中大校定本 《經上》A01-A50 A51-A98
A01 經：故，所得而後成也。
說：故。小故，有之不必然，無之必不然。體也，若有端。大故。有之必(無)然。若見之成見也。
1. At first glance, the five graphs 體也，若有端 in the shūo may seem irrelevant to the subject of this canon. Since the subject of Jīng A2 is 體, some editors (such as 張惠言 and Sun Yirang) transpose these graphs to the shūo for A2. However, interpretively, the graphs do not fit neatly into Shūo A2 either, and it is difficult to see how they could have been displaced to their present position. Moreover, since parenthetical glosses or examples are common in the Jīng Shūo, if these graphs can be interpreted intelligibly in Shūo A1, then it is likely they belong there. The grammar of the string is questionable and raises the possibility that textual damage may have occurred. Yet the string can be interpreted as the parenthetical explanation, “It is a unit. Like having a starting point.”
Graham inserts the graphs 最前之 before 體也; Wu Feibai inserts the graph 尺 between 若 and 有. We believe these insertions are speculative and do not significantly improve the intelligibility of the text. Another possibility would be to emend 體也，若有端 to 若體之有端, taking 若 to have inadvertently been transposed with 體也 and taking 也 to be an graphical error for 之. However, the grounds for this proposal would be purely interpretive, and thus conjectural. We therefore leave the received text as it stands. We believe this decision is moderately well justified.
2. All editors agree that the sentence 大故，有之必無然 in the shūo is faulty, since it contradicts the equivalence formula Jīng A1 gives for 故, “the reason/cause is what must be obtained before something will occur/form” 故，所得而後成也. Two proposals for emendation have been put forward. On the basis of parallelism with the explanation of the 小故, Sun Yirang suggests inserting four graphs: 大故，有之必[然]，無[之必不]然. Liang Qichao proposes instead to omit one graph: 大故，有之必(無)然. We follow Liang’s proposal, on the grounds that it provides the simplest explanation of the text: A scribe accidentally inserted an extra 無 by dittography with the 無 in the first sentence of Shūo A1. By contrast, we see no reasonably simple explanation of how a scribe could accidentally omit just the 6th, 8th, 9th, and 10th characters of an 11-character sentence (大故，有之必□，無□□□然).
An objection to Liang’s proposal is that the resulting sentence states that 大故 is a sufficient condition, but the canon indicates that 故 is always a necessary condition (cf. Graham, 263). However, Liang suggests that repeating 無之必不然 in this sentence is redundant, since the canon makes it clear that this feature of the 小故 carries over to the 大故 as well. Hence our only emendation is to omit the second instance of 無. We believe this emendation is moderately well justified.
A03 經：知，材也。
說：知材。知也者，所以知也而必知。若明。
1. Several editors (梁啟超、吳毓江、姜寶昌、沈有鼎、楊俊江) follow 胡適 in inserting a negative, 不, before 必知 in Shūo A3. The grounds for doing so are that A3 thereby becomes parallel with the phrase 不必得 in Shūo A4, and that the received text does not make sense, since the capacity for knowing does not guarantee that one knows anything.
We take this to be a conjectural emendation, which could be justified only if there were no other reasonable explanation of the text. In this case, we believe the text is probably intelligible as it stands, and so emendation is unjustified. According to the canon and the phrase 所以知, the topic is the capacity for cognitive functions and conscious states. The phrase 必知 can reasonably be interpreted as explaining that as long as this capacity is functioning, then the agent recognizes or is conscious of something. It is unclear, however, whether the verb 知 in 必知 expresses the same concept as that explicated in Jīng A5. Possibly they are the same, but a more plausible interpretation may be that the verb in A3 expresses a distinct notion, roughly like the English “awareness.” The difference would be that the notion of knowing presented in A5 incorporates a normative constraint: only an agent who can describe something correctly counts as knowing. By contrast, to “know” in the sense of A3, we need only be awake.
2. 胡適 and 梁啟超 further propose emending 若明 to 若眼. This too is an unnecessary conjectural emendation. Since 明 can refer to eyesight, a capacity, it is fully intelligible here without emendation.
A04 經：慮，求也。
說：慮。慮也者，以其知有求也而不必得之。若睨。
A05 經：知，接也。
說：知。知也者，以其知過物而能貌之。若見。
Sun’s emendation of 過 to 遇 in the shūo is followed by several editors. This emendation is conjectural and redundant, however, since the interpretation of the sentence is roughly the same either way.
A06 經：
，明也。
說：(恕)[]。(恕)[]也者，以其知論物而其知之也著。若明。
1. By the principle of preferring the more difficult reading, the best explanation
of the text is that scribes unfamiliar with the otherwise unknown graph zhi
in
Shūo A6 corrupted it to the more familiar word shu 恕. Hence
we are highly justified in emending the two instances of 恕 in the shūo
to .
2. The unknown graph
is used here in a context in which we normally would expect the graph 智. Graham
has proposed a hypothesis to explain this and other observations (77). He points
out that 智 is used for “to know” in many of the shūo beginning with
B9, though the graph 知 is used in parallel contexts in the corresponding jīng.
智 even appears as the heading of Shūo B 34, 40, 46, and 48, where the
first graph in the corresponding jīng is 知. Accordingly, Graham proposes
that was
used in Jīng A6 because the Mohists originally used the graph 智 to
distinguish the word “to know” of Jīng A5, which is normally written
知, from the noun “the knowing” of A4, which is also written 知. Then, since this
use of 智 is unusual, at some point during the transmission of the canons a scribe
misguidedly began to “correct” instances of 智 meaning “to know” by emending
them to 知. However, the scribe did not finish this task, instead giving up when
he reached Shūo B9. Hence from B9 on, we observe a discrepancy between
the use of 智 in the shūo and 知 in the jīng.
If justified, this hypothesis would lead us to consider emending several instances of the graph 知 in Shūo A5 and A6 to 智. It would also lead us to systematically emend many instances of 知 to 智 in the jīng starting with B9, especially in the four cases where there is a discrepancy between 智 as the heading of the shūo and 知 as the first graph in the jīng. However, we believe the hypothesis is unjustified, and all of these potential emendations are too conjectural to be convincing. The problem is that Graham’s and other observations about the text seem better explained by the hypothesis that the original text simply did not distinguish these three graphs rigorously or systematically according to the explications in A3-A6. For, in addition to the features of the text Graham observes, we note the following:
(i) All instances of 知 are fully intelligible as they stand.
(ii) Jīng A80 uses only the single graph 知, but it distinguishes several sources and objects of knowledge, not all of which correspond to the notion explicated in A5, nor that in A6.
(iii) Shūo A75 uses all three graphs without any apparent principled distinction between them.
(iv) The shūo after B9 do not use 智 consistently in place of 知. Shūo B33 and B77 both use 知 in contexts whether other shūo use 智. B77 uses both graphs in parallel clauses.
(v) It is not true that, as a rule, the use of graphs in the jīng is consistent with that in the shūo, nor that graphs are always used consistently in different parts of the jīng or shūo. For instance, the graphs 缶 and 正, both expressing the word zhèng 'upright', are used inconsistently. Generally, the jīng uses 缶 and the shūo 正, but this is not a strict rule. The graphs for qū 'segregate' also seem to vary, and may include 樞, 歐, and 傴.
(vi) Nor is it true that, as a rule, the graphs used as headings in the shūo are identical to the first graph of the corresponding jīng. The headings in Shūo A39 (侗 vs. 同), A98 (正 vs. 缶), B12 (俱, perhaps an error for 傴, vs. 歐), and B62 (正 vs. 缶) are not identical to the first graph of the matching jīng.
,
but this assumption is difficult to square with points (v) and (vi). Therefore
we reject the hypothesis about the use of 知, 智 and .
Moreover, we make no emendations here or elsewhere that rely mainly on assumptions
about graphical consistency. A07 經：仁，體愛也。
說：仁。愛己者，非為用己也。不若愛馬。(著若明。)
1. Some editors (伍非百、楊俊光, 89-92) propose to delete the word tǐ 體 from the canon on the grounds that it is unintelligible or ungrammatical. We believe the canon is intelligible as it stands: “unit concern” 體愛 can plausibly be interpreted as concern for the individuals within any group, by analogy to one’s concern for oneself, as mentioned in the shūo.
2. Sun suggests that jǐ 己 may be a graphic error for 民. Again, the text seems intelligible as it stands, so this would be an unjustified, conjectural emendation.
3. The last three graphs in the shūo do not fit the context and are identical to the last three graphs in the preceding shūo. Editors have proposed two approaches for dealing with these graphs. One is to delete them on grounds of accidental dittography from A6. The other is to delete only the final two graphs and emend 著 to 者, thus making the grammatical structure of 愛馬者 parallel to that of 愛己者. We believe that in this parenthetical explanation, unlike in the preceding sentence, 者 would be grammatically redundant. The simpler explanation, requiring no emendation, is that a careless scribe accidentally reproduced all three from the end of Shūo A6. We consider this emendation highly justified.
A08 經：義，利也。
說：義。志以天下為芬而能能利之。不必用。
1. Sūn, followed by Liáng, proposes to emend 芬 to 愛. However, read as 份, meaning one’s role, part, or duty, 芬 is fully intelligible, so this emendation must be rejected as conjectural.
2. Yáng, following 張惠言, does not emend 芬 but interprets it as 紛, arguing that 以天下為芬 refers to taking the world as good or beautiful and overlooking people’s faults. We believe this interpretation coheres less well with the text than the reading as 份.
3. We conclude that no emendation is necessary, but 芬 should probably be interpreted as 份.
A09 經：禮，敬也。
說：禮。貴者公賤者名而俱有敬僈焉。等異，論也。
1. 高亨 proposes to emend 名 to 台 and 焉 to 差 in Shūo A9 on the grounds that a similar remark is found in 《左傳》：「王者公。公臣大夫。大夫臣士。士臣皁。皁臣輿。輿臣隸。隸臣僚。僚臣僕。僕臣臺。」This emendation is conjectural and in any case does not significantly change the meaning of the sentence. Hence we reject it.
2. Several ways of punctuating the shūo have been proposed. Graham reads it as …而俱有敬僈焉，等異論也, translating the last four graphs as “modes of behavior are different for different ranks.” Chen has …而俱有敬僈焉。等異，論也, presumably roughly “for ranks to be different is relation-grading.” Yang has …而俱有敬，僈焉等異，論也. As this is an interpretive issue, we here leave it open.
A10 經：行，為也。
說：行。所為不善名，行也。所為善名，巧也。若為盜。
1. In the shūo, several editors propose to emend the graph 善. 伍非百 emends 善 to 義 and transposes the negative 不 to the second sentence, reading 所為(不)(善)[義]，名行也；所為[不](善)[義]，名巧也，若為盜。高亨 emends 善 to 差 on grounds of their graphic similarity, reading 所為不差名，行也。所為差名，巧也. We consider both of these unjustified, conjectural emendations.
2. Sun, followed by Liu, proposes to emend 巧 to 竊, based on graphic similarity between 巧 and a variant of 竊 and on the semantic relation between 竊 and 盜 in the final sentence. Again, since 巧 is intelligible, we reject this as a conjectural emendation.
3. The shūo is open to several interpretations. We suggest that 為 in the first two sentences be read with the falling tone, wèi, yield the following interpretation: “If what a person acts for (the person’s motivation) does not lie just in valuing a name, that is ‘conduct’. If the motivation lies in valuing a name (he’s seeking the name or title), that is ‘trickery’. In that case, the person is no different from a robber, since, because of this motivation, the name does not fit the actual nature of the behavior.” Against this, contrast Graham’s translation: “What one does, neutrally named, is one’s conduct. Giving a good name to what one does is trickery. (For example, to committing robbery).”
A11 經：實，榮也。
說：實。其志氣之見也，使人(如)[知]己。不若金聲玉服。
1. On the basis of research by 欒調甫, Graham proposes to emend 實 to 誠, on the grounds that the text originally read 誠 but was changed to 實 to avoid a Song dynasty taboo graph. 欒 (155) found six sentences in the incomplete 俞 manuscript of Mozi that read 誠 where the corresponding graph in the Daozang edition reads 實. Also, Wu Yujiang (56) has found a quotation from 〈辭過篇〉of the Mozi in the Tang dynasty 《群書治要》 that reads 誠 where the Daozang edition has 實. We believe these findings provide strong evidence that instances of 實 in contexts concerning ethics (as opposed to epistemology or semantic theory) may well have originally read 誠. However, this textual evidence is all highly indirect; the 俞 manuscript includes only pian 1-13 of the Mozi, and not the Jīng and Shūo, and the graph 實 is intelligible as it stands in received text. Hence any emendation would still be too speculative to be justified. Thus we do not emend the text, but only note the possibility that 實 originally read 誠.
2. Several editors (Gao, Graham, Zhou) emend 使人如己 to 使人知己. Others (Liang, Chen, Yang (123)) emend to 使之如己. Both emendations rest on interpretive grounds, namely the claim that the received text is unintelligible. We believe that emending 人 to 之 is unjustified, partly because the graphs are not especially similar, and partly because 使之如己 does not really improve the intelligibility of the sentence. (The antecedent of 之 is unclear, for instance.)
The received text is indeed difficult to interpret. It could refer to one’s sincerity having a transformative effect on others, and thus “making others similar to oneself.” But the theme of the passage seems to be the outward manifestation of inward attitudes or qualities, which seems unrelated to transforming others and can plausibly be linked to enabling others to know one. Hence we agree that emending 如 to 知 significantly improves the intelligibility of the text. However, since this emendation is based purely on interpretive grounds, we believe it carries only a low to medium degree of justification.
3. A number of editors believe that the negative 不 in 不若金聲玉服 is unintelligible and either delete it (Gao, Liu, Zhou) or emend it to 必 (Sun). In addition, Liu proposes emending 服 to 振. We believe both proposals are unnecessary and the text is intelligible as it stands. The point is that sincerity involves the manifestation of one’s inward intention and qi, and this is unlike the beautiful appearance of the “gold sounds and jade clothes,” which may be beautiful though hollow.
A12 經：忠，以為利而強(低)[?]也。
說：忠。(不利弱子亥足將入止容)[…?…]。
1. Most editors hold that the graph 低 in Jīng A12 is unintelligible in this context and should be emended. However, none of the arguments for emendation seem persuasive.
Sun (331), followed by many editors, emends to 君. He argues that 君 could first have been corrupted to 氐, which is graphically similar in seal script, and then 氐 could have been corrupted to 低. His interpretation of the canon is thus that loyalty is something that benefits one’s sovereign. In our view, both this interpretation and Sun’s hypothesis about how the graph became corrupted are too speculative to be justified.
伍非百 emends 低 to 聒. It is difficult to see how the posited corruption could have occurred, since the graphs in question are not similar, and the interpretive grounds for this emendations is unconvincing. 吳毓江 (491) emends to 民 or 氓. In this case, there is perhaps a graphic similarity between 氐 and 民, but the graph in the text is 低, not 氐. Moreover, the resulting interpretation—roughly, “Loyalty is strengthening the people when one takes something to be beneficial”—does not seem convincing.
Graham emends to 任, based on the similarity of 任 and 
,
a graph for 低 used in Han inscriptions. Presumably his main motivation is interpretive.
Emending to 任 yields an intelligible reading that coheres well with the explication
of 任 in Jīng A19. The hypothesis that 低 is an error for 任 is thus attractive.
But the scenario Graham posits is implausible. In li shu 隸書, rèn
任 is written roughly as it is today, 任, or sometimes 仼. Graham’s proposal
thus entails that a scribe accidentally corrupted the relatively simple original
graph 任, which was part of an easily intelligible sentence, by adding strokes
to form a more complex graph resembling ,
which was then interpreted by a later scribe as 低, rendering the sentence unintelligible.
We reject this proposal on the grounds that it is easy to explain how a scribe
might corrupt a complex graph to a simple one or a difficult sentence to an
easily intelligible one, but difficult to see how the direction of corruption
could run the other way. Indeed, given the considerations Graham cites, the
principle of preferring the more difficult reading would suggest that 低 is the
correct reading, not a corrupt graph.
Several interpretive proposals accept the received text. 畢沅 (cf. Sun 311) interprets the canon as 「以利人為志而能自下」. But it is difficult to see how this interpretation relates to loyalty. 張惠言 interprets 低 as 氐, meaning 'root'; again, it is difficult to see a connection to loyalty. 劉師培 suggests that 低 be read as 抵. This proposal faces two problems. The first is that, as Graham points out (274), all of the other canons in this sequence seem to explicate or give an equivalent for the first term in the canon. On this proposal, the canon would be interpreted roughly as “Loyalty is strongly resisting when one takes something to be beneficial.” It is difficult to see how this could be an explication of loyalty, rather than, as Graham says, an observation about people who happen to be loyal. A further problem is that it is unclear whether there are philological grounds for reading 低 as 抵.
We conclude that we are moderately justified in believing that 低 is an error, but that none of the proposals for emendation are justified. We therefore consider the proper reading of this canon an open question.
2. Most editors agree that the shūo for this canon is unintelligible as it stands. Liang (13), Graham, and Wu Yujiang conclude that the shūo is too corrupt to be reconstructed.
Tan (91-92) proposes a speculative emendation related to the sage king Yu’s flood control work 大禹治水. We reject this on the grounds that there is no reason to think the text has anything to do with the sage king Yu.
Gao, followed by many editors, emends as follows: 不利、弱(子亥)[孩]足將入[井](止)[之]容. Roughly, this yields the interpretation that if the ruler does harm, and the danger imposed is like a child about to fall into a well, it is loyal to stop him. Aside from being highly conjectural, we believe this proposal is implausible for two reasons. First, it is implausible that a scribe writing vertically from top to bottom could accidentally separate the single graph 孩 into two graphs, 子 and 亥. Second, we believe that the proposed emendations do not significantly improve the intelligibility of the text.
As Liang first proposed, the shūo seems to present a contrast of some sort with the jīng. The jīng says that loyalty is a matter of doing some sort of action when we take something to be beneficial; the shūo perhaps says that when we take something to be not beneficial, we should prevent or avoid some action. But even this limited interpretation of the shūo is speculative, and any interpretation or emendation beyond this is highly conjectural. We conclude that without further information, no emendation or interpretation of the shūo can be justified, and the correct reading of the text must remain an open question.
A13 經：孝，利親也。
說：孝。以親為芬而能能利親。不必得。
1. As in Shūo A8, some editors propose to emend 芬. Again, we believe that, understood as a variant of 份, this graph is intelligible without emendation.
2. In the shūo, Chen and Zhou propose to insert 利 before the first instance of 親, on the grounds of parallelism with the jīng. This proposal seems unjustified, since the sentence is intelligible without emendation.
3. On the basis of parallelism with A8, Graham inserts 志 at the head of the first sentence of the shūo. This emendation is attractive, since A8 and A13 are indeed highly parallel. But we believe the text is intelligible as it stands, and inserting 志 does not significantly improve intelligibility. Hence the best explanation of the text is that it is correct as it stands.
A14 經：信，言合於意也。
說：信。不以其言之當也。使人視城得金。
1. In the shūo, several editors (Sun, Liang, Wu Feibai) emend 不 to 必 on interpretive grounds. Since the received text is intelligible, we reject this emendation as speculative.
2. A few editors find the second part of the shūo problematic. Graham takes the shūo to be fragmentary. He posits a lacuna between 使人 and 城得金 and transposes into it a fragment from A29 to complete what he takes to be the truncated first sentence. The sentence ending in 城得金 he believes cannot be reconstructed, though he speculates that it may have been 不若超城得金, on grounds of possible similarity to remarks about “going over a wall” in A75 and A93. The grounds for the proposed transposition are (i) parallelism with A11, which reads …使人(如)[>知]己; (ii) Shūo A29 is unintelligible in its context; and (iii) interpretively, the meaning of the received text is odd, “Send a man to look at a city wall and get gold.”
While agreeing that the received text is puzzling and could be faulty, we believe Graham’s proposal raises more questions than it solves. Shūo A14 and A11 are not strongly parallel, so parallelism between them does not provide grounds for emending A14. A29 is indeed problematic, as it falls in the middle of a long, damaged stretch of text. Yet it refers to 行 ‘conduct’ and appears as the shūo of a jīng dealing with praise, so in fact it does seem relevant to the context of Jīng A29. The received text of shūo A14 is indeed difficult and perhaps faulty, but Graham’s emendation does not really render the text any less obscure and indeed may be grammatically questionable. (His reconstructed first sentence is 不以其言之當也使人視之，必其行也; a more natural way to say this would probably be: 使人視之，不以其言之當也，而必其行.) Most important, however, it is difficult to see how the dislocation Graham posits could have occurred. Some bits of text in the long, damaged section of text between shūo A22 and A39 have surely been accidentally transposed. But these obvious transpositions are from one shūo to others very near by. Graham’s proposal for A11 involves transposing half of a sentence from the middle of the damaged section around A29 into a different part of the text that does not seem obviously damaged. Moreover, the shūo to A11 is separated from A29 by many undamaged shūo. So we reject Graham’s emendation on the grounds that it does not yield an appreciably more intelligible reading and that there is no good explanation of how the fragment of text he transposes was originally dislocated from A11 to A29.
3. Gao and Zhou emend 得 to
on interpretive grounds. We believe 得 is intelligible and so reject this proposal
as conjectural. Liang emends 城 to 誠. Again, we do not see that this enhances
intelligibility.
A15 經：佴[=恥]，自作[=怍]也。
說：佴。與人遇，人眾， 。
This jīng and shūo present difficult problems to which we have no compelling solution. The meaning of the jīng is obscure, and the identity of the graph 佴 is unclear: 佴 is a known graph, but it typically means “duplicate,” which has nothing to do with the context. Moreover, 佴 also appears as the first graph in Jīng A71, suggesting that either A15, A71, or both may be faulty. The shūo contains an unkown graph , and parsing and interpretation of the shūo are difficult.
1. Several editors offer proposals for emending 佴. Liang emends to 侔, Tan to 狂, and 伍非百 to . Without going into detail, we believe that all these emendations are purely conjectural and unconvincing. Graham emends to on grounds of a probably thematic relation to A17, where appears, and the hypothesis that 佴 must be faulty, since it duplicates A71. He interprets as 司. Though tied more closely to the text than the preceding emendations, this proposal we believe is also too conjectural to convince.
Other editors attempt to interpret 佴 without emending, typically by extending the meaning of “duplicate” or by drawing on the 《說文解字》, which glosses 佴 as 佽, supposedly meaning “to benefit.” This is a possible line of interpretation, but on this reading, it is difficult to see how 佴 coheres with the text of the jīng, which appears to read “one does things oneself.” Moreover, it is difficult to see how A15 relates to A16-18.
The most compelling proposal for dealing with this graph is that of 裘錫圭, who suggests that 佴 is a variant of 恥, since 佴 as 恥 is found in the 馬王堆帛書 and 銀雀山漢簡. This philological evidence is persuasive, and since jīng surrounding A15 deal with trustworthiness and honesty, the proposal that A15 is about shame is plausible. Hence we do not emend 佴, but read it as a variant of 恥 ‘shame’. However, we consider the degree of justification for this reading to be only low to moderate.
2. Even if 裘錫圭 has identified 佴 correctly, interpretation of the jīng remains a problem. The jīng appears to say, “Shame is initiating something oneself,” which cannot be correct. Here we should observe, following several editors, that the graph 作 appears repeatedly in the series A15-18. Taking his cue from A18, on 令 ‘commanding’, Graham suggests that the theme of the series is leaders and followers. Gao, on the other hand, suggests that, like the jīng preceding and following them, this series of four jīng is probably about virtues or character traits.
Gao suggests a reading of 作 that coheres well with our reading of 佴 and with the hypothesis that these jīng are about virtues. He proposes that we read 作 as 怍 ‘shame’. One advantage of this proposal, as we will discuss as we treat A16-18, is that it provides a relatively coherent interpretation of all four jīng. Another is that it requires no emendation, since we can probably take 作 as an alternate form of 怍, because the later Mohist texts are recognized as having a wide degree of variation in the graphs used to write common words. However, we emphasize that this solution is partly conjectural, so the degree of justification for it is only low to moderate. The correct reading of A15 should be considered an open question.
Our tentative reading of jīng A15, then, is that the text is semantically equivalent to 恥，自怍也, or roughly, “Shame is taking oneself to be disgraceful.”
3. Interpretation of the shūo is even more difficult. There are two main problems: identifying the graph and determining how to parse the text. We begin with the latter. There are at least four ways to parse the shūo, as follows (‘x’ in our rough translations represents the unknown graph):
與人｜遇人｜眾 (Being together with people, if one meets people, they all are/do x)
與人｜遇人眾｜ (Being together with people, if one meets numerous people, one is/does x)
與人遇｜人眾 (Meeting with people, they all are/do x)
與人遇｜人眾｜ (Meeting with people, if they are numerous, one is/does x)
The difficulty of settling on a parsing is compounded by the obscurity of the thematic relationship between the shūo and the jīng. The shūo appears to be about meeting people and having some sort of response, expressed by the unknown graph, and it is difficult to see what this has to do with shame (or, for that matter, with any other interpretation proposed for 佴).
Among the readings proposed for are 循 ‘follow’, adopted by 吳、Graham and 姜; 遁 ‘retreat’, adopted by 譚 and 周; and 揗 ‘rub’, adopted by 孫. Taking their cue from the heart radical used in the unknown graph, 譚 and 裘 suggest that it may represent an attitude closely related to 遁 ‘retreat’, roughly the intention to retreat, or perhaps shyness or avoidance. Conceivably, an attitude akin to shyness can be thematically linked to shame, so this interpretation may cohere with the interpretation of 佴 as ‘shame’. However, the rival interpretation of 佴 as 佽, or “benefit,” might cohere just as well with an interpretation of as “to follow.” Someone who benefits others might thereby cause them to follow him.
Here we accept the 譚/裘 interpretation of , on the grounds that it coheres better with our interpretation of 佴 as 恥, which in turn coheres better with the themes of the surrounding jīng. Clearly, however, these grounds are quite weak, and our reading of both and the entire shūo at best has only a low degree of justification.
Given this interpretation of , it seems more likely that the correct parsing is that upon meeting people, the person who feels shame would respond by intending to retreat, not that numerous people would retreat. Thus we eliminate the first and third of the four parsings. Of the remaining two, the fourth seems more natural. On the second, for instance, the phrase 與人 seems redundant. Hence we adopt the fourth and read the shūo roughly as, “Encountering others, if there are many, one retreats (feels inhibited).” This interpretation is intelligible but puzzling. It is not obvious how this is an appropriate way to explicate or expand on the notion presented in the jīng. Moreover, the jīng seems to be about a moral notion, but the shūo is not obviously so. Both our reading of and our punctuation of the shūo are partly speculative, and the result is not completely satisfactory. Hence we emphasize that our treatment of the shūo is only weakly justified.
A16 經：
[=狷?]，作嗛[=慊?]也。
說：[=狷?]，為是(為是)之台[=怡?]彼也弗為也。
This is another difficult case. The jīng uses an unknown graph, , and the readings of two other graphs, 嗛 in the jīng and 台 in the shūo, are uncertain.
1. Most editors follow Sun in taking the unknown graph to be 狷 ‘uncompromising, unyielding, scrupulous’, contrasted in Analects 13:21 with 狂 ‘fervent, zealous’. This reading is certainly plausible; insofar as it takes the jīng to be about a character trait, it fits the context, and the Mozi is known for using a language radical in some graphs (e.g., 請 for 情). Moreover, 狷 coheres reasonably well with plausible readings of the shūo. But, partly because the identity of 嗛 is unclear, Sun’s proposal is not fully convincing. We follow it not because it is well justified, but because we lack any better alternative. We consider this reading to have a low degree of justification, and the correct reading should be considered an open question.
Several readings of 嗛 have been proposed, but in our view none of them are convincing. A few editors take 嗛 to be either a variant of (吳汝綸) or an error for (Gao, 40) 兼 ‘jointly, inclusively’. Gao takes this to be a reference to the Mohist doctrine of 兼愛 ‘all-inclusive concern’. Drawing on Lunyu 13:21, 狷者有所不為也 (“…the scrupulous have that which they do not do”), he seems to be suggesting that the scrupulous person does not practice all-inclusive concern, and this is shameful in the eyes of the Mohists. But whether we interpret 作 as 作 ‘do’ or as 怍 ‘ashamed’, it is difficult to see how reading 兼 for 嗛 makes sense of the text. No variation of “Being scrupulous is undertaking [or being ashamed of] joining together with others [or all-inclusive concern]” seems a likely explication of 狷—being an uncompromising person does not suggest that that one either is or is not a nonconformist, or is or is not devoted to all-inclusive ethical concern.
裘錫圭 proposes to read 嗛 as 謙 ‘modest, unassuming’. But “being scrupulous/uncompromising is undertaking to be humble” is hardly an intelligible explication of 狷. 裘錫圭 also tentatively emends 作 to 非. This yields the interpretation, “Scrupulousness is being not modest,” which in our view does not improve the intelligibility of the text.
Following Sun (314), some editors (e.g., Jiang) propose to read 嗛 as 慊 ‘content, satisfied’. This gives, roughly, “Being scrupulous is undertaking to be satisfied,” which may be a plausible interpretation if we think of the scrupulous, uncompromising person as seeking to satisfy himself by conforming to his standards no matter what others think or do. (This is roughly Jiang’s interpretation.) A further possibility is to follow Sun in reading 嗛 as 慊 and to read 作 as 怍, as we do in A15. The result is, roughly, “Being scrupulous/uncompromising is being ashamed to be satisfied/content,” which is plausible, perhaps, but interpretively not especially compelling.
A final possibility that we can think of is to read 嗛 as 廉 ‘honest’, giving us, “Being scrupulous/uncompromising is undertaking to be honest.” However, we find this unconvincing, because 廉 is written as 廉 just two graphs later. This proposal also does not really make sense of 作, and it bears no obvious relation to the text of the shūo.
We believe that the least implausible of this range of proposals is to take 嗛 as 慊, yielding, roughly, “Being scrupulous/uncompromising is undertaking to be satisfied/content.” However, this interpretation is speculative and has no clear relationship to the shūo. Hence we think it fails to reach even a low degree of justification. We conclude that the reading of this jīng is an open question.
3. The shūo begins by repeating the words 為是 twice. Editors adopt two main approaches for explaining the duplication. One is to accept the received text, as Graham does, and punctuate after the first 是, yielding, roughly, “He does this, but when doing this….” The other approach is to omit one instance of 為是 on grounds of dittography. The advantage of the first approach is that it modifies the text as little as possible. Arguably, the received text is intelligible, and we obtain a simpler explanation of the state of the text by accepting the text as it stands. The grounds for deleting the duplicated phrase 為是 are mainly interpretive, and in the absence of a highly justified interpretation of the jīng, it may be hard to see how we can justify deleting one of the duplicates.
Against this line of reasoning, however, we argue that probably the redundant 為是 does not make much sense on any reasonable interpretation of the text. Since the repeated phrase comes at the head of the shūo, it seems open to only one interpretation -- “One does this, [but when] one does this….” -- which renders the text intelligible only if we take the text to imply the phrase we insert in brackets. We believe that on any reasonable interpretation, one of the pair of phrases is strictly redundant. Hence we omit the second on grounds of dittography.
4. The shūo reads 為是之台彼也弗為也. The graph 台 ‘platform’ cannot be correct. Some editions have the variant 治 ‘govern’ for 台, which is adopted by 伍非百. However, these editions do not have independent textual authority, and 治 does not really make sense of the sentence. Following 顧千里, Sun and several other editors read 台 as 詒 ‘sayings’ or ‘hand down’. Sun’s interpretation (336) is 狷者不為欺人之言, roughly, ‘a scrupulous person does not engage in deceptive talk’. We do not see how this yields an intelligible explanation of the shūo.
于省吾 and 裘錫圭 read 台 here as 以. We do not believe that 台 is likely as an error for 以. Also, the grammar of the resulting reading 為是之以彼 is extremely awkward.
Gao and Graham emend 台 to 害, giving, “when doing this [or, in Graham’s case, doing the right thing] interferes with some other thing, he will not do it.” Interpretively, this reading is plausible, but we find it difficult to see how 害 could be corrupted to 台 as graphically the two are not that similar, and the latter is much less intelligible in this context.
Some editors (Chen, Tan (96)) read 台 as 怡, yielding the reading, “Doing this to please others, he will not do it.” 怡 is found written as 台 at least once in the 史記. This interpretation coheres with one common understanding of 狷者, as a person who stubbornly upholds his own principles without regard for others’ likes and dislikes.
The most attractive of these explanations are the readings as 害 and as 怡. Of these, 害 probably provides the more easily intelligible interpretation, but 怡 provides a far simpler explanation of the graph 台. So, without emending the text, we tentatively adopt the hypothesis that 台 should be read as 怡. The meaning of this jīng and shūo text are so uncertain, however, that we consider this reading to have only a low degree of justification.
5. A further question is why the shūo contains two instances of the particle 也, which seem grammatically redundant. We leave this as an open question.
經：廉，作[=怍]非也。
說：廉。己惟[=雖]為之，知其(也) 
[=恥]也。
There are four main problems in treating this jīng and shūo. First, the jīng seems unintelligible as it stands. The received text reads, “Honesty is doing wrong,” which cannot be correct. Second, the shūo contains an unknown graph, . Third, the grammar of the last part of the shūo seems unintelligible: the first instance of the graph 也 is either misplaced or is a graphic error. Fourth, the shūo in the received text ends with the graph 所. We have already argued in our treatment of the textual divisions that this graph is part of A18, so we will not discuss this problem here.
1. Our own solution to the first problem follows directly from our solution to A15. We there argued that a plausible reading of one or more instances of 作 in this group of jīng is as 怍 ‘ashamed’ or ‘disgraced’. Adopting this reading for A17, we get the rough interpretation “Honesty is being ashamed of [what’s] wrong,” a plausible reading of an explication of 廉. (Most of the editors in our sample read 作 as 怍 here, the exceptions being Graham, who is uncertain how to handle this jīng; Chen (218) who reads it as 詛, though without giving a justification; and Jiang (28), who seems to interpret 作非 as doing something wrong (作) and then feeling remorseful (非). Jiang’s interpretation is not that different from our own, but as an interpretation of the unemended text, we believe it is highly forced.)
Another possible interpretation of the jīng is Gao’s (41). He accepts the received text and explains it by suggesting that having done something wrong, an honest person will feel guilty and deem his own action wrong. Thus, “honesty is doing [something] and [deeming it] wrong.” We believe this explanation is too forced; were this really the point of the text, we would expect it to say so explicitly, rather than reading simply 廉，作非也.
裘錫圭 suggests that 作非 should be 非作 (reading 作 as 怍), based on parallelism with A16, 佴[=恥]，自作也. This is plausible, but we do not see that it improves intelligibility over our own, simpler proposal.
Graham tentatively proposes to emend the text to 為非所作也 (277), but he himself considers this reading so speculative that he does not incorporate it into his translation.
In an attempt to interpret A16-A17 more coherently as a pair, Sun and Liang emend 廉 here to 慊, so that it expresses the same word as their reading of 嗛 in A16. (Jiang follows them in reading 廉 as 慊, though he does not emend.) We resist this proposal on the grounds that it does not seem to enhance overall intelligibility. Jīng 11-16 appear to be about various virtues, and the received reading 廉 ‘honesty’ seems to fit this context better than the emended reading 慊 ‘satisfied’. Moreover, though “Being satisfied is being ashamed of [what’s] wrong” is intelligible, it seems a less natural explication than “Being honest is being ashamed of [what’s] wrong.”
We conclude that the jīng should probably be read 廉，作[=怍]非也. But because many readings in this series of jīng are uncertain, and because our grounds for reading作 as 怍 are interpretive only, we consider this conclusion to have only a low degree of justification.
2. In our view, 裘錫圭 has persuasively identified the graph as a variant of 恥. This reading was first proposed by 高亨 (41), followed by
Chen and Zhou. As 裘錫圭 (1995) explains, this hypothesis is supported by archaeological
findings. In 馬王堆帛書《戰國縱橫家書》, the graph 恥 is written as 左耳又鬼 or as 餌. 裘錫圭 points
out that in very early manuscripts, it is common for radicals in 形聲字 to be
switched from left to right and vice versa. He also points out that in the
陸 edition of Mozi, is written as 
, which is likely to be a variant of the same graph as 左耳又鬼. Thus it is possible
that the 思 radical in the Daozang edition is either a variant of, or an error
for, 鬼, since the two are graphically similar. This suggests that expresses the same word as 左耳又鬼 in 《戰國縱橫家書》, namely the word commonly written
in modern script as 恥.
Other proposals include the following. Graham reads as 司, on the grounds that an interchange of 司 and 思 is attested (277). However, this emendation is almost purely conjectural, and Graham himself indicates that his reading of the text is speculative (278).
Sun’s proposal to read as 諰 ‘apprehensive, anxious’ is followed by several editors, including Liang, Tan, and Jiang (though the latter two do not emend the text, instead taking to be a variant for 諰). Tan (97) points out that 諰 has many variant forms, including 葸、偲、鰓 and 左狗又思. So it is possible that is just one of these several variants. The meaning of 諰 could conceivably fit the context here, but we believe that the 高/裘 hypothesis provides a better explanation of the text, in terms of both the provenance of the graph (from 左耳又鬼) and the interpretation of the text (shame seems more closely related to honesty than anxiety is).
A final proposal for treating this graph, adopted by Wu Yujiang, is to emend by dividing it into two graphs, 思 followed by 耳. In fact, several early editions of Mozi do divide the graph this way (郎氏堂策檻本、顧校李本、四庫全書文津閣本). However, none of these editions is as early as the Daozang or has independent textual authority; so these readings are probably later editorial emendations. This proposal would rest on a hypothetical scenario roughly as follows. After copying the text, a scribe might have noticed that he had omitted the graph 耳 and then inserted it by writing it to the right of 思. Then a later copyist accidentally combined the two graphs. However, we find this scenario highly unlikely. No Chinese reader would accidentally take 思 and 耳 to be a single graph, since no such graph is known. Moreover, on this hypothesis, the sentence ends in two particles 耳也, making it grammatically anomalous. Also, the reading that follows from dividing into two graphs (“know its thought”) is much less intelligible than that which follows from taking the graph to be 恥.
We conclude that is an alternate way of writing 恥 and requires no emendation. We consider this conclusion moderately well justified.
3. Most editors follow Sun in omitting the first instance of 也, which appears to be grammatically redundant. The occurrence of the graph can be explained by dittography from the 也 in the preceding shūo or the one at the end of this shūo.
Another possible explanation is that 也 here is a graphic error for another word. Tan (97) suggests that it is an error for 心, which in seal script is indeed graphically similar. Interpretively, however, there are two problems with this hypothesis. First, the Canons generally do not use the word 心 in discussing psychology. Second, and more important, though it is certainly possible that 心 is the original reading, it does not seem to fit the context well. On our interpretation, the sentence reads, roughly, “Although he does it, he knows it’s shameful.” The pronoun 其 appears to share the same antecedent as 之, namely the wrong action denoted in the jīng by the word 非. This reading of the shūo coheres well with our reading of the jīng as “Honesty/integrity is being ashamed of [what’s] wrong.” On the other hand, if 也 is read as 心, then the antecedent of 其 must be the agent, and the sentence becomes, roughly, “Although he does it, he knows his heart’s shame.” Though this reading is plausible, it seems less coherent with the jīng, and it is unclear why the text would refer to knowing, rather than simply feeling, one’s own shame. We conclude that the better explanation is that 也 is redundant, and we follow Sun in omitting it. We take this conclusion to be moderately justified.
It is puzzling why the shūo appears to link honesty/integrity to shame about unethical behavior rather than to reliable conformity to ethical norms. To resolve this puzzle, 裘錫圭 suggests that 己惟為之 “although one does it” may originally have read 己不為之 “One does not do it.” However, this is a conjectural emendation, and it is difficult to see how a scribe could corrupt 不 to 惟. Hence we leave the text unemended, merely pointing out that the shūo presents a minor interpretive puzzle, which could possibly indicate that the text is faulty. With most other editors, we read 惟 as 雖, a reading common in the Mozi.
A18 經：令[=矜]，不為所作[=怍]也。
說：(所)令。[所]非，身弗行。
Interpretation of both the jīng and the shūo is difficult. Moreover, it is not clear whether the word 所 belongs in this shūo or in the preceding one. We have already argued in our discussion of textual divisions that 所 belongs in this shūo, so the following discussion assumes that conclusion.
1. Most editors take 令 to be “to command,” yielding an interpretation roughly like Graham’s, “To command is not to perform what one initiates.” But the meaning of 不為所作也 is obscure, and there are surely more natural ways to express the meaning “to command,” such as 以其言使人為之. Moreover, it is difficult to see how the jīng coheres with the shūo, which appears to read, “What is deemed wrong, one does not do it.”
This leads us to question whether the jīng is in fact about “commanding.” 裘錫圭 proposes to read 令 as 矜, for three reasons. (i) In the《說文解字注》, 段玉裁 mentions that in ancient texts, 矜 was written with 令 instead of the phonetic 今. (ii) In recently unearthed ancient manuscripts, 矜 is written 左矛右令. (iii) 矜 ‘self-discipline’ fits the context. If we read 作 as 怍, as in A15, then reading 矜 for 令 yields “Self-discipline is not doing what’s shameful.” 矜 coheres with the themes of A15-A17, on virtues, and is conceptually related to 廉. However, though this proposal is based on philological evidence, it is motivated by interpretive considerations, and we should emphasize that 裘錫圭 has not cited an actual example of 矜 written as 令 alone, without the 矛 radical.
Against this proposal, one could argue that the theme of commanding coheres with Jīng A19, about bearing responsibility, and that the shūo can be interpreted as “What is not one’s role, one does not do it.” (We argue against this interpretation of the shūo below.)
裘錫圭 also argues that 作 in 不為所作 should be emended to 非, by parallelism with the shūo 所非 and because of the graphic similarity in seal script of 非 and 作. However, we think that reading 作 as 怍 coheres better with the surrounding jīng and provides a simpler explanation of the text.
Our tentative conclusion is to accept the received text, but read the jīng as, “令[=矜]，不為所作[=怍]也 “Self-discipline is not doing what’s shameful.” However, we believe that this reading has at best a low degree of justification and that the reading “To command is not to perform what one initiates” is almost equally plausible. Interpretively, we suggest that our preferred reading coheres better with the surrounding jīng and the shūo. However, our reading depends on the conjecture that 令 can be read as 矜, for which 裘錫圭 has provided a preliminary but not fully convincing argument.
2. Our reading of the shūo is inspired by a sentence in Mozi Book 48: 口言之，身必行之, which leads us to punctuate 所非，身弗行 “What is deemed wrong, he does not do it.” We believe that this interpretation coheres well with, and tends to support, our reading of the jīng.
Another possibility is to read the shūo as 所非身，弗行 “What is not one’s role, one does not do it.” However, on this reading the graph 所 seems redundant, and we question whether 身 is used to mean “role” or “job” in the Mozi and other earlier texts.
Sun and Liang emend 弗 to 所, giving the reading 非身所行 “It is not that one does it oneself.” This coheres well with the reading of 令 as “command,” but we find the proposal that 弗 is an error for 所 too speculative.
A19 經：任，士損己而益所為也。
說：任。為身之所惡以成人之所急。
A20 經：勇，志之所以敢也。
說：勇。以其敢於是也命之，不以其不敢於彼也害之。
A21 經：力，刑之所以奮也。
說：力。重之謂，下與[=舉]重，(舊)[>奮]也。
1. As Sun suggests, the reading of 與 here that seems to fit the context best is 舉 ‘to lift’, since the theme of the jīng is strength. 舉 is written without the radical in at least one other place in the Canons (B5). We do not consider this reading an emendation.
2. Most editors follow Sun in emending 舊 in the shūo to 奮, on the basis of comparison with the jīng. The two graphs are similar, and 舊 is unintelligible without emendation, whereas 奮 yields the plausible interpretation, “Lifting a weight from below is exertion.”
3.Editors parse the shūo in various ways. Wu Yujiang proposes to emend 謂 to 請, equivalent to 情, interpreting the shūo as, “Weight is genuinely below; lifting a weight is exertion.” We do not see that this reading enhances intelligibility. Some editors (Tan, Chen) read, roughly, “It’s said of weight. Below, lifting. It being heavy is exertion.” Sun has, “Of weight, it’s said to be below. Lifting a heavy thing is exertion.” Neither of these seems as clearly intelligible as the reading proposed by Liang and followed by most editors, “It’s said of weight. Lifting a weight from below is exertion.”
4. One editor, Graham, suggests that the first sentence of the shūo is missing a final particle, 也 (147, 279). He calls attention to a similar construction …之謂也 in the 《大取》. We observe that all sentences in the Mozi ending in 之謂 have a final 也, and indeed in a cursory check, we noticed no sentence in the pre-Qin literature that ends in 之謂 without a final particle, either 也, 乎 or 耶. So grammatically, there is good evidence for Graham’s conjecture. However, since the sentence is intelligible without inserting 也, we have left it unemended.
A22 經：生，刑與知處也。
說：生。楹[=?]之生商[=?]不可必也。
The jīng is readily intelligible, reading 形 for 刑, without emendation: “Life is a form (body) located together with knowing (consciousness).” But because of two uncertain graphs, 楹 and 商, the shūo is unintelligible and probably unsalvageable. It marks the beginning of a series of damaged shūo that runs up to A38. We will briefly review several proposals for emending the shūo before concluding that none are convincing.
畢沅 suggested that 楹 be read as 形, and many editors have followed him. Sun further suggests that 商 is an error for 常, because the two were originally homophonous. The reading of 商 as 常 is supported by the 廣雅釋詁, which states:「商常也」 (Gao, 43). These emendations yield the following rough interpretation, “As to the life of the body, its permanence cannot be taken as certain.” As Sun sees it (337), the point is that life is not permanent/constant; when the body and awareness come together, there is life, and when they separate, death (言生無常，形與知合則生，離則死). This interpretation might be plausible if we were confident of the identity of all the graphs. Yet without such confidence, it is puzzling both why only the body, and not the 知 (awareness) would be mentioned in the shūo, and indeed it is puzzling why the text should talk about permanence or constancy at all. Moreover, the emendation of 楹 to 刑/形 strikes us as highly conjectural. The pair are not graphically similar, and as far as we know, interchange between them due to homophony is not attested.
Several editors (Gao, Liu, Zhou, Chen) emend 楹 to 盈 on the basis of the 吳鈔本. Tan too considers 楹 a complex form of 盈 (102). However, the 吳鈔本 regularly omits the radical in many graphs, so the variant 盈 there does not necessarily attest that 盈 is the original reading. We find it difficult to see how reading 楹 as 盈 renders the shūo intelligible: “Filling it, there is life. It is constant, but cannot be taken as necessary.” Gao’s argument for this interpretation (43) is that since jīng A66 states that 盈，莫不有也 ‘none are absent’, the word 盈 here can be understood as implying that neither form nor awareness is absent, and thus there is life (有形體，有知覺，兩者莫不有). So he interprets, roughly, “Filling it [with body and conciousness], there is life.” The second sentence of the shūo he sees as implying that events in life are not necessary, and thus alluding to the Mohist doctrine of 非命, “Rejecting fatalism.” Gao’s interpretation might be feasible if we were certain of the identity of the graphs in the text. But given our uncertainty about these graphs, his rendering is too speculative to be justified.
姜寶昌 (38) takes 商 to refer to merchants, suggesting that the point of the shūo is that life is as undetermined as one’s gain or loss in business. Tan and 楊俊光 (203) interpret 商 as 強 ‘strong, robust’. (On the authority of the《說文解字》, Tan takes 商 to be equivalent to 章, which he suggests has the extended meaning of “strong and flourishing.”) This reading seems speculative, and it does not obviously improve the intelligibility of the text.
Graham contends that the shūo does not fit the jīng and transposes it to Shūo A39, at the end of the series of damaged shūo, where he suggests it fits well. We find this transposition unconvincing but defer discussion of it until A39.
The speculative nature of all of the readings for 楹 and 商 and the forced, unconvincing interpretations they produce testify to the difficulty of the text here. We believe none of the proposals can claim even a low degree of justification. We conclude that we are highly justified in believing that the text of the shūo is faulty, but the correct reading must be left an open question.