\"mozi 31: explaining ghosts, again\"

$: \"Mozi 31: Explaining ghosts, again\"$
© 2013 Koninklijke Brill NV ISBN 978-90-04-23434-5

The Mozi as an Evolving Text

Diffferent Voices in Early Chinese Thought

Edited by

Carine Defoort and Nicolas Standaert

LEIDEN • BOSTON2013


CONTENTS

Introduction: Diffferent Voices in the Mozi: Studies of an Evolving Text .............................................................................................. 1 Carine Defoort and Nicolas Standaert

1. Are the Three “Jian Ai” Chapters about Universal Love? ............. 35 Carine Defoort

2. How to End Wars with Words: Three Argumentative Strategies by Mozi and His Followers ..................................................................... 69

Paul van Els

3. Mozi 31: Explaining Ghosts, Again ....................................................... 95 Roel Sterckx

4. Mozi’s Remaking of Ancient Authority .............................................. 143 Miranda Brown

5. The Ethics of the Mohist Dialogues .................................................... 175 Chris Fraser

6. From “Elevate the Worthy” to “Intimacy with Offfĳicers” in the Mozi ....................................................................................................... 205

Hui-chieh Loy

7. Heaven as a Standard .............................................................................. 237 Nicolas Standaert

Bibliography ..................................................................................................... 271References to the Mozi .................................................................................. 281 Subject Index .................................................................................................... 287


MOZI 31: EXPLAINING GHOSTS, AGAIN

Roel Sterckx *

One prominent feature associated with Mozi (Mo Di 墨翟; ca. 479–381 BCE) and Mohism in scholarship of early Chinese thought is his so-called unwavering belief in ghosts and spirits . Mozi is often presented as a Chi-nese theist who stands out in a landscape otherwise dominated by this-worldly Ru 儒 (“classicists” or “Confucians”).1 Mohists are said to operate in a world clad in theological simplicity, one that perpetuates folk reli-gious practices that were alive among the lower classes of Warring States society: they believe in a purely utilitarian spirit world, they advocate the use of simple do-ut-des sacrifĳices , and they condemn the use of excessive funerary rituals and music associated with Ru elites. As a consequence, it is alleged, unlike the Ru, Mohist religion is purely based on the idea that one should seek to appease the spirit world or invoke its blessings, and not on the moral cultivation of individuals or communities. This sentiment is reflected, for instance, in the following statement by David Nivison :

Confucius treasures the rites for their value in cultivating virtue (while virtually ignoring their religious origin). Mozi sees ritual , and the music associated with it, as wasteful, is exasperated with Confucians for valuing them, and seems to have no conception of moral self-cultivation whatever. Further, Mozi’s ethics is a “command ethic,” and he thinks that religion, in the bald sense of making offferings to spirits and doing the things they want, is of fĳirst importance: it is the “will” of Heaven and the spirits that we adopt the system he preaches, and they will reward us if we do adopt it. He takes it for granted that we will not do what we should (in this sense) if we do not believe in spirits or Heaven or if we think that good fortune depends on

* I would like to thank Carine Defoort and Nicolas Standaert for detailed comments and corrections on an earlier draft of this essay, as well as the participants of the Univer-sity of Leuven conference “The Many Faces of Mozi: A Synchronic and Diachronic Study of Mohist Thought” (25–28 June 2009) for their input on the paper out of which it has grown.

1 See, e.g., Ching, Chinese Religions, 70. JeeLoo Liu speaks of Mozi as the “only ancient philosopher who dealt with philosophy of religion” and “the most religious among ancient philosophers.” See Liu, An Introduction to Chinese Philosophy, 110, 124. Angus Graham char-acterizes the Mohists as being “at once the most religious and the most logical of the ancient thinkers.” See Graham, Later Mohist Logic, Ethics, and Science, 4.

96 roel sterckx


“fate” rather than being a reward for good deeds. So people must believe in spirits and must not believe in fate. In his view of religion there is (as in his ethics ) no inner feeling or awe.2

Sophisticated Ru are juxtaposed here to simple-minded Mo and the impli-cation is that both are rival camps. Michael Puett emphasizes that the Mohists were not bent on sacrifĳicing to transform Heaven . Sacrifĳices for the Mohists, he argues, “are simply a case of humans giving the spirits what the spirits need, just as the spirits give humans what humans need.”3 In Puett ’s reading, neither Heaven nor the spirits can be capricious to the Mohist since they act according to a clear moral calculus. Mohist inter-action with the spirit world therefore is highly perfunctory and efffĳicient since the spirits act in an entirely predictable way: you get out of the spir-its what you put into them.

Another simplifĳication of Mohist religiosity is the idea that their views on spirits are a remnant of an archaic religious world that became gradu-ally superseded by a tide of conceptually more sophisticated philosophies. In this linear view of the history of Chinese thought, Warring States ghosts and spirits make way for rationality and ritual as we move toward the Han. For instance, Burton Watson speaks of “a growing atmosphere of sophisti-cation and rationalism [that] led men to reject or radically reinterpret the ancient legends and religious beliefs that Mozi had so fervently afffĳirmed.”4 Mozi’s “inability to prove that Heaven or the spirits deliver immediate rewards and punishments ,” Benjamin Schwartz writes, “probably left his upper-class hearers indiffferent to his particular religious message.”5 And, according to Lester Bilsky , the Mohist view that spirits give blessings in proportion to the number of offferings they receive had little impact on the overall practice of religion and government.6

Sometimes, the plebeian nature of Mozi’s take on the spirit world is grafted on to a sociology of the Mohist community, a subject for which very little concrete evidence survives in our sources. Angus Graham

2 Loewe and Shaughnessy, Cambridge History of Ancient China, 761. Chris Fraser writes in a similar vein but qualifĳies his statement: “The mundane tone of their religious thought is perhaps partly due to the Mohists’ general disregard of aesthetic and cultural value. Yet it would be indefensibly parochial to expect them to conform to a modern Westerner’s conception of religiosity, since the comparatively mundane character of Mohist religion is typical of much traditional Chinese folk religion.” See Fraser, “Mohism.”

3 Puett, To Become a God, 101–104, quotation on 102.4 Watson, Mo-tzu, Basic Writings, 9–13, quotation on 13. 5 Schwartz, The World of Thought in Ancient China, 170.6 Bilsky, The State Religion of Ancient China, 211–212.

mozi 31: explaining ghosts, again 97


notes that “in passing from the Analects to Mozi one has the impression of descending to a lower stratum of society without access to the higher culture of Zhou.”7 Mozi’s alleged association with the profession of arti-san-carpenters is adduced to support this point. Thus, Graham suggests we think of Mohism “as springing from a class to some extent compa-rable with the merchant class of Renaissance Europe.”8 While, to be sure, there is artisanal imagery present in the received Mozi, it remains risky to infer the provenance of metaphors from an undocumented biography.9 Like others mentioned above, Graham takes the Mohist belief in ghosts and spirits that are capable of punishing and rewarding as a concept that “belongs rather to folk religion.” And like others he contrasts this with the Confucian instinct to keep the spirit world at a distance, appease and offfer sacrifĳice , and refrain from drawing direct moral correlations that inspire do-ut-des type of veneration. “The Mohist,” Graham writes, “. . . comes from a less sophisticated class in which the perennial folk religion of China is still alive, and does not like his masters to forget that they too are subjects of still higher beings.”10 On balance, we are led to believe that Mohist views on the spirit world are to be equated with the lower, or “popular,” strata of society, whereas other “schools” or thinkers deal with conceptu-ally more elevated numinous forces such as Heaven and the Mandate or with strategies for moral self-cultivation. The assumption then is that cul-tic practices were irrelevant to intellectual elites or, indeed, that “popular” or “common” religion can be confĳidently associated with a social class, a premise that is highly questionable.11

To be sure, all these interpretations are invaluable attempts at teasing out the Mohist take on spirits . Yet it is striking that they seem to begin from

7 Graham, Disputers of the Tao, 34. 8 Ibid., 45. See also Graham, Later Mohist Logic, Ethics, and Science, 8. Philip Ivanhoe

and Bryan Van Norden characterize Mozi’s philosophy as “distinctively anti-aristocratic.” See Ivanhoe and Van Norden , Readings in Classical Chinese Philosophy, 55.

9 Indeed, because artisans are also associated with the creation of goods and products that support the luxuries and riches of aristocrats, Mozi would have to disown his own professional afffĳiliation if ever he had one.

10 Graham, Later Mohist Logic, Ethics, and Science, 14. Yuri Pines writes: “Most Zhanguo texts . . . largely neglect Heaven’s will, and no Zhanguo thinker, with the exception of Mozi, assigns divine forces any signifĳicant role in political and social life.” See Pines, Foundations of Confucian Thought, 55. Pines’s statement holds only if one uses a very narrow defĳinition of “divine forces” and excludes, for example, ancestral spirits .

11 For a critique of distinctions in belief systems according to social class, see Harper, “Warring States, Qin, and Han Periods”; Harper, “Contracts with the Spirit World in Han Common Religion”; and Sterckx, “Religious Practices in Qin and Han.”

98 roel sterckx


two assumptions: fĳirst, that views on the spirit world across the received Mozi are unitary and internally uncontested; second, that Mozi’s views on the spirit world should invariably be seen as conversant with or pitched against a countercurrent, the so-called Ru (if there is such a thing as a uni-tary Ru view on spirits ). In what follows I argue that these assumptions are problematic and overlook subtle diffferences and problems embedded in the received Mozi text. I suggest that this is specifĳically the case with Mozi 31, the only preserved chapter in the “Ming gui” 明鬼 (Explaining Ghosts)12 triad and the principal source adduced to represent Mohist doctrine on the spirit world.13 When read carefully against the other Core Chapters , the Dialogues , and, to a lesser extent, the Opening Chapters , Mozi 31 not only leaves signifĳicant question marks over how the Mohists conceived of the spirit world and its workings but also throws doubt on whether we can speak at all of one unitary and agreed Mohist view on the issue. My analysis suggests that Mohist views on spirits evolve or, at least diversify, across the received Mozi text, although the temporal sequence in which these developments took place remains nearly impossible to reconstruct. My claim in favor of a more variegated and multipolar Mohist view of spirits not only is based on an internal analysis of the received Mozi but also is reinforced by possible readings of a manuscript fragment , titled “Guishen zhi ming” 鬼神之明 (Ghostly Percipience) by its editors , among the Chu bamboo-slip materials held in the Shanghai Museum.

12 Most translators take ming as an adjective to mean something along the lines of “percipient” or “conscious.” These are certainly possible. I have opted for a verbal transla-tion since the chapter covers more ground than the question of ghostly consciousness.

13 The textual history of the received Mozi (divided into seventy-one chapters, fĳifty-three of which are extant) continues to be full of unknowns. Pre-Tang bibliographies indi-cate that a version of the text was circulating that included a separate index chapter (mu yi juan 目一卷), but no mention is made of this in sources after the mid-Tang. See Wang Changmin, “Mozi guben mulu ji yipian kaoshuo”; and Wang Changmin, “Mozi pianmu kao.” It remains uncertain when exactly the two missing chapters in the “Ming gui” triad were lost, if they ever existed at all. They are fĳirst listed as missing together with six other titled chapters in the oldest fully preserved Mozi text, the Ming Zhengtong Daozang 正統道藏 (1447), which forms the basis of the presently transmitted version. For the purpose of this essay I circumnavigate some of the more complex text-critical hypotheses on the Core Chapters as a whole that have been proposed by scholars such as Alfred Forke , Chris Fraser , Stephen Durrant , Angus Graham , Erik Maeder , Carine Defoort , Karen Desmet , and others. I will refer to philological and text-critical points made by scholars where they have a signifĳicant bearing on the actual arguments and ideas presented in the text. Most schol-ars have drawn on intratextual signifĳiers to propose hypotheses on internal developments within a Mohist “school.” With reference to Mohist views on the spirit world, intra- and intertextual evidence that permits us to tease out clearly articulated currents of thought within Mohism beyond a reasonable level of speculation remains scarce.



The Arguments in “Ming gui, xia”

Let us look fĳirst at the arguments in the received Mozi 31 and the struc-ture of its narrative. Mozi 31, which, at 3,406 characters, is the longest of the Core Chapters ,14 applies Mozi’s three tests of verifĳication to the ques-tions of “whether spirits exist” 鬼神之有與無之別 and whether “they are capable of rewarding the worthy and punishing the wicked” 能賞賢而罰暴 (31: 50/25–26 ).15 These three tests are (1) spirits exist since there have been reports and eyewitness accounts of them throughout the ages and across geographical space; in other words, common observation and veri-fĳiability through hearing and sight constitute proof; (2) spirits exist since the sages of antiquity believed in their existence; in other words, ancient authority forms the basis of proof; and (3) they exist since spirits are a necessary condition for the establishment of a morally functioning soci-ety, and people will behave better when they know that their actions are sanctioned; in other words, practical utility and applicability for the ben-efĳit of society provide proof. The elaboration of these three tests (called “source,” “root,” and “use” by Chris Fraser ),16 in that order, forms the bulk of the chapter.

Structure of Mozi 31Introduction: The contemporary world and its social structure are in disorder . Why?Because:1. There exists doubt on the existence of spirits .2. People fail to realize that spirits have rewarding and punishing powers.Proof:Test 1: Witness accounts:

1. The unrighteous killing of Du Bo 杜伯 by King Xuan of Zhou 周宣王 (827/25–782 BCE).

2. The spirit Gou Mang 句芒 prolongs the life of Duke Mu of Qin 秦穆公 (659–621 BCE).17

3. The unrighteous killing of Zhuang Ziyi 莊子儀 by Duke Jian of Yan 燕簡公 (504–493 BCE).

14 Following the Sibu congkan edition as in the table presented in Desmet, “All Good Things Come in Threes,” 11; and Desmet, “The Growth of Compounds in the Core Chapters of the Mozi,” 100.

15 The second question no doubt inspired E. R. Hughes’s (in many ways, rather ade-quate) translation of ming 明 as “moral intelligence,” although, strictly speaking, it mis-translates ming gui. See Hughes, Chinese Philosophy in Classical Times, 50.

16 See Fraser, “Mohism.”17 Following Bi Yuan and Sun Yirang , most scholars replace 鄭 by 秦.

100 roel sterckx


4. A sacrifĳicial offfĳicer of Duke Wen of Song 宋文公 (610–589 BCE) is caned by a spirit through the intermediary of an invocator for provid-ing substandard offferings.

5. A minister of Duke Zhuang of Qi 齊莊公 (794–731 BCE) is head-butted by a goat during a blood oath.

Test 2: The sages of the past fĳirmly believed in spirits , as is reflected in their sacrifĳicial practices and documented in their texts.Test 3: Making a belief in ghosts foundational to the state (本施之國家) will benefĳit the people.Concluding reservations:Include a stretched argument on fĳilial piety and an expression of doubt by Mozi concerning his own thesis.

The interlocutors in Mozi 31 are referred to as “those who maintain that there are no ghosts ” (執無鬼者, mentioned seven times). Throughout the chapter the assumption is that the masses (天下之眾) are naturally receptive to believing in ghosts and are in doubt only when their superi-ors (天下之王公大人士君子) question the spirits ’ existence. Note that the proposition of the chapter is not the positive assertion that “ghosts exist”; rather, the chapter concerns a search to confĳirm “whether or not they exist” and “whether or not they reward and punish.” The key of Mozi 31 lies not in providing existential proof but rather in demonstrating the utility of argument. Note also that the proposal to make people believe in ghosts is formulated in the conditional voice, implying that those in charge need not necessarily believe in their existence but can promote creeds and cults as a ploy to influence social behavior:

今若使天下之人借若信鬼神之能賞賢而罰暴也, 則夫天下豈亂哉.Now, if one could cause all people under Heaven to conform to the belief that ghosts and spirits are capable of rewarding the worthy and punishing the wicked, then how could there be disorder in the world? (31: 50/26–27 )

There are no explicit indications in Mozi 31 that Mozi equates “those who maintain that ghosts do not exist” specifĳically with the Ru. The graph 儒 does not occur at all in Mozi 31. The only Core Chapter in which it appears is Mozi 39 (“Fei Ru, xia” 非儒下), which is believed to be a later addi-tion to the ten other doctrines . Beyond that, the graph ru appears only in Mozi 48, a chapter (“Gongmeng” 公孟) among the Dialogues .18 With

18 Naturally, the absence of the graph ru need not mean by itself that the Ru could not be implied, but the statistics are striking nevertheless.



the exception of a covert reference to the three-year mourning period (in witness accounts 1 and 3 of test I), there is little that directly links the ideas presented in this chapter to a clearly identifĳied individual or group of interlocutors anywhere else in the received Mozi. The recurring for-mula that introduces the Mozi 31 interlocutors as “those who hold that . . .” (zhi . . . zhe 執 . . . 者) need not refer to the same group of people across all Core Chapters . It can simply be used to set up a dialogic setting.19

The fĳirst test—witness accounts—is structured along a series of paral-lel stories with clear intratextual parallels. The choice of witness accounts suggests a rhetorical claim for the universal application of the Mozi 31 thesis. The incidents described take place across geographical space on a Spring and Autumn period map: Zhou (central), Qin (west), Yan (north-east), Song (south), and Qi (east).

The unlucky protagonists who are submitted to an unjust execution in accounts 1 and 3 reply to their executioners in similar phrases with only the formula for the embedded conditional clause (underlined) showing signifĳicant grammatical variety:

(Account 1) 吾君殺我而不辜, 若以死者為無知則止矣; 若死而有知, 不出三年, 必使吾君知之. (31: 51/8–9 )(Account 3) 吾君王殺我而不辜, 死人毋知亦已, 死人有知, 不出三年, 必使吾君知之. (31: 51/23–24 )My lord (majesty), you are about to put me to death although I am not guilty of a crime . If the dead do not have consciousness, then it stops here. If the dead do have consciousness, they will make sure that my lord knows this before three years have elapsed.

Even more striking is the recurrence across witness accounts 1, 3, 4, and 5 of the following identical statement:

當是(之)時, X人從者莫不見, 遠者莫不聞, 著在 X 之《春秋》˳ 諸侯傳而語之曰:At that time, there were none among the [name of state X] attendants who did not see what had happened and no one in distant regions who had not caught rumor [heard ] of it; it was written in the Springs and Autumns of [name of state X], and the feudal lords handed down the story saying: [moral lesson X]. (31: 51/26–27 , 31: 52/5–6 , 31: 52/13–14 , and slightly difffer-ently in 31: 51/11–13 )

19 Cf. similar formulas such as 執厚葬久喪者 and 執有命者 in 25: 39/9 , 25: 40/28 , 25: 41/18 , 35: 58/15 , 35: 58/17 , 35: 59/2 , 35: 59/4 , 35: 59/17 , 35: 59/21 , 35: 60/7 , and 35: 60/13 .

102 roel sterckx


The core elements in witness accounts 1, 3, 4, and 5 include (1) the actual cause that provokes ghosts to avenge themselves and (2) the universal relevance of the moral message across all states marked by the fact that the incidents were recorded in the local chronicles and then didactically applied by successive feudal lords. The progression of authority runs as follows: “seeing” (jian 見) → “hearing” (wen 聞) → “recording” (zhu 著) → “transmitting” (chuan 傳).20 Only account 2 stands out in that it describes a spirit rewarding a virtuous character (though not in the language of shang fa 賞罰) and in that it lacks reference to a written record. In sum, these witness accounts, with the exception perhaps of account 2, seem to be built around a formulaic set of nearly identical narrative cues in which an author inserts diffferent events. If broken down to subunits such as the level of the paragraph (ce 冊) therefore, we may have a series of parallel accounts strung together with the narrative formula “it is not only in the explanations found in these writings that it is so, but . . .” 非惟若書之說為然也, a formula that occurs fĳive times in Mozi 31. In the absence of potentially two preceding chapters in the triplet, one might be tempted to infer a temporal sequence across the witness accounts based on the dat-ing of each account’s chief persona. This however reveals little. While the most parallel accounts (1 and 3) are set around three hundred years apart, there appears to be no temporal progression from past to present, unlike in the narrative buildup for the second test, discussed below. Finally, it is worth pointing out that drawing on recorded witness accounts as a way to invoke authority is hardly an original Mohist argumentation strategy. When read as an independent unit, it remains unclear who Mozi’s spar-ring partners are in Mozi 31.

If empirical accounts (耳目之情) fall short of proving the existence of ghosts , the second test ( fa 法) consists of invoking the authority of past sages. Taken on its own, calling on the ancients to bolster one’s argument is a rhetorical technique that does not really set Mohists apart from other Warring States thinkers. In his choice of past exemplars the compiler of Mozi 31 does not specifĳically favor the golden age of Zhou over any other period. This is a pattern that recurs across the other Core Chapters , where Mozi tends to select his examples from among the Three Dynasties

20 Mark Lewis notes that empirical observation here gains the status of proof once it is written down: “The fact of inscription demonstrates the truth of an account. It is granted a weight equal to that of direct perception.” See Lewis, Writing and Authority, 115. I would argue that writing does not necessarily grant equal authority to direct perception but rather that it is the result of a sustained sequence of witness accounts.



(Xia, Shang, Zhou) without a particular preference for the time of Con-fucius .21 Examples adduced to suggest that the ancients did believe in the existence of spirits include the Shang conqueror King Wu ’s 武王 sharing of sacrifĳices with his feudal lords, the fact that each of the Three Dynasties issued rewards at the ancestral temple and punishments at the altar of the soil to publicly demonstrate their justness, and the existence of writ-ten records testifying to the existence of spirits (mentioned are a partial quotation of “Wen wang” 文王, Ode 235 of the “Da ya” 大雅 (Great Odes ), and allusions to the “Yi xun” 伊訓 and “Yu shi” 禹誓 (i.e., “Gan shi” 甘誓) chapters in the Shangshu). Similar to the rhetoric of inclusive geographi-cal space in test 1, the test 2 narrative claims universal applicability across time. Mozi 31 builds up the argument by progressively pushing exam-ples farther back in time: starting with the books of Zhou, then Shang, then Xia.

The third test—utility and social benefĳit —zooms in on the second proposition in Mozi 31, namely the rewarding and punishing powers attributed to the spirit world, a belief that can be used to instill moral order. The general tenet of test 3 is that spirits are all-seeing; they spot all untoward behavior among humans (鬼神見之), and their punishments overcome all. The parallel to 鬼神之明 (ghostly percipience) is therefore 鬼神之罰 (ghostly punishment):

故,鬼神之明不可為幽閒廣澤, 山林深谷, 鬼神之明必知之.

鬼神之罰不可為富貴眾強, 勇力強武, 堅甲利兵, 鬼神之罰必勝之.

Therefore, when it comes to the percipience of ghosts and spirits ,it cannot be (escaped from) in dark creeks and broad swamps, mountain forests and deep valleys,since the percipience of ghosts and spirits will inevitably know about you. (31: 54/11 )When it comes to punishment meted out by ghosts and spirits,it cannot be (overcome) by wealth, nobility, strength of numbers, bravery, physical power, military valor, strong armor, or sharp weapons,since the punishments meted out by ghosts and spirits will inevitably conquer these. (31: 54/11–12 )

21 As noted in A. Cheng, Histoire de la pensée chinoise, 99. See also Miranda Brown’s essay in the present volume.

104 roel sterckx


Unlike the reverse chronology in test 2, examples from the past in test 3 are ranked from past to present: Heaven commands King Tang 湯 to chastise King Jie 桀 of the Xia; Heaven commands King Wu 武 to chas-tise King Zhou 紂 of the Shang . Note that the expression used is 至明罰 (to carry out clear/visible/enlightened punishments), suggesting again that we have a wordplay following the parallel structure sketched above in which, somehow, ming 明 and fa 罰 are treated as coordinate con-cepts, which is difffĳicult to convey in translation. Both accounts end with an identical sentence in conclusion:

此吾所謂鬼神之罰, 不可為富貴眾強、勇力強武、堅甲利兵者, 此也.This is what I mean when I say that punishments meted out by ghosts and spirits cannot be (overcome) by wealth, nobility, strength of numbers, brav-ery, physical power, military valor, strong armor, or sharp weapons. (31: 54/17–18 )

Finally, reference is made to a text entitled “Qin Ai” 禽艾 (31: 54/27 ), pos-sibly a lost chapter from the Yi Zhoushu 逸周書.22

Following test 3, Mozi 31 then takes a sudden thematic turn in which the “ghosts do not exist” party brings up the subject of fĳilial piety. The text at this point is not at all clear, and its link to what precedes is vague, which could suggest that the passage is corrupt or does not belong with the rest of Mozi 31. The claim by the skeptics seems to be that belief in ghosts (generally) implies that one does not benefĳit one’s kin (parents ) and hence is detrimental to fulfĳilling the duties expected of fĳilial offfspring (namely to be fĳilial to one’s own lineage fĳirst). Mozi in turn seems to sug-gest that, “provided ghosts and spirits really exist” 若使鬼神情/誠有, even the obligation to be fĳilial toward one’s own kin (a virtue that could be depicted as partial) hinges on a belief in the existence of ghosts , in this case ancestral spirits, who are offfered food and drink in sacrifĳice . (Clan-based) fĳiliality (xiao 孝) and (altruistic) benefĳit (li 利) therefore need not be mutually exclusive.

The fĳinal part of Mozi 31 then adopts a slightly diffferent tone in which the force of Mozi’s argument seemingly dissipates. Although Mozi can-not answer the central question on the existence of the spirit world, he suggests that even if ghosts or spirits do not exist, one should behave as

22 The fĳigure of Qin Ai occurs in Yi Zhoushu “Shi fu.” Zheng Jiewen speculates that “Qin Ai” is the title of a chapter in a Shangshu version circulating in Zhanguo times. See Zheng Jiewen, Zhongguo Mo xue tongshi, 107.



if they really did since this would benefĳit society. Organized religion, if not theologically watertight, at least has the advantage of bringing people together:

若使鬼神請亡, 是乃費其所為酒醴粢盛之財耳˳ 自夫費之, [非]特注之汙壑而棄之也, 內者宗族, 外者鄉里, 皆得如具飲食之˳ 雖使鬼神請亡, 此猶可以合驩[歡]聚眾, 取親於鄉里˳ If ghosts and spirits truly do not exist, all this may seem to be nothing more than a waste of material resources (cai) used for sacrifĳicial wine and grains. But, over and above, such expenditure is not a case of simply throwing things away into a ditch or gully. For the members of the clan and friends from the village and district can still come together and eat and drink [the offferings]. So even if there were really no ghosts and spirits , a sacrifĳice still enables them to gather together at a party, and people can befriend their neighbors. (31: 55/5–7 )23

Compared with the sustained buildup and consequential logic displayed in the three tests , the argument in the fĳinal part of Mozi 31 lacks vigor; in fact, this part of the chapter seems to represent a strand of Mohism that is less convinced of or indeed less preoccupied with the theoretical proof of the existence of ghosts. In its claim that even if one allows for the pos-sibility that spirits do not exist, one should still perform cult rituals , this fĳinal segment is ultra-utilitarian in that it is no longer concerned with the question of the you 有 or wu 無 of a spirit world. In short, the part of Mozi 31 following test 3 reads more like a discourse about the validity of the thesis itself rather than part of its proof. In that sense it is more akin to the skeptical tone we encounter in some of the passages on spirits in the Dialogues , as we shall see, rather than to the argumentative style that characterizes most of the Core Chapters .

Even if one takes these fĳinal exchanges as somehow related to test 3 on account of their advocacy of the pure social utility of the sacrifĳicial cult, and even if one takes “those who hold that there are no ghosts” to be an implicit reference to the Ru, it remains uncertain how the arguments in Mozi 31 can be seen as diametrically opposed to what one might expect from the Ru. Neither “school” uses, for the sake of argument, mutually exclusive defĳinitions of the types of spirits that make up their pantheon. It is true that in Mozi 31 the master claims: “What constitutes ghosts in past and present consists of nothing other than the ghosts of Heaven ,

23 This passage is also singled out in Schmidt-Glintzer , Mo Ti: Solidarität und allgemeine Menschenliebe, 27.

106 roel sterckx


and also the ghosts and spirits in mountains and rivers, and also dead people turning into ghosts” 古今之為鬼, 非他也, 有天鬼, 亦有山水鬼神者, 亦有人死而為鬼者 (31: 55/1–2 ). Yet these attempts at delineat-ing a clear-cut identity of what should be understood by a “ghost” should not be taken literally. For instance, at no point does Mozi 31 imply that the Mohist spirit world does not include ancestors : ghosts of dead people must include ancestors, the text also mentions the lineage (zongzu 宗族) gathering in convivial sacrifĳice , and it speaks of sacrifĳice as a demonstra-tion of fĳilial piety (31: 55/5–6 ). Furthermore, Mozi 31 insists that spirits as a rule brought rewards to people at the ancestral temple (zu 祖) and punished at the altar of the soil (she 社), both loci not quite congruent with the idea of impartiality found elsewhere in the Core Chapters . Like-wise, throughout Mozi 31 the use of the terms gui 鬼 and shen 神 seems arbitrary, and neither term is used consistently with a clear referent.

The image presented at the end of Mozi 31, namely that of sacrifĳicial ceremonies as feasts and occasions for human interaction, could sit equally well with the classicist take on sacrifĳice . However, the proposal here to maintain the cult, in the absence of theological substance, for the sake of promoting social cohesion might be difffĳicult to square with a comment in Mozi 39 where Mozi castigates the vulgar Ru as beggars who “stufff food away like hamsters,” dragging themselves along to large funerals with their extended family to feast on drink and food.24 Whereas in the Mozi 39 passage the targets of Mohist criticism are explicitly identi-fĳied as Ru, the social aspect of sacrifĳicial religion invoked in Mozi 31 is not far removed from expectations one may fĳind expressed in a Ru context. One could even argue that the above line of argument—“when in doubt about the existence of spirits , sacrifĳice to them communally anyway”—comes close to the highly functional or even utilitarian admonishment to “respect ghosts and spirits but keep them at a distance” 敬鬼神而遠之 of the Analects (6.22).

In short, Mozi 31 poses a number of interpretive hurdles. While its overtone is clear, there are few internal clues in the text that allow us to assume that the chapter is pitched deliberately as an exclusively anti-Ru polemic. Even its own central proposition, the existence of ghosts and spirits , is phrased in the conditional voice and subject to doubt at the end of the chapter. Parts of the chapter, such as the witness accounts, appear like a catalog of nearly identical story units, while the latter half

24 See Mozi, “Fei Ru, xia” (39: 64/19–26 ).



of the chapter is interspersed with moments of self-interrogation aimed at clearing away doubts about the validity of the central proposition, the existence of ghosts. Ghosts may not really exist, but a persuasive argu-ment that they do can be useful regardless.

The Dating of Mozi 31

Since Mozi 31 is the only transmitted chapter in its triplet, it is difffĳicult to draw on it to establish signifĳicant temporal developments in Mohist views on the spirit world. In the case of the “Ming gui” triad, it is impos-sible to test the influential three-sects theory —proposed by Alfred Forke , Luan Tiaofu 欒調甫, Angus Graham , and others—according to which each of the Triplets would represent a distinct school of Mohism. There are indications toward the end of Mozi 31, as I have suggested above, that we might be dealing with a less radical and more self-inquiring group of Mohists. But such speculation only stands if the fĳinal part of Mozi 31 is taken as a continuous narrative following on from test 3 rather than a text unit that has slipped in from elsewhere and is better read together with some exchanges in the Dialogues .

In the 1960s Watanabe Takashi 渡邊桌 proposed that Mozi 31 belonged to the most recent period of the Core Chapters and that it could be dated to the late third century BCE around the Qin unifĳication. Watanabe based his argument for such a late date , among other reasons, on the fact that the three tests as set out in Mozi 31 are conceptually identical to the three tests (san biao 三表 ) in the “Fei ming” 非命 triplet (Mozi 35, 36, 37). The fact that the term itself (san biao or san fa 三法 “three models, standards”) does not occur might indicate, he suggests, that Mozi 31 could slightly predate the “Fei ming” discussions. Watanabe is also not convinced that there ever existed a “Ming gui” triad and argues that the chapter stands by itself as a conceptual adjunct or supplement to the “Tian zhi” 天志 triplet, which he also takes to be late. Watanabe further speculates that the place-names in Mozi 31 indicate that it might have been written in the state of Qin (and he goes as far as to claim that the ideas in these late chapters served the ideology of Qin unifĳication).25 Watanabe ’s 1960s hypothesis of

25 Watanabe Takashi, “Bokushi shohen no chosaku nendai”; and Watanabe Takashi, Kodai Chūgoku shisō no kenkyū, 764. Zheng Jiewen speculates that since Mozi uses exam-ples that involve divination, he is carrying on Song traditions that, in turn, are based on Shang practices. Zheng’s assumption that Shang religion equals the worship of spirits

108 roel sterckx


a Mozi 31 as recent as the 220s BCE remains hard to corroborate. If one takes the Shanghai Museum fragment “Guishen zhi ming,” which some scholars propose to date somewhere between 340 and 230 BCE, as in any way cognate to the themes in Mozi 31, a Qin date for Mozi 31 would be hard to sustain.26 I will discuss the Shanghai Museum fragment and assess whether or not it can be linked directly to the “Ming gui” triad below.

Angus Graham , who sees ideological, rather than temporal, diffferences across the core Triplets, lists Mozi 31 among his H group chapters, that is, chapters belonging to what he calls the “Compromising ” group (Graham identifĳies an H group based on the grammatical principle that in these chapters hu 乎 replaces postverbal yu 於 where possible).27 Without the remaining two chapters in the triplet Graham ’s three-sects theory natu-rally is problematic. One might argue that the fĳinal part of Mozi 31 is more “accommodating” than the more “purist” and “reactionary” exposition of the three tests , but we have little to go on. Furthermore, Carine Defoort ’s thesis, which is based on her study of the “Jian ai” 兼愛 triplet and holds that the doctrine becomes more radical as time went by,28 would turn Graham ’s accommodation theory on its head if one were to apply it inter-nally within one chapter and surmise that the fĳinal part of Mozi 31 is of a later date than the exposition of the three tests .

The mixed distribution of the formulas 子墨子言曰 (occurring three times) and 子墨子曰 (occurring eight times) also reveals little. Bruce Brooks and Taeko Brooks submit a date for Mozi 31 toward the end of the period 390–280 BCE, which is the timeline they propose for the compo-sition of the Core Chapters . In their evolutionary model, Mozi 31 would represent a revised or updated version of the earlier chapter in the triplet, now lost.29

Karen Desmet , who has examined the dating of the Triplets according to the occurrence of compounds , concludes that the H group chapters were written last in the Triplet sequence on the grounds that they consistently

while granting only a minor role to sacrifĳices to altars of the soil and grain, and that Zhou religion emphasizes the opposite, is too simplistic as a basis for proving a Song provenance for Mohist ideas on spirits . See Zheng Jiewen, Zhongguo Mo xue tongshi, 17–24.

26 This is a criticism of Watanabe Takashi recently put forward by Asano Yūichi , who would like to date the Shanghai Museum fragment between 373 and 278 BCE. See Asano Yūichi, “ ‘Guishen zhi ming’ yu ‘Mozi: Ming gui’ ”, 97.

27 Graham, Divisions in Early Mohism; and Graham, “Mo Tzu.” See also the introduction to this volume.

28 See Defoort, “The Growing Scope of Jian 兼”; and her contribution to this volume.29 See A. Taeko Brooks, “The Mician Ethical Chapters,” 100, 118.



contain more compounds than the Y (Purist ) and J (Reactionary ) chap-ters. This would put the composition of Mozi 31, on philological grounds, around 300 BCE.30 Desmet further suggests, with Watanabe Takashi and Erik Maeder , that the earlier two chapters in the triplet may never have existed since the three chapters in which Confucian views are challenged most vociferously (Mozi 25, “Jie zang” 節葬; Mozi 31, “Ming gui” ; and Mozi 32, “Fei yue” 非樂) each represent doctrines of which only the latest, H chap-ter, survives: “All the doctrines of which only one chapter is extant (‘Fei Ru,’ ‘Jie zang,’ ‘Ming gui’ and ‘Fei yue’) are more negative doctrines, not constituting a new Mohist argument but mainly reacting against Confu-cian ideas. As the ‘Fei Ru’ doctrine itself indicates, tensions with the Con-fucians probably sharpened only later in the history of Mohism.”31 The idea that a dichotomy with “Confucians” was not clearly pronounced in the early days of the Mohist movement could tally with the paucity of the label ru in the core doctrines. But even so, the assumption that the interlocutors in the Mozi invariably are “Confucians” needs to be handled with care. Regardless of these diffferences in interpretation, the balance of received opinion would situate the composition of Mozi 31 in the period starting around 300 BCE and ending well into the Qin around 210 BCE.32 I see little room to further refĳine this dating based on internal linguis-tic or thematic arguments, other than that because of its skeptical tone the fĳinal part of Mozi 31 (following test 3) could postdate the three-theses argument.

Ghosts in the Other Core Chapters (Mozi 8–37)

The other Core Chapters in the received Mozi contain several passages that touch on the issue of ghosts and spirits . It is noticeable that the ques-tion of the existence of spirits is largely left untouched in these chapters. Instead, the focus is on the utility of a belief in the spirit world for the advancement of politics and government. In “Tian zhi, shang” (Mozi 26) the initiative to undertake sacrifĳice is presented as an afffĳirmation on

30 Desmet, “The Growth of Compounds in the Core Chapters of the Mozi.”31 Desmet, “All Good Things Come in Threes,” 240, 248–249.32 The only confĳident dissenting voice here is Fang Shouchu 方授楚, who argued that,

because of its length and relative lack of rhetoric, Mozi 31 belongs to the earliest Core Chapters and thus was composed before 329 BCE. For a summary of his arguments, see Desmet, “All Good Things Come in Threes,” 52–55.

110 roel sterckx


behalf of the ruler that he relies on the authority of Heaven . Sacrifĳice to the spirits is seen as an instrumental conduit to the power of Heaven :

天之為政於天子, 天下百姓未得之明知也˳ 故昔三代聖王禹湯文武, 欲以天之為政於天子, 明說天下之百姓, 故莫不犓牛羊、豢犬彘, 潔為粢盛酒醴, 以祭祀上帝鬼神, 而求祈福於天˳ That it is Heaven that entrusts the power to govern to the Son of Heaven is something the common people under Heaven do not yet clearly under-stand. Therefore, among the ancient sage-kings of the Three Dynasties—Yu , Tang , Wen , and Wu —wishing to make it clear to all people in the world that Heaven gives the power for government to the Son of Heaven , there were none who did not feed oxen and sheep with grass, and pigs and dogs with grain, and cleanse the sacrifĳicial grains and ale to offfer sacrifĳice to Shangdi and the ghosts and spirits and to seek and invoke the blessings of Heaven . (26: 43/2–4 , variant at 28: 48/4–8 )

A clear spirit hierarchy is set out here and throughout the “Tian zhi” triad. The ancient kings “in the highest sphere (shang 上) revered Heaven (tian 天), in the middle sphere (zhong 中) served the spirits (guishen 鬼神), and in the lower sphere (xia 下) loved the people,” while their wicked counterparts offfended at all three levels (26: 43/9–14 ).33 Since all in the world exert themselves to prepare sacrifĳices to Heaven , it follows that Heaven loves the people equally (26: 43/18–23 ). The intermediary role of the spirits fĳigures in “Tian zhi, zhong” (Mozi 27), where it is said that a ruler, in response to disease and calamities, will fast and bathe himself and prepare clean sacrifĳicial grains and wine to sacrifĳice to Heaven and the ghosts (gui 鬼). Heaven , in return, has the power to avert disasters (27: 44/17–20 ). Wealth or material well-being (cai yong 財用) is seen as a precondition for people to invest in sacrifĳices (27: 44/29–30 ). Reneg-ing on one’s sacrifĳicial commitments results in the loss of Heaven’s sup-port. Hence, King Zhou lost the support of Heaven since he “would not worship Shangdi and rejected the ancestors and spirits without offfering them sacrifĳices” (27: 46/23–24 , quoting the “Da shi” 大誓). In short, in the “Tian zhi” triad the spirit world acts as a sanctioning voice for the will of Heaven . Spirits are not the ultimate source of authority but the exclusive intermediaries of Heaven.34

33 This tripartite structure is upheld throughout the “Tian zhi” triad. It is possible but unclear whether, as Gunnar Sjöholm speculates, a spatial referent is intended here. See Sjöholm, “Readings in Mo Ti,” 107.

34 John Knoblock’s characterization of the spirits as “informants” comes to mind: “In this Mohist view disease, pestilence, and famine were penalties for contravening the



In “Shang xian, zhong” 尚賢中 (Mozi 9) the investiture of Yao , Shun , Yu , Wen, and Wu as Son of Heaven is presented as a reward from Heaven and the spirits for their impartial love of the people and their benefĳiting them . Conversely, Jie , Zhou , You 幽, and Li 厲 are punished by Heaven and the spirits for their oppressive conduct (09: 12/20–28 ). “Shang xian, xia” (Mozi 10) rehearses the familiar tripartite structure of the world: Heaven on high, the spirits in the middle, the people below (10: 14/24–25 ). In “Shang tong, zhong” 尚同中 (Mozi 12) another explicit reference is made to the need for proper sacrifĳice to prevent disasters and calamities inflicted as punish-ments from Heaven :

故古者聖王明天鬼之所欲, 而避天鬼之所憎, 以求興天下之利, 除天下之害˳ 是以率天下之萬民, 齋戒沐浴, 潔為酒醴粢盛, 以祭祀天鬼˳ 其事鬼神也, 酒醴粢盛不敢不蠲潔, 犧牲不敢不腯肥, 珪璧幣帛不敢不中度量, 春秋祭祀不敢失時幾˳ Therefore, the sage-kings of old understood what Heaven and the ghosts desire and avoided what they resent, in order to seek and increase benefĳits and avoid calamities in the world . Therefore, they led the myriad people in the world by fasting and bathing themselves and by purifying sacrifĳicial wine and grains to sacrifĳice to Heaven and the ghosts. When they were serv-ing ghosts and spirits , they did not dare to use sacrifĳicial wine and grains that were unclean, sacrifĳicial animals that were not fat, or sacrifĳicial jades or silks that did not meet the standard measure . They did not dare to miss the proper timing for spring and autumn sacrifĳices . (12: 18/30–19/2 )

Although the ideology that motivates sacrifĳice here has a defĳinite Mohist ring, it is equally true that, as far as the mechanics of sacrifĳice are con-cerned, the requirements described here—purifĳied and glossy offferings—are common across most Warring States texts.35 Interestingly, the “Jian ai” triad invokes the image of sacrifĳice to implore the spirits on only one occasion, in the case of King Tang ’s famous self-sacrifĳice (16: 29/7–9). “Fei gong, xia” 非攻下 (Mozi 19 ) condemns siege warfare on the grounds that it destroys the material provisions for sacrifĳices :

intentions of Heaven and, as such, did not occur in an orderly age. Ghosts and spirits were informants of unseen misdeeds as well as avenging agents of a wrathful Heaven.” See Knoblock, Xunzi, vol. 3, 6.

35 See Sterckx, “The Economics of Religion in Warring States and Early Imperial China”; and Sterckx, Food, Sacrifĳice, and Sagehood in Early China, chap. 4.

112 roel sterckx


夫取天之人, 以攻天之邑, 此刺殺天民, 剝振神之位, 傾覆社稷, 攘殺其犧牲, 則此上不中天之利矣˳ 意將以為利鬼乎？夫殺之神, 滅鬼神之主, 廢滅先王, 賊虐萬民, 百姓離散, 則此中不中鬼之利矣˳ Yet to gather the people of Heaven in order to besiege the towns belong-ing to Heaven is akin to murdering Heaven’s subjects , to dispossessing the spirits of their tablets, ruining the altars of the soil and grain, and violating and killing their sacrifĳicial animals. This then does not offfer any benefĳits to Heaven up there. Would its intention then be to bless the ghosts? But the people of Heaven are murdered,36 the tablets of ghosts and spirits are destroyed, the former kings are neglected, the myriad people are tortured , and the common people are scattered; it is then not for the benefĳit of the ghosts in the middle realm.37 (19: 33/28–30 )

This third chapter in the triplet is, as Paul van Els has shown in his con-tribution to the present volume, mostly preoccupied with a theologi-cal argument against war.38 Mozi 19 further calls on avenging spirits as punishing agents as an alternative for “offfensive war.” The “Jie yong” and “Jie zang” chapters make no specifĳic reference to the spirit world. In “Jie zang, xia” (Mozi 25), Mozi’s objection to elaborate burials and extended periods of mourning is basically inspired by economics : such practices stand in the way of wealth creation and prevent the increase of the popu-lation.39 Mozi also makes a point regarding cult: when a state becomes poor as a result of overinvestment in funerals and mourning, sacrifĳicial goods will decline in quality, fewer people will have the means to offfer sacrifĳice , and hence the normal sacrifĳicial cycle will be disturbed. As a result, the spirits are upset, punish people with calamities, and abandon them (25: 40/1–8 ). Thus, Mozi marries the idea of sacrifĳicial duty with

36 Emending 夫殺之神 to 夫殺[天]之人 following Sun Yirang , Mozi jiangu, 143.37 Benjamin Wong and Hui-Chieh Loy argue that Mozi’s political philosophy “stands

or falls with the success of his attempt to prove the existence of providential ghosts .” See Wong and Loy, “War and Ghosts in Mozi’s Political Philosophy,” 344. Although warfare, in the passage above, is presented as sacrilegious in that it harms Heaven, the spirits, and the people, the specifĳic mechanism used by the spirits when “rewarding and punishing” does not draw much comment in the “Fei gong” triad, something one might expect if it was central to the thesis. Wong and Loy may be reading too much into the Mohist reliance on ghosts as backup for moral action.

38 The compound guishen or variants thereof occur eleven times in Mozi 19 , as opposed to only once in Mozi 18 and not at all in Mozi 17. Mozi 17 puts forward a predominantly ethical argument; Mozi 18 takes a mostly economic/utilitarian stance. See Paul van Els’s essay in this volume.

39 For a more detailed treatment of this theme, see Riegel, “Do Not Serve the Dead as You Serve the Living.”



funerary economies and suggests that sacrifĳices should not be discon-tinued during the mourning period:

哭往哭來, 反從事乎衣食之財, 佴乎祭祀, 以致孝於親˳ 故曰: 子墨子之法, 不失死生之利者, 此也˳ You may weep on the way to and from a funeral . Yet, once returned, you should engage in earning your livelihood (clothes and food). You should actively continue to perform sacrifĳices in order to express fĳilial piety toward your kin. Therefore, it is said that the rules of Our Master Mozi neglect the benefĳits of neither the dead nor the living. (25: 41/29–30 )

“Fei yue, shang” (Mozi 32) insists that Mozi condemns music (and ban-queting), not because it is aesthetically displeasing, but solely on the grounds of the material and human resources required to produce and play instruments. In “Fei ming, shang” (Mozi 35) Mozi puts righteousness over a belief in fate:

義人在上, 天下必治, 上帝山川鬼神必有幹主, 萬民被其大利˳ When the righteous are in power, the world will have order; Shangdi, hills and rivers, ghosts and spirits will have their lineage elders [to preside over sacrifĳices ];40 and the people will be visited by the great blessings resulting from this. (35: 59/5–6 )

A misguided belief in fate is said to result in a chaotic government and “then, there would be no means to provide sacrifĳicial grain and wine to sacrifĳice to Shangdi, the ghosts , and the spirits ” 上無以共粢盛酒醴, 祭祀上帝鬼神 (35: 60/8 ).

Overall then the relevant passages in the other Core Chapters empha-size the social and political role of the spirit world. The general theme in these chapters does not depart substantially from the theses set out in Mozi 31, but the emphasis is more on practice—more specifĳically, the need to fulfĳill sacrifĳicial obligations. The philosophical issue of the exis-tence of ghosts and spirits (wu/you 有/無) is not engaged.

The Opening Chapters and Dialogues

To what extent are views on ghosts and spirits and references to sacrifĳi-cial practice in Mozi 31 and the other Core Chapters consistent with or

40 Reading gan zhu 幹主 as meaning zong zhu 宗主 following Sun Yirang , Mozi jiangu, 268.

114 roel sterckx


cognate to materials found in the Opening Chapters (Mozi 1–7) and Dia-logues (Mozi 46–49/50)?

It is noteworthy that the spirit world does not fĳigure prominently at all in the Opening Chapters as a subject of discussion (the graphs 鬼 and 神 occur only twice each across the seven chapters). “Qin shi” 親士 (Mozi 1) mentions divination and sacrifĳice , but it does so only as part of a meta-phor noting that those who stand out through their talents are often more likely to fĳind themselves in peril than others: “The numinous tortoises are burnt fĳirst, and the snakes that have spirit power are sacrifĳiced fĳirst” (1: 1/21 ).41 In “Fa yi” 法儀 (Mozi 4), people’s inclination to prepare sac-rifĳicial offferings to Heaven and Heaven’s impartial acceptance of such offferings are highlighted to reinforce the notion of an impartial Heaven (4: 5/1–5 ). “Qi huan” 七患 (Mozi 5) reiterates the need to moderate sacrifĳices in times of famine, a theme found across other late Warring States texts (5: 5/17–24 ).42 If the Opening Chapters are indeed of fairly late origin (mid- to late third century BCE) and hence postdate Mozi 31 and the other Core Chapters , it is striking that a theme so prevalent in the Core Chapters occurs so sparsely in the Opening Chapters . This confĳirms the miscella-neous nature of the Opening Chapters themselves and further suggests that thematic continuities should not necessarily be inferred on the basis of the purported dating of chapter groups in the Mozi, especially since the dating is so uncertain.

It is in the Dialogues (specifĳically, Mozi 46–49) that some of the doctri-nal uncertainties beneath the surface of Mozi 31 and, as we will see, the Shanghai Museum fragment, are on the agenda again. The Dialogues are exchanges between Mozi and his disciplines and opponents . Most schol-ars consider the Dialogues to postdate the Core Chapters on the grounds that the protagonists in these chapters are second- or third-generation disciples of Mozi. It is amid the generally more polemical nature of the exchanges in these chapters that the identity of Mozi’s interlocutors is made more explicit. Several passages in the Dialogues are of particular relevance to the issues debated in Mozi 31: an exchange between Mozi and

41 The reference here is to (rhyming) zhuo/*tiauk 灼, meaning “to stick a hot poker through a turtle carapace,” and pu/*bâukh 暴, a technical term for using a snake in a rain-making sacrifĳice.

42 For appeals to change one’s diet and practice moderation in sacrifĳice during times of natural disaster and famine, see, e.g., Zuozhuan “Lord Cheng, year 5” and “Lord Zhao, year 17”; Guliang zhuan “Lord Xiang, year 24”; Yi Zhoushu “Di Kuang” and “Da kuang jie”; Liji “Qu li” and “Yu zao”; and Zhouli “Shan fu.”



Wumazi 巫馬子 in Mozi 46, several exchanges with Gongmengzi 公孟子 in Mozi 48, and a couple of passages on sacrifĳice in Mozi 49.

“Geng zhu” 耕柱 (Mozi 46) contains a long interchange between Mozi and Wumazi . Commentators traditionally identify this fĳigure as one of Confucius ’s disciples , also known as Wu Maqi 巫馬期. Although his exact identity remains uncertain, his stance in Mozi 46 would certainly befĳit that of a Ru (yet note that the chapter does not explicitly label him a Ru). In the exchange in question Mozi indicates that the perceptive quali-ties of ghosts and spirits are so superior that, were the sages able to avail themselves of them as sensory extensions, they would undoubtedly do so. Sages are represented as bound by time and circumstance, whereas the knowledge of ghosts and spirits is said to transcend time:

巫馬子謂子墨子曰: 「鬼神孰與聖人明智？」子墨子曰: 「鬼神之明智於聖人, 猶聰耳明目之與聾瞽也˳ 昔者夏后開使蜚廉折金於山川, 而陶鑄之於昆吾, 是使翁難卜於白若之龜, 曰: 『鼎成三足而方, 不炊而自烹, 不舉而自臧, 不遷而自行, 以祭於昆吾之墟, 上鄉！』卜人言兆之由曰: 『饗矣！逄逄白雲, 一南一北, 一西一東, 九鼎既成, 遷於三國˳ 』夏后氏失之, 殷人受之˳ 殷人失之, 周人受之˳ 夏后、殷、周之相受也, 數百歲矣˳ 使聖人聚其良臣與其桀相而（諫）〔謀〕, 豈能智數百歲之後哉？而鬼神智之˳ 是故曰: 鬼神之明智於聖人也, 猶聰耳明目之與聾瞽也˳ 」Wumazi spoke to Our Master Mozi: “Who is more perspicacious and knowl-edgeable—ghosts and spirits , or sages?” Our Master Mozi replied: “The per-cipience and knowledge of ghosts and spirits is to the sages what perceptive ears and keen eyes are to the deaf and the blind. In ancient times, King Kai [Qi] of the Xia commissioned [his minister] Feilian to dig minerals in moun-tains and rivers and cast them into vessels at Mount Kunwu. He ordered [diviner] Wengnan [Yi] to divine with the tortoise of Boruo, saying:

‘Let the tripods, when complete, be three-legged and square (方 *paŋ).43Let them be able to cook by themselves, without fĳire (烹 *phrâŋ),to store themselves away without being lifted (臧 *tsâŋ),and move themselves without being carried (行 *grâŋh).Use them for sacrifĳice at Mount Kunwu (虛[墟] *kha).may our offferings be appreciated (鄉 *haŋʔ)!’

Then the oracle spoke:‘I have accepted the offferings. Profuse are the white clouds:At one time they are in the south, then to the north, now to the west, then to the east.

43 For reconstructions of Old Chinese I follow the OCM (Minimal Old Chinese) recon-structions in Schuessler, Minimal Old Chinese and Later Han Chinese.

116 roel sterckx


The nine tripods having been completed, they will be transferred among the three states.’

When the ruler of Xia lost them, the founder of Yin inherited them. When Yin lost them, Zhou inherited them. Now, the transfer from Xia to Yin and Zhou took several hundred years. Even if a sage planned to gather his best ministers and superior assistants to plan, how could he know events that would happen after several hundred years? Ghosts and spirits, however, can know this. And it is for this reason that I hold that the perspicacity and knowledge of ghosts and spirits is to the sages what perceptive ears and keen eyes are to the deaf and the blind.” (46: 100/7–14 )

The divide between Mozi and his interlocutor here is a disagreement on the highest source of authority . For Wumazi, ultimate power of judgment lies with the sages, whereas for Mozi it lies with ghosts and spirits , who act as informants of Heaven that surpass the powers of divination. Later on in the chapter the same Wumazi points out to Mozi that for all his righteous conduct, people do not seem to help him and ghosts do not seem to bring him blessings (46: 100/30–31 ).

In “Gui yi” 貴義 (Mozi 47) Mozi defends the idea that people of humble social station are potential sources of good counsel by referring to the spirit world’s universal acceptance of sacrifĳice : “Now the peasant pays his taxes to his superior, who [uses these to] make sacrifĳicial wine and grain offferings and, with these, sacrifĳices to Shangdi, ghosts , and spirits . Can you claim that [the spirits] would refuse these offferings because they come from people of low standing?” 今農夫入其稅於大人, 大人為酒醴粢盛, 以祭上帝鬼神, 豈曰『賤人之所為』, 而不享哉？ (47: 104/6–7 ). The subject of criticism here is the idea that because the spirit world mir-rors human hierarchies, spirits would therefore reward those who offfered them sacrifĳices by reinforcing those partial social relationships that are based on lineage and social station.

The “Gongmeng” chapter (Mozi 48) generally tallies closely with the key themes in the Core Chapters (and, as we will see, the Shanghai Museum fragment). Unlike Mozi 31, Mozi 48 is clearly set up as a polemic directed against followers of Confucius . Gongmengzi 公孟子 is generally accepted to be a Ru, possibly a disciple of Zengzi 曾子. In Mozi 48 Gongmengzi mentions Confucius by name, Mozi several times disapproves of Con-fucius and the fact that he is extolled as an exemplar, and the Ru are condemned explicitly on four counts: (1) their belief that Heaven is not intelligent (ming) and that ghosts do not have spirit powers; (2) their adherence to the need for elaborate funerals ; (3) their esteem for music and dance; and (4) their fĳirm belief in predestined fate (48: 109/4–8). The



fĳirst theme in the Mozi 48 dialogues that echoes Mozi 31 is the idea that ghosts and spirits control reward and punishment and that antiquity cements the authority of arguments (cf. the second test in Mozi 31):

子墨子曰: 「古者聖王皆以鬼神為神明, 而為禍福, 執有祥不祥, 是以政治而國安也˳ 自桀紂以下, 皆以鬼神為不神明, 不能為禍福, 執無祥不祥, 是以政亂而國危也˳ Our Master Mozi said: “The ancient sage-kings all regarded ghosts and spir-its to be divine and perspicuous and capable of meting out calamity and blessing. They held that [ghosts and spirits ] controlled good fortune and misfortune and that, by means of these, governments were well adminis-tered and states were secure. From Jie and Zhou down [however], they all took ghosts and spirits not to be divine and perspicuous and unable to mete out calamity and blessing. They held that [ghosts and spirits ] did not control fortune and misfortune, and for this reason, governments became chaotic and states were in danger.” (48: 108/1–3 )

A second connection with Mozi 31 is that the Mozi fĳigure in Mozi 48 returns to the philosophical question of the existence of a spirit world. Mozi does this in a much-quoted analogy arguing that the insistence of the Ru on ritual and sacrifĳice is valid only if they accept the basic proposi-tion that spirits actually exist:

子墨子曰:「執無鬼而學祭禮, 是猶無客而學客禮也, 是猶無魚而為魚罟也˳」Our Master Mozi said: Holding that there are no ghosts and yet studying sacrifĳicial rituals is like learning the rites of hospitality when there are no guests or making fĳishing nets when there are no fĳish. (48: 108/17–18 )

Mozi 48 also revisits other relevant themes treated in the Core Chapters , such as excess expenditure on music and burials.

As I pointed out above, where Mozi 48 difffers from Mozi 31 is in its explicit reference to the Ru as advocates of the ideas that the Mohists oppose. We need to remain cautious therefore not to infer that the tenor of the debate in both chapters is exactly the same. In Mozi 48 the Ru are said “to believe that Heaven is unintelligent and that ghosts do not have spirit powers” 儒以天為不明, 以鬼為不神 (48: 109/4–5 ). This is arguably a rather tendentious interpretation of the Ru position since it would efffectively mean that, for the Ru, spirits can be dismissed as hav-ing no numinous power at all.44 Yet the Mozi 48 claim that ghosts do not

44 A point also questioned in Lu Jianhua, “Mozi zhi li xue,” esp. 40–41.

118 roel sterckx


have spirit powers (鬼不神) is not quite the same as the argument put forth by Mozi’s interlocutors in Mozi 31—namely that ghosts do not exist (執無鬼). The inference in Mozi 48 is that the Ru allow for the existence of ghostly beings but that they simply deny the spirits any signifĳicant numinous powers. The philosophical issue of the existence of spirits is sidestepped in favor of a debate on their efffĳicacy as moral agents. Finally, as in the fĳinal part of Mozi 31, Mozi 48 contains two exchanges where Mozi is put on the spot and asked again to clarify the validity of his own thesis. Both episodes start offf with the same premise, expressed by an unnamed interlocutor in the fĳirst story and by someone named Die Bi 跌鼻 in the second: “Sir [Mozi], you hold that ghosts and spirits are intelligent, that they are able to bring about misfortune and good fortune, give blessings to those who do good, and inflict misfortune on those who do bad.”45 How then, the interlocutor in the fĳirst story argues, is it tenable that he has not received any blessings after having served Mozi for many years? Why then, Die Bi in the second story argues, can Mozi, being an exemplary sage himself, be prone to illness (48: 109/30 )? In his answers Mozi seems to suggest that ghosts and spirits do control knowledge and have agency to act upon events but that there are many causes and efffects that escape their control.

“Lu wen” 魯問 (Mozi 49) contains two passages that are of interest to our understanding of the Mohist stance on cult. Both discussions focus on the relationship between the materiality of offferings and the moral intent behind the act of offfering sacrifĳice . The issue is put to Mozi by a disciple who meets him after a three-year interval:

「始吾游於子之門, 短褐之衣, 藜藿之羹, 朝得之則夕弗得, 弗得祀鬼神˳ 今而以夫子之故, 家厚於始也, 有家厚謹祭祀鬼神˳ 然而人徒多死, 六畜不蕃, 身湛於病, 吾未知夫子之道之可用也˳」子墨子曰: 「不然, 夫鬼神之所欲於人者多, 欲人之處高爵祿則以讓賢也, 多財則以分貧也, 夫鬼神豈唯擢黍拑肺之為欲哉˳ 今子處高爵祿而不以讓賢, 一不祥也˳ 多財而不以分貧, 二不祥也˳ 今子事鬼神, 唯祭而已矣, 而曰:『病何自至哉？』是猶百門而閉一門焉, 曰:『盜何從入？』若是而求福, 於有怪之鬼, 豈可哉？」

When I fĳirst came to your gate I had to wear short-sleeved jackets and eat vegetable broth. If I ate it in the morning, I could not have it again in the evening, and I had nothing to sacrifĳice to the ghosts and spirits . Now

45 In the fĳirst story: 先生以鬼神為明知, 能為禍福, [為]善者富之, 為暴者禍之 (48: 109/19–20 ); variant in the second: 先生以鬼神為明, 能為禍福, 善者賞之, 為不善者罰之 (48: 109/29–30 ).



because of your teachings, my family is better offf than we were in the beginning. That being the case, I can respectfully offfer sacrifĳice to ghosts and spirits. Yet, still, many among my close followers have died, my six [types of ] domestic animals do not breed, and I have personally been plagued by ailments. I have yet to be convinced [therefore] that your way is after all to be adopted.

Our Master Mozi replied: “This is not how things work, for the things ghosts and spirits desire of men are manifold. They would wish that when people enjoy high rank and salary, they would give up their position in favor of the worthy ; and that someone with great wealth would share it with the poor. How then could ghosts and spirits merely desire to seize sacrifĳicial grains or snatch away sacrifĳicial lungs?46 Now, you were enjoying high rank and sal-ary, yet you did not give up your position in favor of a worthy. This is your fĳirst misfortune. Being very wealthy you failed to share with the poor. This is your second misfortune. And so now you serve ghosts and spirits merely by offfer-ing them some sacrifĳices while asking yourself, ‘Why do these diseases inflict themselves on me?’ This is like shutting one gate out of a hundred while wondering, ‘Where do these thieves enter from?’ In such a situation how can you seek blessings from these bewildered ghosts ?”47 (49: 114/12–19 )

Contra suggestions that Mohists are merely interested in perfunctory do-ut-des worship, the Mozi fĳigure in this dialogue suggests that appeasing spirits through sacrifĳice without an underlying moral or ethical intent, or indeed, without putting a moral agenda into practice in daily life, is inefffective and lacking in substance. Ghosts are not interested merely in the volume or quality of sacrifĳicial offferings. They continue to watch and assess the conduct of the supplicant and value the integrity that under-pins the presentation of offferings, a message that appears to have escaped Mozi’s own disciple after three years of absence. The argument could be interpreted as a veiled criticism of the Confucian notion that one should respect and offfer sacrifĳice to the spirit world in order to keep it at bay, or their insistence that one should serve spirits “as if ” (ru 如) they are present (祭神如神在). This passage is followed by an anecdote on a pig sacrifĳice that touches on the same theme and to which I will return in the fĳinal section of this essay.

To sum up, the Dialogues offfer fertile ground for tracing debates on some of the central themes of Mozi 31 (note that Mozi 50 adds nothing

46 Liji 26 “Jiao te sheng” 郊特牲 suggests that lungs were valuable organs in sacrifĳice since they were thought to be rich in blood and qi.

47 The expression 有怪之鬼 is slightly puzzling. One could opt to emend it (e.g., John-ston, The “Mozi,” 716 n. 30, reads 靈 for 怪). But it can equally well mean ghosts that harbor suspicion for or that mistrust people who offfer them only partial service—hence, ghosts that are “bewildered.”

120 roel sterckx


new). Yet nothing in the passages discussed above suggests that we are dealing with a direct commentary-type or discursive elaboration of the received Mozi 31. While the discussions are formatted more clearly as a Mo-Ru polemic, it is equally clear that not only may we have diffferent Mozi personas in action but also that Mozi’s core thesis on spirits was the subject of various interpretations and even questioned by Mohists them-selves. Mozi’s position on ghosts appears to be controversial even among his own followers, and likewise, the explicit or implicit Ru counterargu-ments do not speak with one voice. Rather than a sustained commentary on a doctrinal maxim, what emerges is a picture of multiple arguments by multiple actors concerning a thesis that continuously requires clarifĳica-tion. To add further to this diverse polemical landscape, it is now time to bring the Shanghai Museum fragment into our discussion.

The Shanghai Museum Fragment

A brief text fragment among the Chu bamboo slips held at the Shanghai Museum could throw further doubt on one of the central theses in the received Mozi 31, namely the notion that ghosts will invariably punish the wicked and reward those who do good. While the fragment in question possibly adds a signifĳicant new dimension to the Mohist view of spirits, we also ought to be cautious not to extrapolate too much from what is, in essence, a very short manuscript consisting of only fĳive slips and a total of 197 characters. The text was dubbed “Guishen zhi ming” 鬼神之明 by Cao Jinyan 曹锦炎, its editor for the Shanghai Museum series, and it is usually referred to by this working title since its publication in 2005.48 Cao Jinyan , Asano Yūichi 淺野裕一, and several other scholars have linked this fragment to the received Mozi and postulate that it is part of the lost fĳirst two chapters in the “Ming gui” triad.49 Cao takes the text as evidence for the signifĳicant presence of Mohists in the state of Chu, a hypothesis, he

48 Cao Jinyan, “Shanghai bowuguan cang Chu zhushu Mozi yiwen.” I will follow the text as transcribed by Cao Jinyan with emendations and variants suggested by others indicated in notes. For Cao’s text, see Shanghai bowuguan cang Zhanguo Chu zhushu, vol. 5, 149–159 (plates), 307–320 (transcription). Studies that include a transcription and annotations include Asano Yūichi, “ ‘Guishen zhi ming’ yu ‘Mozi: Ming gui’ ”; Xu Hua, “Shang bo jian ‘Guishen zhi ming’ yi wei ‘Dongzi’ yiwen”; Li Rui, “Lun Shang bo jian ‘Guishen zhi ming’ pian de xuepai xingzhi”; Wang Zhongjiang, “Guishen zhi ming yu Dong Zhou de ‘duo yuan guishen guan’ ”; Nishiyama Hisashi, “Shang bo Chu jian Guishen zhi ming de sige bushi”; and Chen Wei, “Shang bo wu ‘Guishen zhi ming’ pian chu du.”

49 Cao Jinyan, “Shanghai bowuguan cang Chu zhushu Mozi yiwen,” 57.



claims, that is strengthened by the discovery of a text fragment possibly linked to the Mozi from Chu Tomb 1 at Changtaiguan 長台關 (Xinyang 信陽, Henan, excavated 1956).50 Asano , who dates “Guishen zhi ming” between 342 and 282 BCE, adduces it as proof for his thesis that the ten core doctrines are to be dated to the early Warring States period, that is, during the latter half of Mozi’s life.51

There are a number of good reasons for associating “Guishen zhi ming” with a Mohist polemic, given its thematic similarities with some of the passages in the received Mozi described above. However, the question of whether or not this piece should be text-genetically linked to Mozi 31 can-not be answered conclusively. It is important not to infer a direct textual afffĳiliation with Mozi 31 based on the provisional title assigned by Cao Jin-yan , which does not appear on the manuscript itself.52 Scholars have sug-gested several other titles , but given that there is no evidence of an actual title on the physical manuscript itself, I will continue to refer to it simply as the Shanghai Museum fragment.53 The opening part of the original text is missing (as is suggested by the connecting expression jin fu 今夫 at the beginning of the fragment). A dialogue setting seems likely,54 but no interlocutors are identifĳied by name:

50 The fragment in question is referred to as the “Shentu Di” 申徒狄 and contains a conversation between the Duke of Zhou and Shentu Di, who famously committed suicide rather than live an immoral existence. Its direct link to the Mozi, however, is controversial. See Li Xueqin, “Changtaiguan zhujian zhong de Mozi yi pian”; and Wang Ning, “Mozi ji Shentu Di shizheng.”

51 Asano Yūichi, “ ‘Guishen zhi ming’ yu ‘Mozi: Ming gui,’ ” 100–101.52 The preserved text runs continuously over fĳive slips (complete slips measure ca. 53

centimeters and were tied together in three registers). While it is likely that some content (including, possibly, an actual title ) preceded these fĳive slips, the end of the fragment must coincide with the end of a pian since a new unit (entitled “Rong shi youcheng shi” 融師有成氏) follows on slip 5 and is marked offf by a separation mark ▃ in black ink (see Shanghai Bowuguan cang Zhanguo Chu zhushu, vol. 5, pl. 5, p. 156). Translations by Erica Brindley and Ken Brashier appeared while my draft was with the editors . See Brindley, “‘The Perspicuity of Ghosts and Spirits’ and the Problem of Intellectual Afffĳiliations in Early China”; and Brashier, Ancestral Memory in Early China, 338–339. My translation broadly agrees but difffers in detail and modalities.

53 Liao Mingchun suggests that, given its central argument, the text should be known as “Guishen you suo ming you suo bu ming” 鬼神有所明有所不明. See Liao Mingchun, “Du ‘Shang bo wu: Guishen zhi ming’ pian zhaji.” Ding Sixin simply dubs the text “Guishen” 鬼神. See Ding Sixin, “Shang bo Chu jian Gui shen pian zhushi”; and Ding Sixin, “Lun Chu jian ‘guishen’ pian de guishen guan ji qi xuepai guishu.”

54 I concur with Erica Brindley , “ ‘The Perspicuity of Ghosts and Spirits’ and the Prob-lem of Intellectual Afffĳiliations in Early China,” that the style of the fragment does not fĳit the dialogical format of the Mozi–disciple exchanges in Mozi 48, but I would suggest that the fragment is simply too short to exclude the possibility that it is a dialogue altogether. Self-reflective and expository moments can be embedded in a dialogue, as can rhetorical

122 roel sterckx


Slip 1今夫鬼神又（有）所明又（有）所不明, 則（以）亓（其）賞善罰暴也˳ 昔者堯舜禹湯, （仁）義聖智, 天下灋 (法) 之˳ 此（以）貴爲天子,

Now, there are things that ghosts and spirits are aware of but equally other things that they are unaware of, and this [can be derived from] their rewarding of the good and punishing of the wicked. In ancient times Yao , Shun , Yu , and Tang were benevolent and righteous and had sagely knowledge,55 and thus, All under Heaven took them as a model. Because of this they were honored to become Son of Heaven ,

Slip 2富又（有）天下, 長年又（有） (譽),56 後世遂之˳ 則鬼神之賞, 此明矣˳ 及桀受（紂）幽厲, 焚聖人, 殺訐者, 賊百眚（姓）, 亂邦家 ˳—[此（以）桀折於鬲山, 而受（紂）首於只（岐）社],57 身不沒, 爲天下笑˳ 則鬼

questions. Regardless, the formula 今夫 suggests that the narrator here is disagreeing with a stance taken previously.

55 Cao Jinyan gives the gloss ren/*nin 仁 for based on the phonetic resemblance of *lhin 身 and *tshîn 千, 忎, the latter being an old form of 仁 in Shuowen jiezi 8A.2a. Nishiyama Hisashi , “Shang bo Chu jian Guishen zhi ming de sige bushi,” suggests pairing renyi 仁義 and shengzhi 聖智 as binomes based on comparative use in Mozi 28, other Ru texts, and the Zhuangzi . He argues that both binomes are stock expressions used by both Ru and Mo (and criticized in Daoist texts) and takes this as an element in favor of a close afffĳiliation between the Shanghai Museum fragment and the received Mozi (in which the binome 仁義 occurs more than twenty times).

56 I read as yu 譽 following Liao Mingchun , “Du ‘Shang bo wu: Guishen zhi ming’ pian zhaji,” rather than ju 擧 as proposed by Cao Jinyan , although their meaning is similar.

57 The thirteen characters between square brackets were written on the bottom half of the back of the second slip and are inserted into the main text by Cao Jinyan ; for the plate, see Shanghai bowuguan cang Zhanguo Chu zhushu, vol. 5, 316. There is nothing that sug-gests that it was written by a diffferent hand. Cao suggests that this is a case of haplography in which the scribe or copyist added on the back of the slip a part of the text overlooked while recording or copying the main text. The black ink mark ▃ could indicate where on the slip the omitted passage should be inserted. There is another possibility: what is written on the back of the slip may be an explanatory comment, almost equivalent to an example in a footnote, since the main text could make sense without the inserted thirteen graphs if one takes the phrase 身不沒 to apply to all four tyrants mentioned earlier.



were enriched with All under Heaven ,58 and lived out their long lives in praise,59 and later generations followed them. And so, that ghosts and spirits rewarded them is evident [from these examples].60 When we come to the kings Jie , Zhou , You, and Li, they burned61 sages to death, killed those who exposed faults, robbed the hundred families, and inflicted chaos on the states. [Because of this, Jie was cut down on Mount Li, and Zhou received his punishment at the Qi Altar Mount.]62 Their bodies did not die a natural death,63 and they became an object of ridicule to the world. And so, that ghosts

Slip 3[神之罰, 此明]64 矣˳ 及五（伍）子疋（胥）者, 天下之聖人也, 鴟夷而死˳ 榮夷公者, 天下之亂人也, 長年而沒˳ 女（如）（以）此詰之, 則善者或不賞, 而暴

58 According to Nishiyama Hisashi , the phrase 貴爲天子, 富有天下, or variations thereof, is attested eleven times in the received Mozi, a statistic that could lend support to textual afffĳiliation. See Nishiyama Hisashi, “Shang bo Chu jian Guishen zhi ming de ‘gui wei tianzi, fu you tianxia.’ ”

59 The term chang nian 長年 occurs twice in the fragment. Liao Mingchun and Cao Jinyan would like it to mean “a long time,” “many years,” in slip 2 and hence take it to have a slightly diffferent meaning from that of “long life,” “old age,” on slip 3. Ding Sixin and Nishiyama Hisashi opt for “long life” in both cases. The contrast here is clearly with the evil quartet that follows on slip 2, who failed to live to a natural old age.

60 One could take ming 明 here in its more pregnant meaning and translate: “this is due to their clear percipience.”

61 Liao Mingchun, “Du ‘Shang bo wu: Guishen zhi ming’ pian zhaji,” emends fen 焚 (*bэn) to fen 僨, meaning jiang 僵, “to collapse dead” or, more generally, “to kill” (rather than burn to death). His gloss does not substantially change the meaning of the sentence. The reference here is most likely to the murder of Bi Gan 比干 by King Zhou and Jie’s killing of Guan Long(p)feng 關龍逢, a loyal and upright minister. See Han Feizi “Nan yan” 難言 and “Ren zhu” 人主.

62 There appears to be no received record of Jie coming to his end at a Mount Li 鬲 (*rêk) other than in another manuscript among the Shanghai Museum corpus (“Rongcheng shi” 容成氏). A lost Shizi 尸子 fragment, however, states that Jie was banished to a Mount Li 歷 (*rêk), which is a homophone. See Shizi yizhu 72 (no. 68, quoting Taiping yulan 太平御覽). Qi She 岐社 must refer to the mountain site (Qi shan 岐山) and geographical heart of the Zhou polity where the (Western) Zhou ancestral altars were located. See Sun Yirang, Mozi jiangu, 152 (“Fei gong, xia”). See also Li Feng, Landscape and Power in Early China, 46–47.

63 Transcribed to read mo 没 following Shuowen jiezi 3B.19b, 11B.23a–b. I follow Cao’s suggestion here that it must mean failing to live out one’s life to the end, in contrast to the virtuous quartet above.

64 Note that these fĳive graphs have been added by the editors based on context; they are missing on the original slip.

124 roel sterckx


and spirits mete out punishment is evident [from these cases]. [However,] when we come to Wu Zixu, he was one of the world’s sage exemplars, yet he ended up in a leather sack and died. Yigong of Rong was one of the world’s troublemakers, yet he lived to old age and died a natural death.65 If one examines the issue66 based on these cases, then [one must con-clude that] among the good there are some who do not receive rewards, and among the wicked

Slip 4[ 者或不罰˳ 古（故）]67 吾因加 68 “鬼神不明”, 則必又（有）古（故）˳亓（其）力能至（致）安（焉）而弗爲

65 Wu Zixu sought refuge in Wu in 522 BCE, when his father and elder brother were unjustly executed in Chu. Eleven years later he persuaded King Helü 闔閭 of Wu (r. 514–496 BCE) to attack Chu to avenge his father and brother. After the death of King Helü, Wu Zixu served the young King Fuchai(i) 夫差 (r. 496–473 BCE) and helped him win a war against King Goujian 勾踐 of Yue (r. 496–465 BCE). He then fell from grace and was forced to commit suicide. The story occurs in many sources. See Johnson, “Epic and History in Early China.” The identity of the second fĳigure in the manuscript is debated. Cao proposes Rong Yigong 榮夷公 (also known as Rong Yizhong 榮夷終), who is attested as an adviser who had a bad influence on King Li 厲 of the Zhou (857/53–842/28 BCE). Mozi 3 refers to 榮夷終 (cf. Sun Yirang, Mozi jiangu, 13, “Suo ran” 所染); see also Lüshi chunqiu 2.95 (“Dang ran” 當染). Others have suggested Duke 穆 of Song 宋 (Song Mu Gong; r. 728–720 BCE). See Yang Zesheng, “Shuo Shang bo jian ‘Song Mu Gong zhe, tian xia zhi luan ren ye.’” In a recent article jointly authored with Li Jiahao, Yang has reviewed his original identifĳica-tion on the grounds that (1) there are no clear references indicating that Song Mu Gong lived a long life, and (2) Song Mu Gong was not as famous as Wu Zixu. Instead, Yang and Li now propose that the individual mentioned on the slip is Duke Mu of Qin 秦穆(繆)公(r. 659–621 BCE). To add weight to the relationship between the Shanghai Museum frag-ment and Mozi 31, Yang and Li present a philological hypothesis that could explain why Qin 秦 appears as Zheng 鄭 in the second witness account of test 1 in Mozi 31. To support their identifĳication they adduce references that indicate that Qin Mu Gong could have lived beyond the age of sixty-fĳive and list four incidents on account of which Qin Mu Gong could have been perceived to misbehave (as implied in the manuscript). See Yang Zesheng and Li Jiahao, “Tan Shang bo zhushu ‘Guishen zhi ming’ zhong de ‘Song Mao Gong.’”

66 I follow Li Rui, Nishiyama Hisashi 西山尚志, Chen Wei 陈伟, and Wang Zhong-jiang in glossing 女 as the conditional 如; Cao Jinyan and Asano Yūichi opt for the per-sonal pronoun 汝, which is possible (the modality of the clause is most likely an implied conditional anyway). Note, however, that the second-person pronoun 汝 does not occur in the received Mozi except in a dubious quotation at Mozi 19: 34/25, 19: 34/27, and 19: 35/5. Whatever option one chooses, to infer that a 汝 reading sheds greater clarity on the agents in the dialogue here (吾 vs. 汝), let alone that we might have a dialogue here between a fĳirst-person Mozi or Mohist grand master and a disciple, is highly speculative. Cf. Ding Sixin, “Lun Chu jian ‘guishen’ pian de guishen guan ji qi xuepai guishu,” 409–412.

67 The fĳirst four graphs between brackets are missing and have been supplied by Cao Jinyan based on context; the lower 口 part of the fĳifth character 古 (= 故) is visible on the slip.

68 I have followed Liao Mingchun, who suggests not to follow Cao Jinyan’s emendation to *krâi 嘉 (“to speak favorably of,” “to commend”) but instead to take 因加 as a binome,



唬（乎）？吾弗智（知）也; 意亓（其）力古（固）不能至（致）安（焉）唬（乎）？吾或 (又) 弗智（知）也˳ 此兩者枳（歧）˳ 吾古（故）

some do not receive punishment. Therefore, when, based on these exam-ples, I advocate the possibility that “ghosts and spirits are not aware” then there certainly are good reasons [precedents] for this.69 Could it be pos-sible that the power of the spirits [to mete out rewards and punishments] is able to manifest itself, but that [spirits can choose] not to act upon events?70 This, I do not know. Or71 is it a certainty that their powers are incapable of manifesting themselves? That I also do not know. These are two diffferent scenarios [lit., “these two options branch out diffferently”]. And so when I

Slip 5[曰: “鬼神又（有）]72 所明, 又（有）所不明˳”此之胃（謂）唬（乎）！

say, “Ghosts and spirits are aware of some things but unaware of others,” this is what it refers to!

In addition to its core theme, the fact that the text invokes Yao , Shun , Yu , and Tang as archetypical sage-kings and Jie , Zhou , You, and Li as their opposites could suggest a common milieu with the received Mozi. Yet it is also clear that the Shanghai Museum fragment, if directly associated with the “Ming gui” triad, would make the received thesis on ghosts and spirits in Mozi 31 more complex since it offfers two hypotheses that are not made explicit there: (1) ghosts and spirits have the power of moral judgment

meaning “to add another explanation” (to the thesis). Nishiyama Hisashi transcribes 加, as does Wang Zhongjiang. Asano Yūichi follows Cao. Li Rui proposes to emend to jie 解. See Li Rui, “Du Guishen zhi ming zha ji.”

69 For Cao Jinyan, this is a rhetorical question: “How would I, because of this, speak favorably of . . .?”

70 安 (焉) expresses modality here rather than a real question. Ding Sixin takes it as a pronoun equivalent to the direct object pronoun 之. Ding Sixin and Wang Zhongjiang gloss 至 as transitive 致. Asano Yūichi, Xu Hua, and Nishiyama Hisashi concur with Cao Jinyan’s intransitive 至, the subject of which should not be the implied guishen 鬼神 but rather qi li 其力.

71 Chen Wei convincingly glosses yi 意 as yi 抑, “or, alternatively, . . .,” and gives paral-lel occurrences in Mozi 31 and 48. See Chen Wei, “Shang bo wu ‘Guishen zhi ming’ pian chudu”.

72 These four graphs are missing on the strip and have been supplemented by Cao Jinyan based on context.

126 roel sterckx


but not necessarily always the intent to apply those powers to each and every case; and (2) ghosts and spirits could lack any sort of intervention-ist power at all, or there are occasions when they simply do not have the capability to punish or reward . In other words, when there is no obvious spirit response, is this because the spirits have the choice to ignore or act upon a certain case, or is it the case that their powers are intrinsically limited? The fĳirst is a choice argument, and the second questions the very possibility of moral judgment by the spirits . The spirits ’ capacity to be percipient and omniscient is separated from their ability to exert these powers.

On that basis one could also question a direct link to the received Mozi, as Ding Sixin 丁四新, Li Rui 李锐, Erica Brindley , and others have done. The fĳirst line of the Shanghai Museum fragment says that there are things spirits know nothing about, the implication being that there are moral flaws that escape their scrutiny. It therefore follows that the actions of spirits can be experienced as whimsical, unpredictable, and capricious. The Shanghai Museum fragment appears to be taking a position difffer-ent from both the received Mozi as well as what are purported to be Ru views: there are events that ghosts act upon, but there are also events that fall beyond their providential radar. The fragment adds nothing to the question of the existence of ghosts and spirits. Their ability to mete out punishments and rewards selectively leaves the ontological question of their existence unaddressed. The Shanghai Museum fragment, Wang Zhongjiang 王中江 points out, could suggest that a spirit response is not inevitable or even that it can simply be a coincidence.73 Here we might have a debate between an interlocutor who takes the Mozi 31 position (ghosts are ming) and someone who questions this and argues that there are occasions when ghosts are not ming. The stance taken in the Shanghai Museum fragment, then, is contra Mozi 31; its line of questioning, including the option that spirits can selectively apply their moral sanction, seems to be more in line with the skepticism seen in the dialogues in Mozi 48.74

73 Wang Zhongjiang, “Guishen zhi ming yu Dong Zhou de ‘duo yuan guishen guan,’ ” 50.74 This point has also been noted by Okamoto Mitsuo 网本光生 in “Shang bo Chu jian

‘Guishen zhi ming’ yu ‘Mozi: Gong Meng’ suo jian liangduan duihua” and, independently, by Brindley in “ ‘The Perspicuity of Ghosts and Spirits’ and the Problem of Intellectual Afffĳiliations in Early China,” 218 n. 19. Note that comparisons based on direct textual par-allels with Mozi 31 are further complicated by the fact that the passage in Mozi 31 that introduces the expression guishen zhi ming 鬼神之明, at the beginning of test 3, occurs in a unit of twenty-one characters (是以莫放幽閒, 擬乎鬼神之明顯, 明有一人畏上誅



If parallels with the received Mozi come to mind, the acknowledg-ment in the Shanghai Museum fragment that spirit powers have a limited reach resonates with the two specifĳic dialogues in Mozi 48 mentioned above. In Mozi 48 Mozi indicates to his fĳirst interlocutor (the one failing to understand why he has not received blessings from the spirits despite years of dedicated service to Mozi) that humans may deceive themselves in thinking that they are morally accomplished. In the second dialogue Mozi acknowledges that even sages can sufffer ill health because not all factors that cause bad luck can be attributed to the spirit world. The latter implies that spirits have only partial powers, a position that would tally with the manuscript’s suo bu ming 所不明 thesis and the proposition that being perspicuous (ming 明) is not necessarily the same as having the ability (neng 能 ) to act upon it (wei 爲). Ding Sixin goes so far as to date the Shanghai Museum fragment as contemporary to Mozi 48 on the basis of these similarities (i.e., for Ding it has to be later than Mozi 31 and Mozi 26–28 and datable to Xunzi ’s time, ca. 313–230).75

In sum, rather than postulate that the Shanghai Museum fragment is a lost piece of the “Ming gui” triad, it might be safer to interpret this frag-ment as yet another example of the self-searching dialectics that surround the Mohist doctrine on ghosts and spirits . It could be evidence for the existence of difffering views on spirits among diffferent strands of Mohism. It could even reflect dissident views that were not exclusively Mohist. If the Shanghai Museum fragment is Mohist, rather than associating it with Mozi 31, its tone suggests that it is more akin to the skeptical reflection in the dialogue passages of Mozi 48 and 49. As Benjamin Wong and Hui-chieh Loy point out in their reading of a central line in Mozi 31 before the publication of the Shanghai Museum fragment, Mozi’s thesis is to a large extent an aspiration: he wishes he could make everyone in the world

罰) that is out of place or corrupt according to commentators. See Sun Yirang, Mozi jiangu, 243–244.

75 Ding Sixin emphasizes this potential fĳiliation in “Lun Chu jian ‘guishen’ pian de guishen guan ji qi xuepai guishu,” 417–422; and in “A Study on the Dating of the Mo Zi Dialogues and the Mohist View of Ghosts and Spirits,” 74–81, 84. The similarities that he fĳinds between the Shanghai Museum fragment and these two dialogues in Mozi 48 are not far-fetched. However, Ding overemphasizes the monovocal nature of Mohist views on spirits in the received Mozi and generally downplays the occurrence of skepticism in Mozi 31 itself. Ding too easily assumes clear-cut school afffĳiliations as the basis of his argument, which forces him to conclude that the Shanghai Museum fragment (1) is not Mohist, but (2) if it has to be Mohist, then it can only belong to a dissident Mohist group; (3) on the other hand, since Mozi 48 has conceptual similarities to Xunzi ’s “Tian lun” 天論, the manu-script could be “Ru,” but (4) it is not quite “Ru.”

128 roel sterckx


“conform to believe” (ruo xin 若信) that ghosts and spirits have the pow-ers to reward and punish , which is not the same as claiming that they already and invariably do so:

Mozi is not saying that people ought to believe that ghosts exist, but that “ghosts with the power to reward and punish the wicked” exist. Compare also his statement of the prevailing attitude towards ghosts: there are per-sisting doubts about their existence, and there is also a misunderstanding about their nature and power. This points at least to two diffferent kinds of “unbelievers”: those who do not believe that ghosts exist, and those who believe that they exist but not that they have the power to reward the good and punish the wicked.76

Mozi’s ambition is not to prove the existence of ghosts but rather to promote the sociopolitical utility of a natural or indeed force-fed belief in their punishing and rewarding powers. His aim is fĳirst and foremost to demonstrate the utility of having a persuasive argument that favors their existence rather than to come up with ontological proof. Whether, as Wong and Loy claim, this impulse to invoke the workings of the spirit world without seeking to prove that spirits exist necessarily constitutes a “challenge to classical Confucianism” is far less certain in my view. One might argue that an informed aporia on the question of the existence of spirits combined with a practical attitude to keep spirits at a distance or use them to one’s advantage tallies well with the Confucian tenet that one should sacrifĳice to spirits as if they are present (祭神如神在, Lunyu 3.12). Even Confucius ’s ambition in the Lunyu to keep the spirit world at bay through ritual mimicry or discursive abstention hinges on the assumption that humans are able to influence the spirit world and prevent them from intervening in human afffairs through deliberate neglect. Leaving aside the thorny issue of what should be taken as the canonical basis for so-called Ru views of the spirit world, the implication could be that Mozi misun-derstands the Ru position on ghosts and spirits . The Ru nowhere deny the

76 B. Wong and Loy, “War and Ghosts in Mozi’s Political Philosophy,” 348. On the issue of “belief” Chad Hansen comments that Mozi “does not literally argue for believing in spirits any more than Confucius does! . . . He especially worries about wasting the traditional mor-alizing literature that describes spirits redressing injustice. How, he asks, can it be appro-priate in this linguistic tradition to wu [無] spirits ? After a string of such literature based arguments, Mozi concludes with an argument that pointedly stresses the pragmatic nature of the issue. He adopts a tone that suggests that he himself actually doubts that there are spirits . He is arguing that we have to adopt and follow this dao of you-wu. Because of the popular literature, folk beliefs, word of mouth, and so on, the proper and utilitarian guiding language will use you [有] of spirits .” See Hansen, A Daoist Theory of Chinese Thought, 118.



existence of the spirit world outright but rather leave the question open and merely propose an alternative method to deal with them. For the Ru, disengaged reverence will keep spirits at bay, and sacrifĳice therefore seeks to appease in a negative way (you want them to leave you alone); for Mozi, sacrifĳice forges a direct link with the spirits in a positive way (since they are there anyway, you may as well make sure they work for you).

If the fĳigure bringing chaos to the world on slip 3 of the Shanghai Museum fragment is in fact identifĳied as Duke Mu of Qin , who, in the received Mozi 31, is granted an additional nineteen years of life, one pas-sage in the Lunheng 論衡 (Discourses Weighed in the Balance) might even suggest that the Shanghai Museum fragment could be associated with a Ru tradition. The passage in question forms part of Wang Chong ’s 王充 (27–ca. 100 CE) overall argument that Heaven does not intervene either to extend or to curtail someone’s life span on account of his or her deeds:

A follower of the Ru (儒家) named Dong Wuxin 董無心 and a disciple of the Mo (墨家) named Chanzi 纏子 met each other and spoke about the Way. Chanzi praised the Mohist doctrine that one can rely on the assistance of ghosts and spirits , and in support he brought up the case of Duke Mu of Qin , who, because of his enlightened virtue, had been granted an additional nineteen years of life by Shang Di. Dongzi77 objected to this because Yao and Shun had not been granted extra years and because Jie and Zhou did not die young.

Yao , Shun , Jie , and Zhou can be considered to belong to the distant past, but more recently, one could [equally] take the cases of Duke Mu of Qin or Duke Wen of Jin as examples to refute the [Mohist] thesis. A posthumous name is like a footprint of someone’s deeds. Traces of one’s deeds during one’s lifetime form the basis for a posthumous name in death. Mu 穆 is a name that stands for error and chaotic conduct; Wen 文 is an expression of virtue and grace. Is it the case then that Heaven granted additional years to someone with erroneous and disorderly conduct, or that it took away the life of someone who behaved virtuously and with grace? The reign of Duke Mu did not surpass that of Duke Wen, and Duke Wen’s posthumous name is more beautiful than that of Duke Mu. Yet Heaven did not extend the years of Duke Wen of Jin; it only granted extra years to Duke Mu. This means that Heaven rewards misconduct and disorder , much the same way as Duke Mu himself did.78

77 Note that the received text here says Chanzi but most commentators agree that it should be Dongzi (alternatively one could read “Chanzi’s argument was objected to on account of . . .”).

78 See Lunheng “Fu xu” 福虛.

130 roel sterckx


To be sure, the labeling of Duke Mu of Qin as an unworthy person because his posthumous name (or a variant thereof) can be decoded as such is an argument initiated by Wang Chong to support his own (Dongzi’s) belief that Heaven behaves in a whimsical way. Yet this Lunheng passage also comes strikingly close to the contents of slip 3 of the Shanghai Museum fragment, granted of course that the fĳigure there is identifĳied as Duke Mu of Qin . Wu Zixu in the Shanghai Museum fragment provides the equiva-lent example of Duke Wen of Jin in the Lunheng passage. Some scholars therefore take these parallels as evidence that might link the Shanghai Museum fragment to the persona of Dongzi, who is distinctly identifĳied as Ru.79 Naturally, the similarities are there, yet this hypothesis is also problematic on several grounds: fĳirst, reading a Han passage back into a fourth- or third-century BCE manuscript can be seen as Hineininterpre-tierung; second, the identifĳication of the fĳigure of Dongzi as a Ru in a Han or Tang source need not reflect a Warring States reality, especially when this fĳigure is unattested in sources contemporary with the received Mozi.80 In short, these attempts by Chinese scholars to make a detour through the Lunheng with a view to attaching a “school” label to the Shanghai Museum fragment illustrate the complexities involved when one infers conceptual resemblances based on fragmentary terminological parallels.

Many questions therefore remain. Without further evidence to help flesh out its context, the value of the Shanghai Museum fragment lies primarily in that it should encourage us to refrain from labeling Mohist views on spirits as generically unifĳied, uncontested, univocal, and, conse-quently, representative of some kind of “school.” If proven to be Mohist, at most this new piece of evidence suggests that there were voices ques-tioning the absolute nature of the Mohist thesis on certain spirit inter-vention. Such criticisms could suggest a rivalry of ideas among diffferent branches of Mohists, yet there is nothing that excludes the possibility that the skepticism and doubts expressed in the Shanghai Museum fragment

79 See Xu Hua, “Shang bo jian ‘Guishen zhi ming’ yi wei ‘Dongzi’ yiwen” (who also adduces another similar dialogue between Chanzi and Dongzi preserved in Ma Zong’s 馬縂 [d. 823 CE] Yilin 意林 as evidence); Yang Zesheng and Li Jiahao, “Tan Shang bo zhushu ‘Guishen zhi ming’ zhong de ‘Song Mao Gong,’ ” 185; Zheng Jiewen, Zhongguo Mo xue tongshi, 198.

80 The bibliographical treatise in the Hanshu 漢書 mentions a work entitled Dongzi 董子 in one scroll. His name is given as Wuxin 無心 and he is identifĳied as an opponent of Mozi. See Hanshu 30.1726. The same work is mentioned in the bibliographical treatises of the Suishu 隋書 (34.997) and Xin Tangshu 新唐書 (59.1510). Xu Hua suggests, on the basis of quotations, that it may have been in circulation until late Ming times. See Xu Hua, “Shang bo jian ‘Guishen zhi ming’ yi wei ‘Dongzi’ yiwen,” 105–106.



were uttered in a Ru context or in an environment that was not explic-itly identifĳied as Mohist. At best the Shanghai Museum fragment demon-strates that the idea of spirit intervention in response to human behavior was debated. My reading of this manuscript therefore concurs with and reinforces the assessments reiterated recently by Li Rui and Erica Brind-ley : fragmentary conceptual parallels or the overlapping of terminology between received and excavated texts do not constitute sufffĳicient evi-dence to attach a particular intellectual or “school” afffĳiliation to a text, a concept that itself is problematic.81 Following the same logic, one should allow for the possibility of a similarly porous interplay and divergence of arguments within textual units of a received text. Reconstructing what is newly excavated should inspire us to deconstruct what is received.

Ghostly Demands versus Frugal Rituals

The ontological problem of the existence of a spirit world and questions surrounding the basis of their interventionist powers are not the only themes in Mozi 31 that are cast into doubt elsewhere in the received Mozi and beyond. Mozi 31 sits uncomfortably among other Core Chapters that advocate thrift in ritual expenditure . To argue that the spirit world looms permanently on the doorstep as a “supramundane surveillance system” of human conduct, to quote Ian Johnston ,82 but to support moderation and frugality in funerary or ritual expenditure elsewhere might be seen as inconsistent. Several voices, past and present, have noted this apparent inconsistency, starting with the eponymous Mozi fĳigure himself in Mozi 48, who compared rituals without spirits to throwing nets when there are no fĳish. Wang Chong takes Mozi to task on this issue at great length:

The Mohists (Mo jia 墨家) advocate thrifty burials yet also honor ghosts , so their Dao is perverse and stands in mutual contradiction to their deeds; it follows then that given this [contradiction], their teachings are hard to follow. Wherein now does the contradiction lie? Suppose that it is not the case that ghosts are the seminal essence ( jing 精) of dead people, then they would not be aware of the fact that people honored them. Now, the Mohists claim that ghosts are the essence of dead people; they treat a dead person’s essence [i.e., his ghostly manifestation] generously, yet treat a dead person’s

81 Li Rui, “Lun Shang bo jian ‘Guishen zhi ming’ pian de xuepai xingzhi,” 31–33; Brind-ley, “ ‘The Perspicuity of Ghosts and Spirits’ and the Problem of Intellectual Afffĳiliations in Early China.”

82 Johnston, The “Mozi,” lx.

132 roel sterckx


body tightfĳistedly. They are generous to their spirits (shen 神), yet are stingy in treating their bodies (ti 體). When stinginess and generosity do not can-cel each other out, and outward and inward do not assist each other, then [ghosts] will be furious and send down calamities. Even if one esteemed ghosts , in the end one [does this] because of the feelings of hatred [that might be harboured by] the dead. It is human nature to desire generous treatment and to despise ungenerous treatment; the same applies to the hearts of the spirits (shen xin 神心). If one uses the method of Mozi and serves ghosts to seek blessings, blessings will rarely come about but calami-ties will arrive frequently. And, ab uno dice omnes, the methods of the fol-lowers of Mozi are all like this. That their doctrines disappeared and were not transmitted has its basis in this.83

For Wang Chong , Mozi’s inconsistent stance results from a misunder-standing of the nature of death. His criticism, however, should be seen in the context of his own time, the Han, and in the context of his own work, the Lunheng , which contains several chapters that are preoccupied with the nature of death and the moral continuity between the living and the dead. In contrast, in the received Mozi, the doctrine on spirits rarely involves theoretical discussion of the nature of death. The only noticeable exceptions are the witness accounts 1 and 3 in test 1 in Mozi 31, where the accused threaten that if the dead can be presumed to have conscious-ness (which is implicitly admitted), then they will take revenge upon their executioners within three years. Wang’s claim therefore that the declining popularity of Mohism in his own time was to be attributed to the inherent doctrinal contradiction he highlights in the passage above must be exag-gerated, or at least, internal contradiction can only be one among several factors that might explain why Mohist ideas had fallen into disfavor in his time.84

Modern scholars too have commented on Mozi’s, at fĳirst sight, para-doxical stance. Feng Youlan 馮友蘭 speaks of a “seeming inconsistency” in the positions of not only the Mohists but also the Ru, who insist on funerary and sacrifĳicial rituals “yet did not believe in the existence of the spirits .” Feng reconciles these inconsistencies and calls them unreal on the grounds that Confucians perform sacrifĳicial rituals as a “poetic” sentiment of respect toward the departed rather than as an afffĳirmation of a religious

83 See Lunheng “An shu” 案書.84 Michael Nylan cautions not to extrapolate too much from this Lunheng passage

about the fate of the Mohists in Wang Chong ’s time. See Nylan, “Kongzi and Mozi,” 3–4. On the problems of interpreting the paired reference to Ru and Mo in Han, see also Chen Jinxia, “Lun Shiji zhong Han Wudi ‘yi guang Ru Mo.’ ”



belief in the existence of spirits. Mohists, for their part, Feng argues, are ultra-utilitarian in that they seek to prove the existence of spirits merely to support their core doctrine of universal love. They employ any theory that can be used to that end, including advocating economies in ritual expenditure .85 Zheng Jiewen 郑杰文 broadly concurs: logically speaking, Wang Chong ’s criticism of the Mohist inconsistency is accurate, but it is not relevant since the main purpose of the Mohists is to use their stance on spirits as a service doctrine to promote impartial care, frugal burial, and so on.86 The implication here is that some of the doctrines in the Core Chapters would stand as independent tenets, while others merely serve to lend weight to these. It is not clear to me on what basis the “Ming gui” triad should be singled out to do just that.

Feng Youlan’s contention that the Ru do not believe in spirits, except perhaps in the form of an empathy toward ancestral spirits outside the common pantheon maintained through simple do-ut-des sacrifĳices , is equally problematic. It assumes that Ru sacrifĳicial ritual distinguishes itself from “popular” or “common” religion because Ru ritual is inspired by an advanced moral program. It also implies that sacrifĳicial practice in a Ru context should be seen as a manifestation of elite culture, while, by contrast, Mohist cultic activity can be reduced to a gift-giving dynamic popular among the masses, who are mostly concerned with the material aspects of cult expenditure or the question of whether humans can expect a direct and immediate return for their expressions of reverence.

This attempt to distinguish Mohist religiosity from Ru or “Confucian” attitudes on the basis of their views on ritual is tenuous. It also appears in the work of Benjamin Schwartz , who, in his explanation of the notion of li 禮, distinguishes a secular sphere of “respectful manners” from “reli-gious rites of sacrifĳice .” Schwartz speaks of a Mohist disbelief in the “magi-cal” spiritual-ethical function of ritual and music : “In carrying on religious sacrifĳices [the Mohist] does so out of gratitude towards the spirits and in order to please the spirits, not because ritual makes him a nobler person. In this sense, ritual is not ‘magic’ in Fingarette’s sense.”87 It follows that Mohist acts of service to the spirit world (i.e., the offfering of sacrifĳice ) do not transform the performer and, hence, are more “secular” (or a-reli-gious) than ritual and music as advocated by the Ru. Yet how useful is it to

85 Fung, A Short History of Chinese Philosophy, 57–58.86 Zheng Jiewen, Zhongguo Mo xue tongshi, 209–210.87 Schwartz, The World of Thought in Ancient China, 152–153 (my italics).

134 roel sterckx


separate sacrifĳicial cult from a category of high ritual and associate the for-mer with Mo and the latter with Ru? If ritual, as Schwartz suggests, covers the sphere of basic etiquette, are we to assume then that, say, table man-ners are less “secular” and more morally transformative than the prepara-tion of sacrifĳicial offferings? And does Mohist criticism of ritual imply that the Mohists are disinterested in ritual altogether? This is highly doubtful. More convincing in this respect is Mark Csikszentmihalyi ’s suggestion that Mohist criticism was directed to ritual as practiced by the Ru (because it is inconsistent) rather than ritual in general. In the Mohist perspective, rituals can serve a utilitarian purpose provided that one adopts the right ones and is consistent in their application.88 This certainly appears to be a view held by Han times, when the complexity of Ru ritual is highlighted as objectionable in the eyes of Mohists.89

In short, the Mohist advocacy for the simplifĳication of ritual should not imply that Mohists considered ritual to be unimportant altogether. By the same token, it cannot be assumed that the so-called Ru conceived of ritual as devoid of utilitarian motives. Indeed, offfering sacrifĳice to keep the spirit world at bay and refraining from discourse about them (Zi bu yu 子不語) equally serve a purpose. The Mohists indeed constantly weigh and nego-tiate the material requirements of ceremony, but so do most thinkers of the period. If we accept the premise that, as my internal analysis of the received Mozi shows, Mohist perceptions of the spirit world were dynamic, internally contentious, and evolving, there is no need to explain a Mohist belief in service to the spirit world and their advocacy of frugality in ritual as a doctrinal contradiction. Putting together passages expressing doctrinal doubt in Mozi 31, exchanges preserved in the Dialogues , and the Shanghai Museum fragment, a picture emerges of a Mohist landscape in which the polemic on spirits appears to be as much an internal debate as it is a discussion pitched against external parties. And if we refrain from assuming that Mohist views on the spirit world and ritual in general were exclusively formulated in the context of a polemic against an identifĳiable group known as Ru, the contradiction Wang Chong highlights might dis-solve. It is impossible to establish in each case who are understood to be the Ru in the received Mozi (e.g., none are mentioned in Mozi 31) and

88 Csikszentmihalyi, Material Virtue, 36–37.89 See, e.g., Huainanzi “Yao lüe,” which notes that Mozi found the rituals of Confucius

uselessly complicated and difffĳicult (煩擾而不說).



whether or not the mentions of the Ru across the received Mozi all refer to the same group of people.

The apparent contradiction between ritual and materiality is not solved in the Mozi. The point the Core Chapters seem to make is that rituals , if performed inconsistently, could be a waste of time and resources, but the material requirements of ritual are not invariably rejected. At the end of Mozi 31, when the skeptics object to spending wealth (cai 財) on sacrifĳicial offferings, Mozi counters their claim by insisting that such expense would be justifĳied even if it only brings people together (31: 55/1–11 ). In the sec-ond test in Mozi 31, Mozi makes ritual expenditure part of an ontological proof for the existence of ghosts and spirits , or at least, he uses it to back up his claim that the ancients believed in them:

故曰: 官府選效, 必先祭器祭服, 畢藏於府, 祝宗有司畢立於朝, 犧牲不與昔聚群, 故古者聖王之為政若此˳ Therefore, it was said that when the government offfĳices provide the imple-ments, they must fĳirst ensure that the proper sacrifĳicial vessels and robes are fully stocked in the warehouses, that the invocators of the ancestral temple and all other offfĳicials in charge [of sacrifĳices ] have all been appointed in the court, and that the animals to be used as sacrifĳicial victims are no lon-ger grouped together with the common herds. Since the sages of antiquity conducted their government in this fashion, it must be the case [that they believed in the existence of ghosts and spirits ]. (31: 53/2–3 )

Moderation in ritual expenditure is not universally acclaimed across the received text. In a passage in Mozi 25 he argues that sumptuous funerary expenditure has a theological efffect since it might deplete resources that could otherwise be used to maintain regular sacrifĳices :

今惟無以厚葬久喪者為政, 國家必貧, 人民必寡, 刑政必亂˳ 若苟貧, 是粢盛酒醴不凈潔也˳ 若苟寡, 是事上帝鬼神者寡也˳ 若苟亂, 是祭祀不時度也˳ 今又禁止事上帝鬼神, 為政若此, 上帝鬼神始得從上撫之曰: 「我有是人也, 與無是人也, 孰愈？」曰: 「我有是人也, 與無是人也, 無擇也˳ 」則惟上帝鬼神降之罪厲之禍罰而棄之, 則豈不亦乃其所哉˳ Now, if one follows those who support elaborate funerals and lengthy mourning to conduct government, then the state will necessarily become poor, the people will become few, and the government will necessarily be in chaos . If the state is poor, then the sacrifĳicial grains and wine will not be of the required purity. If the people are few, there will be few to serve Shangdi and the spirits . And if the government is in chaos , then sacrifĳices will not be conducted at the appropriate times or in the appropriate mea-sure. If now one conducts government in such a way that one efffectively

136 roel sterckx


prevents the proper services to Shangdi and the spirits , they will be the fĳirst to look down from above and, considering how to soothe the people, might say: “What is better for us, to have these people exist or to have them not exist?” Or: “Whether they exist or not does not make any diffference to us!” Consequently, Shangdi and the spirits will send down cruel punishments for the people’s misdemeanors and abandon them. And if they do so, would not that just be the appropriate thing to do! (25: 40/18–22 )

Lavish expenditure not only exhausts the state’s resources but, more importantly, incites discontent in the spirit world itself. Indirectly Mozi here invokes the spirits themselves as moral arbiters of ritual expenditure. In Mozi 27 he draws on religious obligation to promote his views of an egalitarian society: social harmony and the impartial division of wealth ensure that the state will always have the necessary resources to provide offferings to the spirit world (27: 44/28–30 ). So the material basis of ritual obligation is fĳirmly acknowledged by Mozi. It supports his ideal of the impartial society: a division of wealth leads to an equally shared burden in the sustenance of religious obligation , which, in turn, prevents ritual expenditure from undermining conventional economic productivity or burdening one social class more than another. To label the Mohist world-view, in this respect, as dominated by religious conservatism is therefore exaggerated.90

In its most basic form one would assume that the Mohist expects a return from the spirits that is proportionate to the generosity of his offfer-ings. But Mozi takes a morally superior stance when, in a passage in the Dialogues (in Mozi 49), he insists that sacrifĳice should not degenerate into a calculated quest for blessings but must reflect moral conduct:

魯祝以一豚祭, 而求百福於鬼神˳ 子墨子聞之, 曰: 「是不可˳ 今施人薄而望人厚, 則人唯恐其有賜於己也˳ 今以一豚祭而求百福於鬼神, 鬼神唯恐其以牛羊祀也˳ 古者聖王事鬼神, 祭而已矣˳ 今以豚祭而求百福, 則其富不如其貧也˳ 」A priest from Lu offfered one pig in sacrifĳice and asked for a hundred bless-ings from the spirits . Our Master Mozi heard about this and said that this cannot be done. Now, to give to others sparingly yet to expect much from them would cause them to be afraid of the gifts given to them. Now, since one pig is offfered and yet a hundred blessings are sought for from ghosts and spirits , they would be quite afraid if they received [an even more valuable] sacrifĳice of oxen and sheep. Anciently, when the sage-kings served ghosts and spirits , they just offfered sacrifĳice and that was all. Yet today, by offfer-

90 See Knoblock, Xunzi, vol. 1, 58, for an example of overemphasis on Mohist religious conservatism.



ing one pig while seeking a hundred blessings, one would be better offf to remain poor than to become rich by it. (49: 114/21–23 )

Here a priest from Lu (heartland Ru territory) is reprimanded for expect-ing a spirit response that is disproportionate to the value of the sacrifĳicial offferings. Sacrifĳice should be about moral intent rather than the quest for blessings from the spirits by means of gifts. The offfering of sacrifĳice (do) should not necessarily come with an expectation of greater recom-pense (des). In emphasizing the role of human conduct Mozi’s attitude here could be comfortably construed as “classicist.”91 His exhortation to the priest of Lu is that sacrifĳices should morally transform a person. In the analysis of Feng and Schwartz , then, the Mozi fĳigure in this passage would be the perfect Confucian.

In sum, despite the fact that in certain debates in the Mozi, Mohist views are clearly contrasted with those of the Ru, this dichotomy must be handled with care in our analysis of the text. In the case of views on the spirit world, it would be hard to claim that the Ru simply dismissed ghosts

91 Zheng Jiewen constructs a hypothesis to explain this seeming discrepancy and argues that these are debates that took place in the state of Lu at a time when Mozi had yet to carve out a clear doctrinal niche. Mozi, according to Zheng, is still learning from the Ru and has yet to make up his mind. His sojourn in Lu should therefore be seen as a dialecti-cal arena in which Mohist core doctrines gradually became distinct from Ru teachings. The implication for Zheng then is that the Core Chapters must be later than the Dialogues, which would represent an early phase, when Mohist ideas were in an embryonic and inar-ticulate state. See Zheng Jiewen, Zhongguo Mo xue tongshi, 3, 22–23, 41–42, 46–47. Zheng’s scenario unfortunately is based on tenuous extrapolations, as is shown by the fact that he needs to rely on quotations on sacrifĳice from the “Li qi” 禮器 chapter in the Liji 禮記 and commentaries by Zheng Xuan 鄭玄 (127–200 CE) and Kong Yingda 孔穎達 (574–648 CE), texts and commentators dated centuries post factum. Zheng Jiewen’s summary of the Mohist position is extremely speculative on other fronts. First, the doctrinal doubts expressed by Gongmeng suggest to Zheng that the theory of morally conscious spirits was established early on when the Mohists were setting themselves apart from the Ru. Zheng makes this argument with an overly simplistic analysis. According to Zheng, the “school” of the Ru does not like to talk about ghosts and spirits and prefers instead to concentrate on self-cultivation. Yet for the lower classes of the population, whose main concern is survival, such elevated doctrines are of little help. Mozi thus provides the common folk with ideas that they can connect with and that might assist in encouraging rulers to toe the line. Second, Zheng speculates that the early Mohists insist on proving the existence of ghosts and their capacity to reward (and hence the need for sacrifĳices, etc.). Later on, however, as the status of individual disciples changes, Mozi’s doctrine takes on a more social character in that its main focus lies in encouraging disciples to cultivate themselves and promote the selection of talented folk. It is unclear to me what basis Zheng draws on to make these claims. Finally, Zheng asserts that Mozi’s own disciples question his theories about providential spirits because of a disjuncture between his ideals on the one hand and on the other hand the social reality and spirit of the time, which was increasingly human centered. Underneath Zheng’s last hypothesis lurks again a narrative that echoes the so-called triumph of Confucianism.

138 roel sterckx


and spirits . Their insistence on ancestral rituals illustrates an acute aware-ness of the spirit world. Perhaps the reality was more complex. Rather than simply discrediting agnostic Ru and accusing them of hypocrisy when investing in rituals without believing in spirits , the Mohists may have sought to “democratize” the pantheon and divorce the notion of spirit power from its ancestral basis as emphasized by the Ru. The dif-fering views of the role of the spirit world by what are purported to be Mohist and Ru camps would therefore be not so much of substance but rather of degree: rather than being preoccupied with the question of the existence of spirits, their shared concern was to understand what aspects of the spirit world should be deemed spiritually potent and when. The Mohist would argue that ancestral spirits are partial and, therefore, that their spirit response fails to have universal or impartial relevance (unlike nonancestral spirits). Mozi 31 can then be read not merely as a chapter that seeks an imprimatur for sacrifĳicial cult of the folk-religious type but also as a plea for a more inclusive spirit pantheon. The Mo-Ru disagree-ment hinges, not on whether or not spirits are potent, but rather on the question of what signs should be taken as signifĳicant and hence acted upon. Mohists would argue that the spirits should be monitored con-stantly since the spirit world is impartial in judging human conduct; the Ru instead assume that the spirit world is biased toward certain groups or individuals. For them the crux lies in deciding when to ignore signs from the spirit world and when to interpret them as socially signifĳicant. As a result, Mohists prefer to see cultic activity and expenditure dedicated to the entire pantheon so as to benefĳit as many as possible, whereas the Ru prefer to invest resources in cultic activity on behalf of those enti-ties that are of immediate concern to clan, lineage, or locality. Whereas Mohists want resources for the sustenance of the spirit world deployed for the benefĳit of the greatest common good (an attitude that could be construed as “moderate” or “utilitarian” ), the Ru do not object to the idea that resources can justifĳiably be concentrated and invested in a cause that is partial (and concentrating goods for use of the few can indeed be construed as “excessive”). At any rate, we must assume that many at the time were undecided on the efffĳicacy of the spirit world, as the Shanghai Museum fragment corroborates.

Conclusion

Evidence does not allow us to reconstruct the sociology of Mohism with sufffĳicient detail. Yet there are elements both internal and external to the



received Mozi that suggest that the Mohist position on ghosts and spirits may not have been straightforward or internally uncontroversial. Mohist views on the spirit world were not static but evolving, or at least, they were more nuanced and subject to debate than scholars have suggested. I have shown that it is possible to arrive at a picture of a more polyphonic Mohist view of the spirit world. I also hope to have shown that the absence of all but one of possibly three original chapters that made up the “Ming gui” triad does not prevent us from doing so, and that a close reading of units at the subchapter level, as well as comparisons with passages in the other Core Chapters and the Dialogues and the Shanghai Museum frag-ment, reveals a more complex picture. This picture suggests that there were Mohists who were skeptical about the prescience of the spirit world. It also reveals that the philosophical issue of the “existence” of ghosts and spirits , on the one hand, and the more pragmatically inspired question of whether or not the spirit world is capable of punishing and rewarding , on the other hand, were seen as separate issues. The fĳirst may have been an ideological or theological debate that occurred between Mozi and his crit-ics , the latter not necessarily identifĳied as Ru in each and every case. The second concern may have been debated as part of what some perceived to be a “Mo versus Ru” polemic, with the latter advocating a fatalism far more detached from individual moral (or immoral) action than the for-mer. Yet even here, the interlocutors in these exchanges are not always clearly defĳined.

The question then arises as to what encourages scholars to continue to present the Mohist stance on spirits in the form of a terse and seemingly unproblematic acknowledgment that Mohists simply believed in ghosts ? In this respect Nicolas Standaert has made the incisive observation that our reading of the Mohist Core Chapters has been overly influenced by the titles these chapters carry. Given that the phrases used as chapter titles almost never occur in the text itself, Standaert suggests that they were not conceived as part of the original chapters but that they may have been assigned by an editor who drew on the list that appears in a passage in Mozi 49. In that particular passage the doctrine on ghosts is referred to as shi gui 事鬼, “serving ghosts.”92 And so the relationship between a title and a chapter is not necessarily clear and, in fact, sometimes seems random. The case of Mozi 31 and the Shanghai Museum fragment support Standaert ’s hypothesis, since one could argue that the ideas presented in

92 See Standaert, “Problems with Titles.” For the passage on the ten doctrines, see Mozi 49: 114/7–10, quoted in the introduction to this volume.

140 roel sterckx


the received Mozi are very elastic and swing between moments when the theoretical issue of the spirits ’ consciousness or existence is at play and moments when the texts seem to advocate nothing more than a generalist exhortation to respect ghosts (even if one is not convinced of their exis-tence). Naturally, we do not know what title was assigned to the Shanghai Museum fragment, if any. But the title assigned to Mozi 31, “Ming gui,” if taken as a verb–direct object phrase, “explaining ghosts ,” does appear to be a weak choice if it was meant to reflect the core content of the chapter. In this respect a reading of ming as a stative verb (“percipient ghosts ”) would be equally plausible.93

This, together with a seeming paradox between the vociferous advo-cacy of sacrifĳice and a criticism of ritual opulence, indicates that, when situating Mohist views on spirits in their context, we should also refrain from isolating these views as either uniquely Mohist or uniquely inspired by or crafted in opposition to Ru views. Furthermore, in approaching the Mozi, we may also have to be more careful in handling analytical concepts such as religion versus ritual and popular (folk, common) versus elite or indeed in assuming that we understand the contours of what “sacrifĳice ” or “ritual ” meant at the time. A much more systematic analysis of religious life in practice is required for the period in question. The anecdotal and often-dispersed nature of much of the discourse on ghosts and demons in Warring States texts indicates that debates on the spirit world rarely happened exclusively as a philosophical exercise. We may not be able to identify the voices in these discussions with certainty, but it is imperative that we remove expectations that their views can be readily pigeonholed in terms of a polemic between philosophical “schools.” Warring States debates on sacrifĳicial practice and cult in general were rarely held in iso-lated ideological or theoretical terms, and practical utility seems to unite them all, whether the polemic is associated with Mohism or not. The intellectual and religious world of fourth- and third-century BCE China was certainly more complex than a dialectic described by Herrlee Creel in 1951, in which “vulgar” or “superstitious” Mohists contended with “ratio-nal” Confucians.94

Ironically, many of the strategies one sees at work in the polemics sur-rounding purported Mohist and Ru views on spirits in the Warring States period reappear a few centuries later among vociferous critics active in

93 Johnston allows for both readings; see Johnston, The “Mozi,” lvii–lviii.94 Creel, Confucius, 198.



the Eastern Han. The writings of Wang Chong , Wang Fu 王符 (90–165 CE), and Ying Shao 應劭 (ca. 140–204 CE) are peppered with ambiguities on the issue. Although they appear to dismiss a preoccupation with ghosts and supernatural phenomena on the grounds that it is “superstitious” or vulgar (su 俗), at the same time the very existence of the spirit world is accepted or at best left unquestioned. Ghostly activity is not discounted a priori but made subject to human response. Wang Chong does not deny the potential influence of the spirit world on human afffairs, but like some of Mozi’s more skeptical interlocutors , he seeks to establish which forms of spectral involvement humans should accept as being of a genu-ine demonic nature. And a fĳigure such as Ying Shao stresses the dangers inherent in either an obsessive preoccupation with the spirit world or the absolute neglect of demonic appearances.95

It appears then that, in Warring States as in Han China, disagreement arose not so much over the presumed appearance or status of ghosts and spirits or the impact of the divine on human afffairs; rather, the debate focused on whether such signs from the spirit world were to be taken as genuinely ominous and hence socially or politically relevant and whether there were occasions when the spirit world could be responsibly neglected. But whereas in the Warring States period concern with these issues was often associated with the Mozi fĳigure, by Han times such questions had taken on a more general existence and their links to the world of the his-torical Mozi had faded.

95 The core arguments occur in the “Jie chu” 解除, “Lun si” 論死, “Si wei” 死偽, “Ding gui” 訂鬼, “Si yi” 祀義, and “Ji yi” 祭義 chapters of the Lunheng; the “Si dian” 祀典 and “Guai shen” 怪神 chapters in the Fengsu tongyi 風俗通義; and the “Bu lie” 卜列 and “Wu lie” 巫列 chapters in the Qianfu lun 潛夫論.

\"mozi 31: explaining ghosts, again\"

Documents