of X’s unimportance, acedia then becomes self-contradictory. That’s what I was thinking of in #42.

  45. Bill, this is very helpful: “Akrasia is the presence of little enemy motives, acedia is the absence of little friend motives.” But I’m not sure “little friend motives” captures everything. Shouldn’t it be more like “… absence of little friend motives *or* of sufficient strength of (the principal) motive” (inclusive “or” )? I take it one could use friendly motives to form a coalition of the willing; wouldn’t another strategy be to try to strengthen the motive itself? I’m not sure how the latter could be achieved, but it seems like Nivison and others talk that way, in interpreting Mencius, more than in terms of friendly other motives.

    Separately: I suppose denial of judgment internalism results in the absence of any motive to φ whatsoever that necessarily accompanies the judgment that I should φ, and hence is too strong to be (or, alternatively, is an extreme case of) acedia. I think I understand acedia and Nivison much better now.

  46. Actually, it occurs to me that akrasia and acedia probably occur together, in the same agent, given that the common cause seems to be the insufficient strength of the motive attached to the judgment that one ought to φ.

  47. Manyul,

    Your last point seems right to me – the two conditions would commonly coincide. But full acedia would never coincide with typical akrasia (I mean the kinds of case of akrasia people typically discuss), insofar as the latter involves motivational conflict and the former doesn’t.

    Separately: I guess there are strong and weak kinds of judgment internalism. Strong internalism says the motivation necessarily attached to all judgment is adequate for action; weak internalism says motivation is necessarily attached to all judgment but might sometimes not be adequate for action. The denial of strong internalism is consistent with the acceptance of weak internalism.

    I agree that a good rectifier of names would say that the falsehood of strong or weak internalism does not amount to everyone’s having mild or extreme acedia, respectively. I think the point of the term ‘acedia’ is to describe a non-universal condition. If judgment internalism is false, still there will be normal correlations between judgments and motivations that mimic what a judgment internalist thinks is necessary in all cases.

    And yet … the philosophical problematic about akrasia is supposed to include the possibility of arguing that since judgment internalism is true, nobody really has akrasia. To preserve that problem we must not make it true by definition that akrasia is sometimes had. But maybe a good name-rectifier would not preserve that problem about akrasia.

    A good name-rectifier should preserve the possibility that nobody has acedia, because that might be at issue among internalists; but need not preserve the possibility that everyone has it, because the latter idea would not serve discussion by being distinct in any interesting way from the falsehood of internalism.

    In #51 I was arguing primarily that Nivison’s words can suggest a diagram that is slightly misleading about his distinction, and at the end I was developing that diagram toward the exegesis of earlier comments by Boram and me. The little-friend account was part of the latter exegetical project. But I think it’s a possible way of conceiving things. You’re right to point out that it suggests and allows a strange sort of power-steering view of normal action, but I didn’t mean it to imply that view. In the rest of this comment I’ll set out the diagram more clearly.

    AKRASIA: Despite the fairly standard definition of akrasia as involving a conflict between judgment and motivation, usually when people are talking about akrasia they’re talking about a conflict of motivation, reflecting a distinction between two things that motivation might attach to or push toward. To simplify, let’s call these two things end E and means M. (E might be “bringing about X”, and M might be a means to that; or E might be “always following principle P”, and M an instance of that; or E might be a complex course of action, and M a piece of it; etc.) The motivational conflict is that I like E but I dislike M. Accompanying the puzzle about how akrasia is compatible with judgment internalism is a second philosophical puzzle about how it can make sense to say E moves me if M doesn’t therefore move me. But the phenomena do make us want to say something like that. To leave room for both puzzles, and to leave conceptual room for judgment internalism and its denial, we might have the following two-part picture of akrasia.

    (a) There is a judgment “E is right”
    and perhaps an associated motivation toward E
    (this account is noncommittal on
    judgment internalism).

    (b) There is motivation against M.
    One does not do M.

    ACEDIA. Now suppose one wants an elegant distinction or parallelism between akrasia and acedia. One can describe acedia in a parallel way, as follows:

    (a) There is a judgment “E is right”
    and perhaps an associated motivation toward E
    (this account is noncommittal on
    judgment internalism).

    (b) There is little or no motivation toward M.
    One does not do M.

    In these accounts of akrasia and acedia, (b) speaks of the presence of little enemies or the absence of little friends, but is noncommittal as to whether the little friends in normal (non-akratic, non-acedic) action are direct reflections or outgrowths of the motivation toward E mentioned in (a).

    As you suggest, discussions of acedia have less reason to stress a distinction between E and M than do discussions of akrasia. In that sense this account of acedia may be misleadingly complex.

    How’s that?

  48. In (a) above, which is the same for akrasia and acedia, the second line says “and perhaps an associated motivation toward E”. That possible association might be because judgment internalism is true, or it might just be because a judgment that E is right is normally or commonly associated with a motivation toward E. I meant to be noncommittal about that. Sorry.

  49. I like this way of describing the conditions:

    Paradigmatic acedia is “I judge X right but I don’t care”. Peripherally, acedia can include “I judge X right but I don’t care enough to do it.” At least core acedia doesn’t involve motivational conflict.

    The paradigmatic models of akrasia are two: “I judge X right but I don’t do X,” and “I intend X but because of some other motivation I don’t do X.” In the literature, I gather, definitions lean toward the first and examples lean toward the second. Etymologically ‘akrasia’ and ‘weakness of will’ both suggest the second rather than the first. So a definition fully faithful to the use and point of the word will be hard to come by.

    I’m not prepared to offer references to support my impression about examples in the literature.

  50. Thanks for continuing the discussion in my absence, and I think Bill has clarified the distinction between akrasia and acedia very well, esp. in Comments 54 & 55 (especially in noting that we should define these in ways that are non-committal on judgment internalism).

    When I last wrote on this stuff 4 years ago, my favorite example of acedia was my judgment that it was wrong to kill animals for meat: I cared that animals not suffer, but didn’t care enough about it to actually stop eating meat. And as Manyul points out in Comment #53, this condition occurred together with akrasia: sometimes the temptation of succulent steaks is too overpowering, and one gives in to temptation against one’s better judgment.

    I too like the way Bill puts it: acedia is the absence of “little friend motives”. Here’s how I see it, as a judgment internalist who admits the possibility of acedia. When I acknowledge something to be right, it is impossible for me to have absolutely no motivation to do it. But it is possible for me not to be motivated enough to do it.

    The cure to acedia (or maybe just one of the possible cures), then, is to find little friend motives that jointly will move you all the way to doing what you believe to be right. That’s in fact how I coped with my judgment that I ought not to eat meat. This judgment came with motivation that was not strong enough to move me all the way to action, but was strong enough to enable me to look for unrelated motives that I can recruit and whose motivatonal force I can channel into doing the right thing. E.g., I tell myself that philosophers ought to be better, finer humans than celebrities, and if celebrities can abstain from eating meat because they believe it’s wrong, then shame on me as a philosopher if I can’t. I incorporate arguments against eating meat into my Phil 104 syllabus, and then the sense of obligation I have as a teacher, and my dislike of being labeled a hypocrite by my students, may be channeled into the cause of a vegetarian lifestyle. Then I tell as many of my friends, family members, and acquaintances that I’ve become vegetarian, and this too helps me to abstain from meat, again from the dislike of being known as a hypocrite.

    All of this is compatible with judgment externalism. Mill was a judgment externalist, who believed that acknowledging the right thing does not automatically get you to do it, and that there was need of internal and external sanctions. But it is also compatible with what Bill has labeled the weak version of judgment internalism (in the 2nd para. of Comment #54), and that is the position I endorse.

    However, judgment internalism faces an additional problem, namely the onus of explaining how one can come to appreciate a weakly motivating judgment as a better reason for action than other overpowering motives (in cases of akrasia). I will address that in my next comment.

  51. Let me argue here that Mencius was a judgment internalist; I believe he was also a reason internalist.

    Let me explain what I mean by these terms. By judgment internalism I mean the position that Frankena describes as follows, that “motivation is somehow… ‘built into’ judgments of moral obligation” (Frankena, “Obligation and Motivation in Recent Moral Philosophy” in Melden (ed.), ESSAYS IN MORAL PHILOSOPHY, 1958). Generalize Frankena’s characterization from judgments of moral obligation to judgments of other kinds of obligation–and you have judgment internalism as I intend it.

    By reason internalism I mean the position that Williams argues for in “Internal and External Reasons” (in his MORAL LUCK, 1981), that reasons for action must be connected to one’s subjective motivational set in the relevant way. In my opinion Williams himself specifies “being connected… in the relevant way” a bit too narrowly, in a way that assumes instrumentalism (the position that all practical reasoning is instrumental reasoning). But I won’t dwell on that all that much here. I’m concerned with explaining why Mencius is a judgment internalist.

    My impression is that ancient Chinese philosophers for the most part implicitly adhere to the dictum ‘ought’ implies ‘can’ (conversely, that ‘cannot’ implies ‘not-ought’, though different philosophers (even within the Confucian school) may adhere to it in different ways. So in what sense does Mencius adhere to it, and how would he use the dictum? Certainly not in the sense that we must compromise what we acknowledge to be moral demands, no matter how taxing it may be on our abilities. Here’s what Mencius says (7A41 Lau trans.):

    Gongsun Chou said, “The Way is indeed lofty and beautiful, but to attempt it is like trying to climb up to Heaven which seems beyond one’s reach. Why not substitute for it something which men have some hopes of attaining so as to encourage them constantly to make the effort?”

    “A great craftsman,” said Mencius, “does not put aside the plumb-line for the benefit of the clumsy carpenter. Yi did not compromise on his standards of drawing the bow for the sake of the clumsy archer.”

    The way in which the dictum ‘ought’ implies ‘can’ constrains Mencius’s recommendation is as follows: our judgment of obligation implies our emotionally rooted appreciation of it, and motivation to act according to it. So one’s ought-judgment implies one’s being moved to act according to that judgment, and this seems to be implicit in Mencius’s conversation with King Xuan of Qi in 1A7 (especially the lifting Mount Tai vs. lifting a feather analogy). In short, Mencius is a judgment internalist, one who maintains that if an agent is able to acknowledge the claims of morality (or other normative claims), s/he already has the motivational resources to act according to those claims.

    At the same time, Mencius seems to recognize cases of weakness of will, both akratic (King Xuan’s admission to Mencius, “Great are your words”, and then his immediate confession of weakness, “I am fond of valor” in 1B3) and acedic (those who reject themselves in 4A10, saying “I do not think I am capable of abiding by benevolence or of following rightness”, or even more clearly Ran Qiu in Analects 6.10/12 who says “It is not that your Way does not commend itself to me, but that it demands powers I do not possess”). [Of course, we may also read Mencius or the Analects as saying that these akratic or acedic excuses are simply false claims, but the way Mencius handles these excuses, by skilfully channeling the overpowering or non-moral motives of the king toward rightness or benevolence, suggests to me that Mencius did recognize these cases of weakness of will, and suggested cures to it on a case-by-case basis.]

    As I suggested earlier, Mencius’s judgment internalism and his recognition of weakness of will can be reconciled in part by noting that in acedic cases one is concerned and motivated to act in accordance with what one judges to be good or right, but the motive force is insufficient to move the agent all the way to action.

    But how about akratic cases where there are overpowering temptations preventing one from doing what’s right? Why is it that, according to judgment internalism, one judges the overpowered reason for action to be better than the stronger motives for action? Here again I defer to (what I recall to be) the Humean resolution of the puzzle. On the one hand we have concerns rooted in more stable and basic affective dispositions, and on the other hand we have less stable/basic concerns are more immediately present to us, or are kept more constantly in our view, so that they tend to affect us more viscerally and overpoweringly at any given moment. Whenever we recognize one course of action as a superior reason for action, but fail to act on it due to some overpowering temptation, let me suggest that what we recognize as the superior reason is rooted in our more stable/basic affective dispositions, and that’s why we recognize it as such, even when we fail to act on it. Mencius too, I suggest, can have recourse to this type of explanation.

  52. Boram, this is really interesting, but very rich and complex, and I’m not sure what the basic argument is. In the sentence right after the long quote it almost looks like a slide from “Mencius thinks the truth of my judgment implies my motivation” to “Mencius thinks my making my judgment involves my motivation.”

    Here are some other questions/worries.

    I wonder how strong the evidence is that Mencius recognizes such an entity as (a) a judgment that one ought to do D, as distinct from (b) some intention or concern to do D, so that he can recognize the view that the two necessarily coincide. You do give some evidence: the king said “great are your words!” But that’s not the Mencius character speaking, and it could be empty politeness, and it’s words rather than judgment.

    One way of describing internalism is that it is the claim that each person necessarily takes morality as her practical standard or as one of her practical standards, by which she decides on actions. But in 1A7 and 6B4, Mencius seems to think that there is a real possibility that people will led by example to make their standard of action li (profit) rather than renyi (morality). “Your majesty says ‘How can I profit my state?’ — the Counsellors say ‘How can I profit my family?’ …” (1A1, Lau revised). The worry here doesn’t seem to be about the transient akrasia of Hume’s violent passions. What do you think?

  53. The word ‘acknowledge’ toward the end of the paragraph after the long quote suggests to me that maybe you’re thinking of moral knowledge rather than moral judgment. Moral knowledge implies “can”, not because knowledge is judgment, but because knowledge is true and “ought” implies “can”. ??

  54. Dan, further to your #32 about zheng 正, which answered my question in #30:

    If Mencius’ “wu zheng xin 勿正心” is advising against direct aiming at something, then either (a) direct aiming is zheng 正 and he’s advising against aiming directly at xin 心, or (b) direct aiming is zheng xin 正心 and he’s advising against aiming directly at something – the context suggests yi 義.

    The usage you report from Xunzi is significant support for (a) and insignificant support for (b).

    Legge favors (a), for he translates the phrase as “without the object of thereby nourishing the passion-nature.” But an argument against (a) is that when distinguishing sprouts or whatever from full virtue, Mencius uses the term “heart” for the former rather than the latter (2A6, 6A6). At least that’s my first impression — you’ve looked much more carefully at this sort of thing.

    Bryan seems to favor (b), since he translates the phrase “do not aim at it directly”.

    So Legge’s reading is problematic, and Xunzi’s usage doesn’t significantly support Bryan’s reading.

    Another possibility I guess is that zheng xin 正心 is a craft metaphor – the kind of violent “straightening” that Slingerland talks about somewhere in the Confucius section of Effortless Action. But that seems quite a stretch.

  55. Bill, thanks for your great questions, as usual!

    My line of argument takes as its starting point the following: like most ethical philosophers (including Chinese philosophers), Mencius implicitly holds the dictum ‘ought’ implies ‘can’. And the gist of the argument is this: the only plausible sense in which Mencius holds this dictum (given what he tells King Xuan of Qi in 1A7 and Gongsun Chou in 7A41) is that judgment of obligation to φ implies motivation to φ. And this, of course, is judgment internalism, and the desired conclusion of my argument was to show that Mencius was a judgment internalist. (Aside from this, in Comment #58, I also argue that Mencius allowed for weakness of will and in fact took it to be a serious problem, and that Mencius’s judgment internalism can be reconciled with his recognition of weakness of will.)

    As to the possible slide you mention between:
    (a) the truth of my judgment implying my motivation, and
    (b) my making my judgment involves my motivation.

    What I had in mind all along is (b), not (a). Does my argument (sketched above) rely on a slide from (a) to (b)? I hope not, but if you can point it out that would be very helpful to me, and I would appreciate it.

    Now, of course, judgment internalism can be fleshed out in different ways. Does one’s judgment that one ought to φ imply the motivation to φ (a) because the judgment gives rise to the motivation, or (b) because the motivation gives rise to the judgment, or (c) is there some some other explanation? My answer is (c): one’s affective dispositions in their cognitive role give rise to the judgment, and in their conative role provide the motivational push. Here my use of the word “cognitive” is quite misleading, so let me clarify: all I mean is that emotions have intentional/representational content that can figure in practical deliberation, and not that I endorse a cognitivist account of practical judgments (though I do think we need to provide a cognitivist account of theoretical judgments). I hope it’s clear that I’m trying to give a noncognitivist, perhaps even expressivist account of practical judgment. This is how I read Mencius. Convergence towards the same moral judgment is not due to the truth-aptitude of our practical judgments, but because we all have the same human nature, and have the same basic affective dispositions to respond to paradigm scenarios in similar ways, and the same capacity to extend these dispositions via rational deliberation.

    So by “acknowledging” that one ought to φ, all I mean is that one is admitting the normative force of that ought, or owning up to it. I don’t mean it in the cognitivist sense.

    Bill, you also ask: “I wonder how strong the evidence is that Mencius recognizes such an entity as (a) a judgment that one ought to do D, as distinct from (b) some intention or concern to do D….”

    My response: I don’t think it’s necessary to make a distinction between the judgment that one ought to φ from the intention to do φ, in order to claim that Mencius was a judgment internalist. Belief is not just judgment, but judgment held with conviction. In analogous fashion, I would claim that an intention or a desire is a motivating judgment.

    You also raise the worry that the king might not be sincere when he says “Great are your words!” Perhaps. All I can say is that I think the editor(s) intended the king’s words to be taken at face value. I read these words as expressing the king’s endorsement of morality (i.e., the “hooray!” part of expressivism), and the book Mencius does seem to contrast such endorsements with condemnations of morality (see 4A10 on the self-violators). As opposed to those who condemn morality (self-violators), there are those who apparently acknowledge the claims of morality but insist that they are unable to abide by such claims (self-rejecters). King Xuan, in claiming “Great are your words”, is definitely not a self-violator, but he is possibly a self-rejecter. Mencius’s lifting-feather:lifting-Mount-Tai::feeling-compassion-for-people:feeling-compassion-for-animals can be seen as a preemptive argument against King Xuan’s self-rejection, not against self-violation. Against self-violation, there seems to be no argumentative remedy (“it is impossible to talk with those who violate themselves”; also see Confucius’s reaction to Zai Yu in Analects 17:21).

    You make an excellent point in the concluding paragraph of Comment #59. Yes, the mental habit of engaging in profit-driven instrumental reasoning is different from “transient akrasia of Hume’s violent passions” as you so aptly put it. I think here too we can extract answers from Hume and Mencius, though I suspect the answers will not be entirely adequate to your challenge. This is something I should reflect on more deeply, and I won’t be able to respond over the weekend and the next couple of weeks due to grading and other responsibilities. But I will eventually return to it, I promise!

  56. Thanks Boram. You’ve complicated for me the question whether Mencius has the terms (e.g. “judgment of ought”) to articulate judgment internalism, to the point that I don’t know what to say or ask about that now.

    But your #63 doesn’t help me out of the worry I described briefly in the middle of #59 and again in #60.

    If I’m missing a key point, I think what I’m missing is the following, and I’m still missing it: I don’t understand *how* internalism can be a form of the dictum that ought implies can. The move goes by very fast in #63: “the only plausible sense in which Mencius holds this dictum … is that judgment of obligation to φ implies motivation to φ. And this, of course, is judgment internalism.”

    My impression is that you are attributing to Mencius Dictum Prime: “If a person ought to φ, that person is motivated toward φ.”

    Dictum Prime implies (A): “If a person knows that she ought to φ, that person is motivated toward φ.” For knowledge is true. But (A) is not internalism.

    Dictum Prime implies (B): “Propositions about obligation have implications about motivation.” Hence (B1) Alice’s judgment that a person ought to φ implies that the person is motivated toward φ. Hence (B2) Alice’s judgment that Alice ought to φ implies that Alice is motivated to φ – in the sense that if her judgment is true, then she will in fact be motivated to φ. One could put the point this way: “judgment of obligation to φ implies motivation to φ.”

    But I don’t see how Dictum Prime implies (C): “If a person judges she ought to φ, that person is motivated toward φ.” And (C) is judgment internalism.

  57. Oh I see what you are getting at, Bill, you put it crystal-clearly in your last comment. (Sorry it took me so long to see it, I’m very slow most of the time.)

    I do not subscribe to Dictum Prime, because it seems to me false. A third-personal judgment that “King Xuan ought to φ”, even if true, cannot motivate King Xuan if the king himself does not acknowledge the obligation.

    So I should have added that in ‘ought’ implies ‘can’ interpreted as judgment internalism, the ‘ought’ judgment must be made from the first-personal standpoint (King Xuan judges “I ought to φ”), though the ‘can’-statement may be made from a third-personal standpoint (Mencius states “King Xuan has the motivational resources to φ”).

    Thank you for pressing me on this point! I’m willing to clarify further if my position is still unclear.

  58. Just to make it clear what I think Mencius’s judgment internalism amounts to (though the statement below probably isn’t clear):

    X’s first-personal judgment “I ought to φ” implies that X has the motivational resources to φ.

    (I get the nagging feeling I’m still not explaining it clearly. If so, I hope you will all be patient with me and challenge me further.)

  59. Boram, I don’t see a difference between the view you lay out in #67 (and #66) and the view I called “(B2)” in #65.

  60. Except that the view in #67 is a little farther from internalism than is (B2), since the view in #67 speaks only of motivational resources, not motivation.

  61. Boram, you write in #66: “I do not subscribe to Dictum Prime, because it seems to me false. A third-personal judgment that “King Xuan ought to φ”, even if true, cannot motivate King Xuan if the king himself does not acknowledge the obligation.”

    I don’t understand your argument against Dictum Prime. DP does not imply that my judgment “Xuan ought to φ” can motivate Xuan. Rather, DP implies that if the judgment is true, Xuan is motivated.

    I thought you were ascribing to Mencius either Dictum Prime or a weaker view that speaks only of motivational resources. Dictum Prime, being stronger, has more chance of implying internalism.

  62. Hi Bill, I’m finally done with grading (was done with it yesterday, but my brain was too numb and stupid after grading all those exams). Anyway, I’m back again to resume our conversation, though this thread seems to have run out of steam!

    Thanks very much for your questions, which forces me to sharpen my vague thoughts. Now, here’s why I’m wary of Dictum Prime, looking at it as an expressivist. As Hamlet put it, “there’s nothing either good or bad, but thinking makes it so”. Along the same lines I want to say, “There’s no φ that X ought either to do or not to do, but Y’s feeling obligated to φ makes it so (where it may or may not be the case that X = Y)”. Let’s call this expressivism.

    Given expressivism, claiming “X ought to φ” is a lot like claiming “Soymilk tastes good”. With the latter claim, we want to ask, tastes good to whom? It may taste good to Tom, but not to Dick and Harry. Ascending from the material to the formal mode of speech, we want to ask, the claim is true for whom?

    As with the claim “Soymilk tastes good”, so with the claim “Tom ought to φ”. For an expressivist, the claim is not baldly true, it’s not true simpliciter. It may be true for Tom, but not true for Dick and Harry.

    So from this expressivist standpoint, the truth value of the antecedent of Dictum Prime is indeterminate, unless you specify who the antecedent is true for. If Alice judges truly-for-her (because she feels obligated to φ) that Lewis ought to φ, then in my view it does not necessarily follow that Lewis is motivated to φ. My position is that, if Lewis judges truly-for-him that he ought to φ, then it necessarily does follow that Lewis is motivated to φ. Or, if Alice judges truly-for-Lewis (because Lewis feels obligated to φ) that Lewis ought to φ, then too it necessarily follows that he is motivated to φ.

    For these reasons, I don’t think your (B2) follows as a special case of (B1), assuming the truth of expressivism.

    Let’s continue assuming the truth of expressivism. But let me restate it, as follows. For some person(s), X and Y, where it may or may not be the case that X = Y,

    EXPRESSIVISM: If X judges that Y ought to φ, then X is emotionally disposed to φ.

    Then combine this with the following psychological thesis, about the connection between emotional disposition and motivation:

    If X is emotionally disposed to φ, then X is motivated to φ.

    Add to this the qualification that, even if X is motivated to φ, X may not be moved all the way to φ.

    Then by hypothetical syllogism (if p then q, if q then r, therefore if p then r), we get:

    If X judges that Y ought to φ, then X is motivated to φ.

    Which I take to be judgment internalism.

  63. Oops. In light of my qualifications about true for, let me revise as follows. For some person(s) X and Y, where it may or may not be the case that X = Y,

    (1) If X judges-truly-for-Y that Y ought to φ, then Y is motivated to φ.

    (2) If Y is emotionally disposed to φ, then Y is motivated to φ.

    Then by hypothetical syllogism (if p then q, if q then r, therefore if p then r), we get:

    (3) If X judges-truly-for-Y that Y ought to φ, then Y is motivated to φ.

    Which I take to be judgment internalism. (This is getting complicated…. My head hurts!)

  64. Hi Boram, thanks! I’m looking forward to thinking through this. I’m not sure how to read “judges-truly-for-Y.”

  65. Sometimes I really dislike the fact that you can’t edit your own comments. Read (1) in Comment #72 as follows:

    (1) If X judges-truly-for-Y that Y ought to φ, then Y is emotionally disposed to φ.

    So, Bill, looking at your (B1) and (B2) again I will acknowledge that (B2) can follow as a special case of (B1), if (B1) is read as my (3).

  66. Boram, this is great mind-bending fun. I don’t think I understand you yet! Maybe these questions and names-of-statements will help.


    First I’m going to restate my old position, including names of propositions and speculations about your line of thought about how a Mencian version of “Ought Implies Can” could support Judgment Internalism.

    (OIC): Alice ought to φ implies Alice can φ.
    (JI): “If Alice judges that Alice ought to φ, then Alice is motivated toward φ.”

    Mencius’ version of this Dictum might be
    (DPP): If Alice ought to φ, Alice can be motivated toward φ.”
    Or, a little less plausibly,
    (DP): “If Alice ought to φ, Alice is motivated toward φ.”

    Then there are two ambiguous formulae.

    (B1) as I presented it is of course an ambiguous formula.
    (B1a): “If Alice’s judgment that Smith ought to φ is true, then Smith is motivated toward φ.”
    (B1b): “If Alice’s judgment that Smith ought to φ exists, then Smith is motivated toward φ.”

    (B2) in my very first expression of it is similarly ambiguous:
    (B2a): “If Alice’s judgment that Alice ought to φ is true, then Alice is motivated toward φ.”
    (B2b): “If Alice’s judgment that Alice ought to φ exists, then Alice is motivated toward φ.”

    (B2b) = (JI).

    (B1a) formally implies (B2a), and (B1b) formally implies (B2b)

    My worry about your line of thought was this: “Let’s grant that (DP) is a version of (OIC). That’ just not enough for judgment internalism. (DP) plainly implies (B1a) and (B2a). I don’t see how (DP) would imply (B1b) and (B2b).”

    I gather your Expressivism is supposed to explain how DP yields judgment internalism.


    You say this is judgment internalism:
    (JIL): “If X judges that Y ought to φ, then X is motivated to φ.”

    That doesn’t look like judgment internalism to me. To make it into judgment internalism, one has to change the Y to X, so that only one person is involved. Yes? And is that what you meant?


    I’m not at all worried about the differences among these: feeling obligated to φ, being emotionally disposed to φ, being motivated toward φ, and being motivated to φ. I’ve taken no notice of such differences in structuring my thoughts and arguments. So, for example, from the premse (e) “If Alice judges that she should φ, Alice feels an obligation to φ” I am happy to infer (JI) “If Alice judges that she should φ, Alice is motivated to φ”. I’m never questioning any such move.

    Now, whatever Expressivism is, it probably implies (e) and thereby (JI). So from my point of view that leaves you with these tasks: state Expressivism and show that it follows from some Mencian version of (OIC).

    I don’t understand what you mean by Expressivism. (Here’s an assumption I make in reading you: I think that often when you say “for some X…”, what you really mean is “for all X …”.)

    In #71 you seem to offer two different expressivisms, one near the beginning and one later.

    (E1) “There’s no φ that X ought either to do or not to do, but Y’s feeling obligated to φ makes it so (where it may or may not be the case that X = Y)”.

    (E2): If X judges that Y ought to φ, then X is emotionally disposed to φ.

    (I guess that to see how these are related, one should reverse the X and Y in one of them.)

    (E1), like (DP), talks aout how obligations are related to feelings, and does not explicitly broach the topic of judgments.
    (E2), like (JI), does the reverse: it talks about how judgments are related to feelings, but does not explicitly broach the topic of obligations.

    Here’s another interesting theory:
    (E3): Alice’s judging “Smith ought to φ” is Alice’s having a certain positive feeling about Smith’s φ-ing.

    I don’t understand (E1). Does it imply that for all persons X and Y, if Y feels obligated to φ, X is obligated to φ? Does it mean that for all persons X, X is not obligated to φ unless at least one person feels obligated to φ? If the latter, why not make the view this: “For all persons X, X is not obligated to φ unless X feels obligated to φ”?

    I’m not sure I understand (E2). Does it imply that nobody thinks obligations differ from person to person?


    “X judges truly-for-Y that Z ought to φ.”

    That’s a locution for use within an Expressivist framework. Still, I feel I can’t tell quite what it is supposed to mean. Here are some possibilities one might construct within an (E3) framework:

    X and Y feel positively about Z’s φ-ing.
    X believes that Y feels positively about Z’s φ-ing.
    X and Y feel positively about φ-ing by X, Y, and Z.

  67. I said early in #77C that you have to show that Expressivism follows from some version of (OIC). Why did I say that?

    Because I suspect that once I find out what you mean by Expressivism, I’m going to think that (JI) follows directly from Expressivism. If we need Expressivism to get from his (OIC) to (JI), as you seem to say, but his version of (OIC) doesn’t imply Expressivism, then it would seem that his (OIC) does no work as evidence of his accepting (JI). But you seem to hold that it does the key work.

  68. If we use the term ‘expressivism’ in the traditional way, expressivism is the idea that moral language that seems to be of the right form to express judgments that could be true or fase don’t in fact express judgments. There aren’t moral judgments—not the kind of “judgments” that can be true or false, anyway. To say “It is right to φ” is not to state anything; rather it is to express a positive feeling or attitude about φ; it’s equivalent to “Hooray for φ”. Let’s call traditional expressivism “(E4)”.

    (E4) doesn’t seem to leave room for a notion of “true-for-X”, though it leaves room for correct and incorrect uses of moral language by X. X uses the language correctly when she chooses words that fit the feelings she has. Thinking the words would, I suppose, count as expressing something to oneself.

    Thus if Alice says “Smith ought to φ” but Alice is against Smith’s φ-ing, Alice has misused the words. She has said “Hooray” for something she is against.

    (That doesn’t imply that she was being insincere. She could have been mistaken about her own feelings or attitudes, so that her sincere sentence did not fit her feelings. On the other hand, maybe she was being insincere. Maybe she was trying to mislead about her feelings, or anyway not trying to express her feelings.)

    I’ll give another name:

    (XYZ): “X judges truly-for-Y that Z ought to φ.”

    Friends of (E4) might read (XYZ) this way:

    (XYZe): “X sincerely says ‘Z ought to φ’, and Y feels positively about Z’s φ-ing.”

    Note that (XYZe) as it stands makes no claim about whether X’s language fits X’s feelings about Z’s φ-ing. Also (XYZe) does not imply that X has Y in mind at all. Other readings of (XYZ) might take different stands on those matters.

  69. Bill, thanks again for your detailed response! Largely owing to your patient questions my thoughts are crystallizing into something, hopefully something clearer and better.


    As for judgment internalism, I finally see the problem you are getting at, and I also see how I seem to shift my position back-and-forth on the matter. The position I’m gesturing towards is complex, but I hope not confused (though it very well may be).

    Since I believe that emotions have intentional or representational content, content that practical deliberation can operate on, I also believe that being emotionally disposed to φ involves the judgment to φ. Since this judgment is part and parcel of the emotion, it cannot be false (and it cannot be true either, because the judgment has a world-to-mind, not a mind-to-world direction of fit, i.e., the world must be changed to fit the judgment). Also, I don’t think the content of that judgment necessarily takes the form that one ought to φ. The oughtness, or rather one’s feeling obligated, comes from the conative force of that emotion expressing itself as a motivating desire in response to outside stimuli. (Perhaps this can be compared to Hume’s analysis of “necessary connection” between cause and effect.)

    Let’s call the judgment that’s part and parcel of the emotion “emotional judgment”. One’s emotional judgment may or may not give rise to the meta-judgment “I ought to φ”, which explicitly contains the word “ought”. This meta-judgment may be true or false, and the expressivist position is that the ultimate truth-makers for the meta-judgment are emotional judgments. Also, the meta-judgment is something that one can truly make of another.

    The judgments involved in judgment internalism may be emotional judgments (EJ) or meta-judgments (MJ), and I formulate judgment internalism differently with respect to these two types of judgments:

    (EJ) If X judges that X [ought to] φ, then X is motivated to φ.

    That is to say, X’s making the judgment implies X’s motivation, because the judgment is part and parcel of being emotionally disposed to φ, and this in turn implies being motivated to φ.

    And, where it may or may not be the case that X = Y:

    (MJ) If X meta-judges-truly-for-Y that Y ought to φ, then Y is motivated to φ.

    Here X’s making a judgment that’s true-for-Y implies that Y is motivated to φ, because the truth of the meta-judgment implies that Y has an emotional judgment, which together with (EJ) and modus ponens implies that Y is motivated to φ.

  70. Boram, that’s extremely clear and helpful.

    “Since I believe that emotions have intentional or representational content, content that practical deliberation can operate on, I also believe that being emotionally disposed to φ involves the judgment to φ. Since this judgment is part and parcel of the emotion, it cannot be false (and it cannot be true either…)”

    That’s two inferences. I wonder whether the first of them goes both ways. That is, do you distinguish at all between being emotionally disposed to φ and judging to φ?

    I wonder why you find the second inference valid. Language, like emotion, has representational content, so sincerely accepting a sentence commonly amounts to having a belief. But that doesn’t suggest that the belief has no truth-value. Why would the fact that having an emotion amounts to accepting a judgment (because emotions are representational) suggest that the judgment has no truth-value?

    The first potato chip from the bag is, in my mouth, a representation. Its tastiness and saltiness is a vehicle by which I have a sense of the chips to come. I am perceiving what they’ll be like. Or, if it’s a bad batch and the others aren’t like that, then I am misperceiving: my judgment is false.

    Hume ought to have allowed that passions can represent passions on grounds of similarity, as the one potato chip represents the rest.

    The anticipatory pleasure before a kiss is a representation of the pleasure to come. It’s pleasure with representational content. Like the potato chip, it’s a Peircean icon. But it could be false (insert funny story here).

    The sympathetic distress I feel when I see a tot by a well is, I think, the same sort of icon. It depicts to me the distress in store for the child; it’s at once evidence for that and a representational vehicle of my judgment or knowledge of that.

    But my judgment, being a judgment, could be wrong.

    I think that when I use the right procedures to arrive at my feelings about my options—procedures such as imagining “what if that were done to me?”—my feelings embody judgments of the overall net pleasure knowably at stake in those options (weighting each unit of possible pleasure by its degree of predictability using the best methods, which include sympathetic imagination). And because what I’ve just said is very vague, there’s room for an interpretation of it on which the judgments turn out to be infallible. But in fact I think we can be mistaken about what we should do. In trying to make my view more precise (over the years) I’ll try to save that phenomenon.

  71. 1. I missed at first your distinction between “emotional disposition” and “emotion,” but I don’t think that ruins what I said. I have this disposition: when I think of iced tea, I think of it with pleasure. The image is pleasant and brown. By the vehicle of those dispositions I believe that iced tea is pleasant and brown.

    2. Is it your view that these three are identical:

    X ought to φ
    X judges to φ
    X is emotionally disposed toward φ

    Let’s call the claim that they’re identical (3I).

    3. What if Alice has some emotional disposition toward each of several courses of action that she recognizes as incompatible alternatives? Do you hold that in that case Alice judge to each of them, and Alice ought to do each of them?

    4. Is the following an accurate paraphrase of your (MJ)?:

    (MJa) If Y ought to φ, and X sincerely asserts “Y ought to φ,” then Y is motivated to φ.

    Note that (3I) implies that in (Mja)’s antecedent, the second conjunct is superfluous.

  72. Here is a follow-up comment, explaining how your (DP) or Dictum Prime relates to my (EJ).

    Supposing that expressivism is true, then the antecedent of (DP) and the antecedent of (EJ) are materially equivalent. Namely, I ought to φ iff I make the emotional judgment that I φ. But the right-hand-side explains the left-hand-side: “I ought to φ” is true because I make the emotional judgment that I φ. So, given expressivism, (EJ) explains the truth of (DP).

    Bill, in Comment #77 (A:Review), you attribute to me the position that “a Mencian version of ‘Ought Implies Can’ could support Judgment Internalism” (emphasis mine). But that has never been my position! As far as I am aware, I have never stated that there’s a relation of argumentative support between the dictum and judgment internalism. What I claimed is this: I can see a number of ways of interpreting the dictum ‘ought’ implies ‘can’, and I’ve suggested that judgment internalism is one way of interpreting the dictum. For instance, ‘ought’ implies ‘can’ might mean:

    (1) obligation implies permissibility

    (As deontic logic tells us, on analogy with the rule in modal logic that necessity implies possibility.)

    (2) responsibility implies freedom

    (If I am responsible for having φ-ing, then I can φ or not-φ.)

    And so on.

    More precisely, I’ve suggested that, insofar as Mencius is thought to subscribe to the dictum, we must take him to be subscribing to the judgment internalist interpretation of it. And that’s because of the following puzzle.


    If one accepts that ‘ought’ implies ‘can’, then one ought to accept the converse, ‘cannot’ implies ‘not-ought’.

    Now, Mencius seems to accept the dictum that ‘ought’ implies ‘can’. I.e., he does not stop at saying that we ought to be moral sages, but is very much concerned to show that we can be: e.g., in 6A6 Mencius says that as far as the essential part of our nature (the part that distinguishes us from animals) is concerned, we are capable of becoming good (乃若其情,則可以為善矣).

    But in 7A41, when Gongsun Chou implicitly uses the converse of the dictum, i.e., ‘cannot’ implies ‘not-ought’, to argue that since the Way is too difficult to follow with our powers we should attempt something easier, Mencius rejects the argument.

    So what’s really going on here? Is Mencius contradicting himself, both endorsing and rejecting the dictum? I think not. What we have to do is to provide a plausible interpretation of the dictum on which Mencius is not contradicting himself.

    And I offered judgment internalism as a plausible way of interpreting the sense in which Mencius endorses the dictum. My ought-judgments imply that I am motivated to act as my judgments dictate. And the converse, that if I cannot be motivated to act in a certain way, then there is no obligation for me to do so. (Here Mencius seems to share common ground with Confucius. Confucius, in Analects 17.21, converses with Zai Wo, the Confucian version of Thrasymachus or the Foole or the sensible knave. Confucius says to Zai Wo, in effect: if you are not emotionally disposed and motivated enough to engage in prolonged mourning, then by all means you need not do so… but it’s inhuman not to feel emotionally disposed to engage in prolonged mourning! Mencius, in his conversation with Yi Zhi in 3A5, observes how people are emotionally disposed to react to the exposed bodies of their parents, and from there arrives at the judgment that it is right to inter the bodies.)

    Here then is my proposed solution to the puzzle:


    Gongsun Chou himself admits that the Way is lofty and beautiful, and this judgment implies that Gongsun Chou is emotionally disposed to follow the Way, and thus motivate to follow it. Thus he can follow the Way. On the judgment internalist interpretation, Mencius’s response to Gongsun Chou is entirely consistent with the dictum.

    Confucius would add that Gongsun Chou is drawing the line even before he has begun. And Confucians in general would advise him that so long as he makes the effort, and does all that is within his power to follow the Way, then he is doing right, and whether he fails or succeeds is a matter that he should leave up to Heaven or fate. As far as I can see judgment internalism is consistent with all these responses.


    (This comment I had worked out Comment 81, so here I am working with what I’ve labeled (MJ). Work-in-progress!)


    As for expressivism, I cannot undertake to give a full-blown account of it here, only catchphrases that give snapshots of the intended position. My Hamlet-ripoff line seems to have complicated things unnecessarily. Hamlet’s statement says that goodness or badness ascribed to states of affairs is not a mind-independent property, and my ripoff statement says the same about the property of oughtness ascribed to actions. There must be some mind or other that feels and thinks that Hitler ought not to have brought about the Holocaust, and that’s all I meant to indicate with my X’s and Y’s.

    I offered another snapshot of the position via comparison with “Soymilk tastes good”. Taste varies from one person to another, and from culture to culture, contra Mencius 6A7. But it’s not the variability that primarily concerns me, but the relativity of the judgment to a perceiver. Take “Grass is green”. That judgment does not vary from one person to another, so long as they don’t have jaundiced vision, etc. Nevertheless the truth of the judgment is relative to a perceiver, and we arrive at the same judgments because we have the same visual mechanisms. If grass looks red to a Martian, and he says “Grass is red”, I would be wrongheaded to insist that the Martian is mistaken. I meant to capture the relativity of such mind-dependent judgments to perceivers through the notion of “true-for”.

    Of course you are right, in Comment #80, to point out that, according to “traditional expressivism” (we may also call it boo-hooray expressivism), to say “One ought to do this or that” is not to claim anything true or false, but to express one’s feelings. Since I’m not a verificationist like Ayer, I need not subscribe to the boo-hooray version of expressivism. Besides, there is the Frege-Geach problem, which notes that ought-statements can be embedded in more complex sentences that are true or false, say material conditionals. A more complex version of the problem (which I think is due to Christine Tappolet) is that ought-statements can be embedded in deductively valid arguments with at least one factual statement that can be true or false, where the conclusion necessarily follows from the premisses. Since a valid argument is truth-preserving, the ought-statements must be true or false after all.

    Given these problems, and the downfall of logical positivism, it seems reasonable to revamp expressivism, say by combining it with relativism (implicit in my use of the notion “true-for”). Which reminds me, I’ve got to read David Wong’s MORAL RELATIVITY and NATURAL MORALITIES this summer!


    Let’s now get to my awkward statement of expressivism:

    If X judges-truly-for-Y that Y ought to φ, then Y is emotionally disposed to φ.

    This is just another snapshot statement of expressivism. The antecedent is meant to capture the requirement that the truth value of ought-judgments must be relativized. So let’s call this “relativistic expressivism”, to be distinguished from boo-hooray expressivism.

    Now I allow that X and Y may be distinct, and perhaps a concrete (that we all know so well) will help you see why. Take Mencius and King Xuan in 1A7. There Mencius judges truly-for-the-king what the king ought to have done (to have spared the unseen sheep just as he spared the ox: 王若隱其無罪而就死地,則牛羊何擇焉), and explains the emotional disposition in the king that moved him to feel that the ox ought to be saved, which, had the king reflected, would have extended to the sheep as well. Then the king says to Mencius, “The Book of Odes says, ‘The heart is another man’s, / But it is I who have surmised it.’ This describes you perfectly. For though the deed was mine, when I looked into myself I failed to understand my own heart. You described it for me and your words struck a chord in me….” (Lau trans.)

    Of course there can be less striking, very mundane, examples of my judging-truly-for-others what they ought to do: simply by accident, or owing to our biologically shared or culturally shaped emotional makeup, and so on. But with someone like Menicus judging-truly-for-another can be a cultivated art much like psychotherapy, demanding the same sort of reaction from the patient as in psychotherapy (Wittgenstein somewhere says that in psychotherapy the patient himself must recognize the diagnosis and cure as such, and in the same way King Xuan himself must recognize that he ought to do what Mencius tells him he ought to do, given his own emotional dispositions).

  74. Hi, Bill! I’ve been typing Comment #84 as you were posting Comments 82~3. I don’t think I can keep up with your mental stamina! (I’ll have to rest a lot and then read your comments next week.)

    Anyway, I’m benefitting tremendously from your comments, and I hope to work this stuff out fully and include it in my dissertation. Do you happen to know any relevant literature on this that I should be looking up?

  75. Bill, just a quick comment on your:

    “I wonder why you find the second inference valid. Language, like emotion, has representational content, so sincerely accepting a sentence commonly amounts to having a belief. But that doesn’t suggest that the belief has no truth-value. Why would the fact that having an emotion amounts to accepting a judgment (because emotions are representational) suggest that the judgment has no truth-value?”

    That’s because of the direction of fit. Belief has mind-to-world direction of fit, i.e., mind must change to fit the world. Emotional judgment that I φ, like a desire, has a world-to-mind direction of fit: the world must change to fit my emotional judgment or desire.

    Think of imperative statements. Those are neither true nor false, right?

  76. Boram, I think I’m going to be out of commission here for a week or so too. Catch you later!

  77. Hi Boram, I know you’re very busy and haven’t finished replying to what I’ve already posted above, but I’ll add this …


    In my comments I’ve been using such locutions as “X judges that X ought to φ” and “X is motivationally disposed to φ” as though they are clear. I’ve been pushing them around like tokens on a game board. I think anyone who is committed to the truth or intelligibility of Judgment Internalism is committed to the intelligibility of such talk. If those locutions don’t make sense, then neither does Judgment Internalism.

    Here are some questions about your locution ‘X judges to φ’.

    1. Does it address only X’s φ-ing, or is it about φ-ing in general, e.g. Y’s φ-ing?
    2. If it’s only about X’s φ-ing, can X do something regarding Y’s φ-ing that is just like “judging to φ” except that it’s specifically about Y’s φ-ing? And what would you call that thing that X does?
    3. To avoid prejudging the answer to question 1, I’ll here use ‘Q’ to stand for the party or parties whose φ-ing is at issue when X judges to φ. Here’s my question: Since I think you think it’s better to say “X judges to φ” than to say “X judges that Q ought to φ,” and since I gather you are identifying X’s judgment to φ with X’s having an emotional disposition, I wonder why you don’t simply put your view this way: “There are no ought-judgments.”


    Re your #84 re my #77—I suppose by “support” I just meant “imply.” Sorry! I meant to be attributing to you the view that some statement that is prima facie a version of the Dictum implies Judgment Internalism.

    Examples of prima facie versions of the Dictum: The (1) and (2) you list in #84, and DPP.
    Examples of statements that are not prima facie versions: DP and Judgment Internalism.

    What I still don’t get is how Judgment Internalism is supposed to be implied by, or equivalent to, some prima facie version of the Dictum (or even DP).

    I think in the big paragraph in #84 you’re assuming that Judgment Internalism can be regarded as a version of the Dictum. That’s what I don’t get.

    Here’s a new fallacious argument from a prima facie version of the Dictum to Judgment Internalism:

    1. That I ought to do X implies that I (am able to) appreciate that I ought to do X.
    2. Appreciating that I ought to do X involves both (a) judging that I ought to do X, and (b) being motivated toward doing X.
    3. Ought implies judgment, and ought implies motivation. (1,2)
    4. Judgment implies motivation. (5)

    The inference to (4) is invalid on its face. That’s the main problem.

    The inference to (3) is invalid because it forgets the parenthesis in (1). The parenthesis in (1) is what makes (1) prima facie a version of the Dictum, and in my view it’s what allows (1) to be plausible.

    “Hamlet’s statement says that goodness or badness ascribed to states of affairs is not a mind-independent property, and my ripoff statement says the same about the property of oughtness ascribed to actions. There must be some mind or other that feels and thinks that Hitler ought not to have brought about the Holocaust, and that’s all I meant to indicate with my X’s and Y’s.”

    “Feels and thinks” reminds me of the new fallacious argument above, both because it sounds like (3), and because it seems to forget the parenthesis in (1). If the wrongness of Hitler’s act was easy to see and feel, but nobody actually went to the trouble of seeing and feeling that it was wrong, it could still be wrong; just as a sunset can be beautiful even if nobody sees it.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.