Throughout our work we have largely followed A. C. Graham (1978) in recognizing that both the jing ‘Canons’ and jing shuo ‘Explanations’ of the Mobian text divide themselves into two major categories by type of passage, (a) sections that can be identified as ‘definitions’ (A 1 through A 87) and (b) sections that can be identified as ‘propositions’ (A 88 through B 82). The Mohist text itself does not use any kind of meta-language to refer to these two different types of section, but the distinction becomes apparent from the structural features of each. As will have become clear from the preceding material, the ‘definitions’ characteristically end with a sentence final yě 也, as is befitting a proper definition, and the ‘propositions’ typically end with a sentence of the form shuō zài X (說在 X) “the explanation lies with X” where X is a single word or a short phrase that seems intended somehow to epitomize the sense of the Canon, though the intended explanation is not always obvious. The last 12 of Graham’s ‘definitions’ sections, A 76 through A 87, do not conform to the wording structure or content type of the rest of the definitions; in particular they tend not to be in full-sentence form and there is no sentence final yě 也. Nor are they propositions in any obvious sense, and they do not end with the shuō zài X phrase expected of Mobian propositions.Footnote 1 Graham includes them among the ‘definition’ sections, but says of them that they are “words with more than one usage” and refers to them generally as ‘ambiguous’ (Graham 1978, 323). Christoph Harbsmeier, by contrast, prefers to regard these sections as the first 12 sections of the ‘propositions’ (Harbsmeier and Needham 1998, 328, fn 2).

Whether one follows Graham’s view that these are definitions of ambiguous words or Harbsmeier’s that they are propositions, equally ambiguous, the 12 sections are anomalous with respect to the rest of the Mohist Canon, and that anomaly has something to do with their lexical nature. In view of this we examine these sections here as a group, which we call the “Lexical Appendix.” Graham’s characterization of them as ‘ambiguous’ is perhaps a little less specific than their content would call for.Footnote 2 The Canon of each typically gives one or more glosses for the entry word. These glosses vary sometimes in their meanings in ways that we might not immediately see as linking them semantically to the entry word. The Explanations usually provide contexts and specific examples for the glosses that show what the intended inter-connections are and allow us to see how the semantic scope of the entry word might be conceived. In several of the sections we find glosses and explanations that seem to distinguish the use of words when referring to a kind of language behavior from their use in referring to real-world contexts. In other cases it looks like the Canon or the Explanation, or both in tandem, are presented in a way to draw attention to an apparent semantic “polarity”, i.e., oppositional senses, inherent in a single word’s usage. These aspects of the Mohist Canon taken together with those parts of the text that Graham has identified or reconstructed as “Names and Objects” (which we do not deal with here) would suggest a conscious interest in language as a part of the Mohists’ larger concern with real-world knowledge.

Harbsmeier has said that the “Later Mohists did not show a special interest in the meanings of words as such, but primarily in the relation between words and objects” (Harbsmeier and Needham 1998, 329). It is not clear what kind of qualification Harbsmeier’s “as such” brings to the statement, but it seems unmistakeable that the Mohist was not only interested in the relation between words and objects, but also that between speech acts and real-world acts.Footnote 3 This entails recognizing a distinction between using words as linguistic acts of some kind and using them to refer to objective, non-linguistic things. Both of these involve a concern with ‘words’ and ‘language’ in a way seen only in one other pre-Han text, viz., chapter 22 of the received text of the Xunzi 荀子, titled “Zheng-ming 正名.” This Xunzi chapter is certainly the primary pre-Han text dealing with the question of “what is the right use of names?”Footnote 4 Graham says that this chapter “contains very little not in the Mohist summa” (Graham 1978, 39). But the “Mohist summa,” i.e., the Mobian, conversely, contains a great deal not in the Xunzi, “Zheng-ming” chapter. This is of course because the latter is a text centered exclusively on the right use of names, chiefly, as Christoph Harbsmeier has pointed out, as a matter of proper social prescription and behavior (1998, 322), whereas the Mohist Canon, in spite of its scattered implications on the difference between the use of words as speech acts and their use to refer to external phenomena, is clearly not primarily centered on either words or language usage. This is what makes the linguistic aspects of the sections that constitute what we have called a “Lexical Appendix” (as well as of sections elsewhere in the Mobian and of Graham’s reconstructed “Names and Objects”) so interesting; they reveal a kind of awareness that language functions as a representation of knowledge, where knowledge is as much of the physical world as it is of the logical and lexical.

FormalPara A 76
C::

已。成, 亡。

E::

已:為衣, 成也。治病, 亡也。

C::

yǐ ‘coming to an end’: may be ‘completing’, may be ‘getting rid of’.

E: :

yǐ ‘coming to an end’: in making a garment one ‘completes it’, in curing an illness, one ‘gets rid of it’.

Graham (1978, 197 and 323–24) points out that yǐ 已 ‘ending, coming to a stop’ is the only word that is used in the Mohist Canon for ‘end, stop’. The common word zhōng 終 ‘to come to an end’ does not occur in these dialectical parts of the Mozi. Graham goes on to suggest that this may have something to do with how the Mohists think of time. In fact, because of its likely etymological affinity with dōng < *ttung 冬 ‘winter’, the primary sense of the word zhōng < *tung 終 ‘end’ is probably ‘cyclical end’ (though the presence of the “silk” classifier suggests rather a linear end). This distinguishes it semantically from yǐ 已 ‘come to an end’ which, as A 76 here clearly shows, has the sense of a completion that entails a change from one state to another (e.g., as the text says, ‘completing an unfinished garment’, ‘curing an illness’). In this connection it is likely that yǐ 已 ‘end’ is the underlying lexical form of homophonous yǐ 矣, the sentence final particle that marks perfective/change of state aspect in the classical language. In other words the perfective aspect marker yǐ 矣 is a grammaticalized form of the verb yǐ 已 ‘end, complete’. This sense accounts, as Graham points out, also for the adverbial use of yǐ 已 as ‘already’.

The Canon is not simply pointing out an ambiguity or semantic breadth associated with the word yǐ 已 ‘coming to an end’, but shows what are actually opposite senses of the word; ‘to bring to an end’ is to ‘terminate’, i.e., to “get rid of” whereas to ‘complete’ is in one sense precisely the opposite, and the point of the section may be to draw attention to that sense and that fact.

Compare A 51.

FormalPara A 77
C::

使。謂, 故。

E::

使:令謂a也。不必成。濕, 故也。必待所為之成也。

C::

shǐ ‘causing’: may be ‘calling’, may be ‘a basis’.

E: :

shǐ ‘causing’: ordering something is ‘calling something’; it may not inevitably be completed; being moist is a ‘basis’; one inevitably expects the completion of what it brings about.

  1. (a)

    R: 謂謂, emendation Graham 1978, 297.

Understanding 故 gù in C as ‘basis’ > ‘providing a basis for’ > ‘cause’ is consistent with A 1; note in particular the Shuowen definition for gù: 使為之也 ‘to cause something to bring some other thing about’. Duan Yucai (1735–1815), perhaps the most important of the Qing philologists to study and comment on the Shuowen, says in his note to this Shuowen entry 墨子經上曰故所得而後成也 “The Canon, part A, of the Mozi says: gù ‘basis’ is what must be the case before something will come about,” repeating exactly the text of the Canon of A 1 (SWGL 1329).

The somewhat redundant sense of 使為 is removed by Xu Kai (920–974), the younger of the two Xu brothers who edited the Shuowen text in the Song period. He emends the definition to simply 使之也 ‘to cause something to happen’.

Note that the inevitability mentioned in the Explanation pertains to the expectation, not the result. This is equally applicable to either the ‘minor basis’ or the ‘major basis’ as set out in A 1.

Here, as with A 76 above, the two glosses given in the Canon do not represent simply semantic scope or nuance, but are intended, we suspect, to reflect fundamental differences in meaning and usage of a single word. In this case the glosses are used precisely to distinguish ‘causing’ as a speech act (wèi 謂) from ‘causing’ as a physical phenomenon (gù 故), an example of the recognition of an intrinsic distinction between linguistic behaviors and external physical events. Beyond this, as set out in the Explanation of these two glosses, the opposition between causation ‘not inevitably’ leading to something in one case and ‘inevitably’ leading to something in the other suggests the same kind of opposition seen in A 76 above, and the point of the section may again be to draw attention to the fact that a single lexical item can in its usage entail this kind of seeming opposition.

FormalPara A 78
C::

名。達, 類, 私。

E::

名:物達也。有實, 必待之a名b也。命之馬, 類也。若實也者, 必以是名也。命之臧, 私也。是名也止於是實也。聲出口, 俱有名。若姓字c。

C::

míng ‘naming’: may be unrestricted, categorizing or personal.

E: :

míng ‘naming’: the word wù ‘thing’ is unrestricted; given any object, it inevitably allows for this kind of name. Giving it the name ‘horse’ is categorizing; any object of this sort will inevitably take this name. Giving it the name Zang is personal. The particular name remains fixed with the particular object. When sounds emerge from the mouth they all have naming-power, either like a surname or like a given name.

  1. (a)

    R: 文, emendation Graham 1978, 325, 之 is used regularly in the Later Mohist texts as the demonstrative modifier ‘this’.

  2. (b)

    R: 多, emendation Graham 1978, 325.

  3. (c)

    R: 宇, emendation Graham 1978, 325.

The word dá 達 ‘unrestricted’ means here having no restrictions on the applicability of a name. Our translation ‘unrestricted’ for the word dá 達 is taken from Graham 1978, 325; it is rendered ‘all-obtaining’ by Rafael Suter, who says that the Mohists “reserve” the word wù 物 for a “thing of any kind.” This is in contradistinction to the term shí 實 which Suter identifies in the Mohist context as the word for ‘things’ as “non-linguistic ‘objects’.” (Suter 2017) This in turn entails an implicit difference between ‘name’ as a linguistic device, contrasting one thing from another by virtue of its name and ‘name’ as an external, real-world label for something. For the sense of ‘kind’ as a category, the Mohists use the word lèi 類, as the Explanation shows with respect to naming something ‘horse’ mǎ 馬.

The last part of the Explanation, “when sounds emerge from the mouth they all have naming-power,” seems to mean that all utterances have a naming function, i.e., they will inevitably be associated nominally with something. In the phrase “either like a surname or like a given name” ‘surname’ is used to designate a category, and ‘given name’ is used to designate the particular. The same observation is found in Mencius 7B.36, where instead of zì 字, the particular is referred to by the word míng 名: 姓所同也, 名所獨也 “Surnames are considered to be held in common; given names are considered to be held individually.”

The word mìng < *mreng-s 命 ‘to give a name to’ is derived from míng < *mreng 名 ‘name’ by the addition of the OC *-s that derives transitive verbs from nouns. By Warring States period times the derivational relation between the two words seems already to have been non-productive and not widely recognized. At the same time the semantic scope of the word mìng < *mreng-s 命 ‘to give a name to’ came to include ‘to call, to order, to issue a mandate’. As a consequence, the word míng 名 ‘name’ was commonly used for the transitive verb ‘to name’ and perhaps read mìng < *mreng-s.

FormalPara A 79
C::

謂。移, 舉, 加。

E::

謂:麗a 狗犬命, 移b 也。狗犬舉也。叱狗加也。

C::

wèi ‘referring’: may be transferring, adducing or applying.

E: :

wèi ‘referring’: pairing ‘whelp’ and ‘dog’ in giving the name is transferring; ‘whelp’ or ‘dog’ [either used alone] is ‘adducing’; shouting at a dog is ‘applying’.

  1. (a)

    R: 灑謂; the order of the head character of E and the first character of the text has become reversed, see Graham 1978, 95–6; 灑 stands for 麗 lì ‘to pair with’, see Graham 1978, 326.

  2. (b)

    R: null, emendation Graham 1978, 326.

yí 移::

“pairing ‘whelp’ and ‘dog’” may be a reference to the common use of wèi 謂 ‘to refer to something by a different name’, e.g., 謂 狗犬 “to refer to a ‘whelp’ as a ‘dog’.”

jǔ 舉::

The term ‘adduce’ entails the sense of 謂 used to raise a topic for further comment. Sometimes this is grammatically the same as the first case, to refer to some object with a comment rather than a different name for the object.

jiā 加::

shouting at a dog is ‘applying’, i.e., addressing the dog directly (with some kind of command).

The Mohist is again here concerned with the difference between what can be recognized as a ‘speech act’ on the one hand versus a ‘real-world event’ on the other. The Explanation suggests that the verbal sense of yí 移 ‘transferring’ is understood as a speech act in that it involves only a relation between lexical items, whereas jǔ 舉 ‘adducing’ and jiā 加 ‘applying’ refer to real-world behaviors that pertain to a real dog.

FormalPara A 80
C::

知。聞a, 說, 親。名, 實, 合, 為。

E::

知:傳受之聞也。方不運b 說也。身觀焉親也。所以謂名也。所謂實也。名實耦合也。志行為也。

C::

zhī ‘knowing’: may be having heard, by explanation or personal; may be about names, about objects, about [their] conjunction or about acting.

E: :

zhī ‘knowing’: having received something transmitted is ‘having heard’; that a square object will not roll is ‘explicable’; being an observer to it yourself is ‘personal’; that whereby something is referred to is a ‘name’; what is referred to is the ‘object’. The name and the object as a pair is a ‘conjunction’; when one’s intentions are carried out it is ‘acting’.

  1. (a)

    R: 間, emendation Graham 1978, 327.

  2. (b)

    R: , emendation Graham 1978, 327, basing himself on earlier explanations of Wu Yujiang and Sun Yirang; is taken as a misprint for 庫 which is then taken as a variant of the unfamiliar character ; this is in turn understood as a graphic variant of (= 軍 jūn), the phonophoric of 運 yùn ‘to roll’. The latter part of this emendation is confirmed conclusively by the same occurrence of 庫 for 運 yùn ‘to rotate’ in the Canon of A 48, which emendation is unambiguously indicated by 運 as the head character of the Explanation.

While it is self-evident that a square object will not role, the comment in the Explanation is not trivial because it is intended to illustrate a kind of knowing that can be easily explained (說), distinct from the things you know having heard about them or observed them personally.

Notice that the Canon uses the word shí 實 ‘(non-linguistic) object’, conforming to the distinction seen in A 78 between the word wù 物 as the term that is used to name referentially a “thing of any kind” and shí 實 the word for a non-linguistic ‘thing’, i.e., ‘object’, and the Explanation follows in exactly this same sense. See the discussion at A 78. By including the statements “that whereby something is referred to is a ‘name’; what is referred to is the ‘object’” the Explanation is consistent with the yí 移 ‘transferring’ sense of wèi 謂 ‘calling’ in the immediately preceding section. Finally, the Explanation concludes by pairing the lexical item ‘name’ and the real-world ‘object’, making the speech act together with what it refers to a form of knowing.

FormalPara A 81
C::

聞。傳a, 親。

E::

聞:或告之傳也。身觀焉親也。

C::

wén ‘hearing’: may be transmitted, may be personal.

E: :

wén ‘hearing’: someone telling you something is ‘transmitted’; being an observer to it yourself is ‘personal’.

  1. (a)

    R: 愽, emendation Graham 1978, 328; the 愽 in R is a non-standard variant of 博 bó ‘wide-ranging’. While the E text with chuán 傳 ‘transmit’ makes it likely that Graham’s emendation is correct, it is not impossible to read the text as it stands with 愽 / 博 bó ‘wide-ranging’, thus “wén ‘hearing’ may be ‘wide-ranging’ (i.e., widely known), or may be ‘personal’.” That the E text unambiguously writes chuán 傳 ‘transmit’ would then have to be understood as a secondary change from 愽/博 bó ‘wide-ranging’; thus “when others announce it, it is ‘widely known’” to “someone telling you something is ‘transmitted’,” an example of the lectio facilior supplanting the lectio difficilior.

FormalPara A 82
C::

見。體, 盡。

E::

見:特a 者體也。二者盡也。

C::

jiàn ‘seeing’: may be elemental, may be exhaustive.

E: :

jiàn ‘seeing’: in particulars, is elemental; in the aggregate, is exhaustive.

  1. (a)

    R: 時, emendation Graham 1978, 329, based on conjecture of Sun Yirang.

In A 2 體 ‘element’ is explained as 二之一 “one of two,” where we can understand the ‘two’ as a composite set. By the same token we can understand the 二者 here also as a ‘composite set’, and thus standing for elements in the aggregate.

FormalPara A 83
C::

合。正, 宜, 必。

E::

合a:與b 立, 反中, 志工, 正也。臧之為, 宜也。非彼必不有, 必也。聖者, 用而勿必。必也者, 可而c 勿疑。反d 者, 兩而勿偏。

C::

hé ‘conjoining’: may be ‘straightforward’, ‘appropriate’ or ‘inevitable’.

E: :

hé ‘conjoining’: standing together, opposites in accord, intention and result, are examples of straightforward conjoining; Zang behaving as he does, is appropriate conjoining; were it not for something, something else inevitably would not come about, is inevitable conjoining. As for the sages, take advantage of them but do not regard [what they say] as inevitable. As for ‘inevitability’, accept it, do not be suspicious of it; as for opposites, regard them as dual, but do not privilege one over the other.

  1. (a)

    R: 古, emendation Graham 1978, 330.

  2. (b)

    R: 兵, emendation Graham 1978, 330.

  3. (c)

    R: null, emendation Graham 1978, 330 on the grounds that the phrase 可勿 is grammatically impossible, given the precise sense of 勿 negating only transitive verbs with third-person pronominal direct object.

  4. (d)

    R: 仗, Graham emends this to 仮 and treats it like the 反 of the earlier sentence (1978, 330).

The use of 正 in these lexical appendix entries is somewhat formulaic and serves to specify the meaning of the word in question in an unqualified and absolute sense, thus, here, ‘straightforward conjoining’, and so mutatis mutandis in the subsequent sections.

The phrase 反中 (‘opposites in accord’) is grammatically subject-verb, fǎn 反 ‘opposites’—中 zhòng ‘to hit the mark exactly’. The sense may refer to the kind of experiences that later in an analytical context come to be called ‘action-reaction’.

The reference to an individual (here named Zang, the rough equivalent of “John Doe”) and his behavior is called yí 宜 ‘appropriate conjoining’, i.e., the conjunction of a person and the behavior typically identified or associated with that person; yí < *ngraj 宜 ‘to behave in a way proper to an individual according to status or position’, akin to wǒ < *nggaj 我 ‘I, we, me, us’ and to yì < *ngrajs 義 ‘ritual or ceremonial behavior proper to an individual’.

The meaning of the line 非彼必不有, 必也 in the Explanation here is effectively the same as the definition of a ‘major basis’ in A 1, 大故, 有之必然 (‘Major basis: having it entails the inevitability of (a certain thing) becoming so’). The point here in A 83 is to show that the necessary condition and its inevitable consequence constitute a conjunction defined by the two elements’ causal relation.

The reference to the ‘sages’ and what they stand for or say seems to be intended to encourage recognizing a distinction between what they advocate and what may in fact be the case. The former does not necessarily imply the latter. So the Mohist endorses an open-minded response to their teachings, take advantage of them when it makes sense to do so, but do not regard them as inevitably right. Inevitability, the text asserts, should be accepted and not doubted, but the doctrines and claims of the sages vis-à-vis contrary opinions or facts are one of two opposites, to be recognized for what they are but not given preferential treatment. This train of thought generally seems to arise from the category of ‘inevitable conjoining’. This line of the E sounds atypically polemical in this part of the Mozi, though it is not inconsistent with the more narrative parts, especially, e.g., chapters 38 and 39, Fei ru 非儒 “In opposition to Confucians.” Here it may be a kind of marginal note that has inadvertently been included in the primary text.

FormalPara A 84a
C::

欲。正, 權利, 且。

E::

[null]

C::

yù ‘desiring’: straightforward desiring, appreciating the effectiveness of the benefits, being on the verge of.

E::

[null]

This and the following section have been put together by Graham as A 84. Given the fact that there is no other example of a pair of opposite terms being identified or disambiguated in a single Canon, coupled with the absence of any Explanation to provide a guide to the identification of Canon boundaries, we prefer to treat them as two separate Canons, splitting them up as A 84a and A 84b.

Including the term qiě 且 ‘being on the verge of, about to’ in C here is likely intended to suggest a connection with yù 欲, which also has the meaning ‘about to’; cf. English will (i) marker of future tense (“It will be so”) and (ii) ‘want’ (“I will that it be so”).

FormalPara A 84b
C::

惡。正, 權害。

E::

[null]

C::

wù ‘disliking’: straightforward disliking, appreciating the effectiveness of the harm.

E: :

[null]

For the identification of this section, see the note at A 84a immediately above.

FormalPara A 85
C::

為。存, 亡, 易, 蕩, 治, 化。

E::

為:中^a, 存也。病, 亡也。買鬻, 易也。霄燼b, 蕩也。順長, 治也。 買d, 化也。

C::

wéi ‘acting’: may be sustaining, eliminating, switching, dislodging, putting in good order or transforming.

E: :

wéi ‘acting’: for according exactly with a counterpart it means ‘sustaining’; for illnesses it means ‘eliminating’; for buying and selling it means ‘switching’; of snow-slush or ashes it means ‘dislodging’; of conforming and leading it means ‘putting in good order’; of the frog and trading it means ‘transforming’.

  1. (a)

    R: 早臺, emendation Graham 1978, 332; 中^zhòng dí understood as verb-object, ‘to accord exactly with a counterpart’.

  2. (b)

    R: 盡, emended on the basis of suggestion by Graham 1978, 332.

  3. (c)

    wā ‘frog’ as in A 45.

  4. (d)

    Based on the E text of A 45, which specifies the chún 鶉 ‘quail’ as the thing into which the frog is transformed, the expected word here should be chún 鶉 ‘quail’. As reasonable as such an emendation sounds, it is also possible to understand the text as it stands, which by default we are obliged to do.

At first glance the Explanation’s reference to 買 in the last example 買 ‘frogs and trading’ as 化 ‘transforming’ would seem to be inconsistent with the preceding reference to 買鬻 ‘buying and selling’ as 易 ‘switching’. It appears that the two references to trading reflect different aspects of that kind of transaction. Buying and selling seen as a bilateral exchange is an example of ‘switching’, the same action viewed from opposing perspectives. By contrast 買 in its second occurrence is not simply ‘buying’ but is the whole of a commercial transaction, 買賣 mǎi mài, i.e., trading one thing for another, turning one thing into another. The two verbs mǎi < *mmraj-Ɂ ‘buy’ and mài < *mmraj-s ‘sell’ are no more than morphological variants of a single word; the *-s of mài ‘sell’ marking the causative sense of mǎi ‘buy’, that is ‘sell’ < ‘cause [someone] to buy’. In other words, what seem to be two distinct verbs can be understood as two parts of a single action, and the character 買 in the E text here could stand for the word for the transaction in general without specifying direction. It thus seems to refer to trading as the transformation of a commodity from one form to another, just as the frog was described in A 45 as being transformed into a quail. If we understand it in this way, the sense of 買 mǎi in 買鬻 mǎi yù ‘buying and selling’ is “half” of the sense it has in 買 wā mǎi ‘frogs and trading’, and this may be another instance of drawing attention to apparently opposite (or in this case quasi-opposite) meanings of the same word.

For the semantic relation between 化 huà ‘transform’, ‘convert’, cf. 貨 huò ‘commodities, goods convertible into something of equivalent value’.

FormalPara A 86
C::

同。重, 體, 合, 類。

E::

同:二名一實, 重同也。不外於兼, 體同也。俱處於室, 合同也。 有以同,類同也。

C::

tóng ‘same’: may be of a duplicate kind, an elemental kind, a conjoined kind or a categorical kind.

E: :

tóng ‘same’: two names and one object means a ‘duplicated kind of sameness’; not external to the composite whole means an ‘elemental kind of sameness’; all located in a room means a ‘conjoined kind of sameness’; having something whereby they are the same means a ‘categorical kind of sameness’.

FormalPara A 87
C::

異。二, 不a 體, 不合, 不類。

E::

異:二必異, 二也。不連屬, 不體也。不同所, 不合也。不有同, 不類也。

C::

yì ‘different’: may be of a being-two kind, a not-functioning-as-elements [of a single composite] kind, a not-conjoined kind or a not-in-a-single-category kind.

E: :

yì ‘different’: two inevitably different means ‘being two’; not connected or attached means ‘not functioning as elements [of a single composite]’; not in the same place means ‘not conjoined’; not having something whereby they are the same means ‘not in a single category’.

  1. (a)

    R: null, restored on the basis of E (Graham 1978, 334).