Is this me?A story about personal identity from the Mahāprajñāpāramitopadeśa / Dà zhìdù lùn

ABSTRACT

In a Buddhist treatise from around the fourth century CE there is a very remarkable story which serves as a thought experiment calling us to question the nature of self and the identity of persons. Lost in Sanskrit, the passage is fortunately preserved in a Chinese translation, the Dà zhìdù lùn. We here present the first reliable translation directly from the Classical Chinese, and discuss the philosophical significance of the story in its historical and literary context. We emphasise the philosophical importance of embedding the story in two framing narratives, and demonstrate that the story taps a range of intuitions, and indeed fears, about the survival of the self which have also played a large role in the history of the topic in the West, and which continue to be of great contemporary concern.

KEYWORDS:

• personal identity
• selfhood
• Dà zhìdù lùn
• framing narratives
• Mādhyamika Buddhist philosophy

View correction statement:

Correction notice

Acknowledgements

We would like to thank Michael Beaney and two referees for their helpful feedback on our manuscript.

Jing Huang is grateful to Wai-Hung Wong, Jason Yonover and especially Chong Fu and Bo Shi for their support. She has also greatly benefited from discussion of earlier drafts with Wei Cheng. Finally, she would like to thank Jonardon Ganeri for his trust, patience, and thoughtfulness that made this project possible.

Notes

1 This would date the preservation of the ship of Theseus to the time of Alexander. Alexander's invasion of northern India in 325 BCE, as is well known, led to the formation of a syncretic Graeco-Buddhist tradition, as witnessed in Gandhāra sculpture and The Questions of King Milinda, a treatise purporting to record a conversation between the Greek king Menander and a Buddhist monk, Nāgasena.

2 The title of the text was initially reconstructed as Mahāprajñāpāramitāśāstra. However, after considerable scholarly attention has been paid to several manuscript fragments discovered in Kucha (庫車), which give the original title of the text in Chinese transliteration “摩訶般若波羅蜜優波提舍”, there is now a consensus that the title should be Mahāprajñāpāramitopadeśa; see Chou, “庫車所出《大智度論》寫本殘卷之研究”.

3 The literature is very extensive. See, for instance, Chou, “The Problem of the Authorship of the Mahāprajñāpāramitopadeśa”; Hikata (ed.), Suvikrāntavikrāmi-paripṛcchā-Prajñāpāramitā-Sūtra, lii–lxxv; Katō, “羅什と『大智度論』”; Lamotte, Der Verfasser des Upadeśa und seine Quellen.

4 An English translation of the extant Sanskrit text preserved in the Divyāvadāna is available in Strong, The Legend of King Aśoka. Our story’s precursor is not found in the surviving Sanskrit text but can be traced to a fourth-century (?) Chinese collection of miscellaneous Aśoka legends (see §2).

5 Contemporary philosophers have developed similar scenarios in the decades subsequent to the publication of Lem’s story. Parfit (Reasons and Persons) offers a similar thought experiment involving what he called a ‘teletransportation’ machine. John Perry’s A Dialogue on Personal Identity and Immortality structurally resembles Lem’s story, and the theme is echoed in Daniel Dennett’s science-fictional essay “Where am I?”. This reconfirms the importance of our story to the history of the philosophy of personal identity.

6 For details, see Westerhoff, Nāgārjuna’s Madhyamaka, 67–90; Tillemans, How Do Mādhyamikas Think?, 95–110.

7 Affinities have been found between the philosophical method of Nāgārjuna and that of the Phyrrhonian sceptic, Sextus Empiricus (second/third century CE). For two very recent studies, see Beckwith, Greek Buddha; Neale, Madhyamaka and Pyrrhonism. See also Kuzminski, Pyrrhonism, and Garfield, “Epoche and Śūnyatā”, 285–307. Regrettably, there does not appear to be a discussion of the ship of Theseus in the extant works of Sextus.

8 The most important accounts of the production of the Dà zhìdù lùn that survive are those of Sēngruì 僧叡 and Huìyuǎn 慧遠 recorded in Sēngyòu’s 僧祐 Chū sānzàng jì jí 出三藏記集 (T2145.55.52c27-53b27, 74c11-76b18). In the case of the Chinese canon, our references are always to the Taishō Tripiṭaka. For a helpful introduction to the life and work of Kumārajīva, see Robinson, Early Mādhyamika in India and China, 71–95.

9 The Dà zhìdù lùn itself explains what an upadeśa (also translated as “章句” or “論議經”, or transliterated as “優波提舍”) is, see T1509.25.308a17-b4. Regarding the genre of upadeśa and its influence on Chinese commentary traditions, see Wáng, “南北朝佛教解經學文體源流略考”.

10 See Chou, “庫車所出《大智度論》寫本殘卷之研究” on the dating of these fragments.

11 Huìjiǎo 慧皎, Gāosēng zhuàn 高僧傳, T2059.50.332c26-333a6.

12 This is from Kumārajīva’s assistant Sēngruì and his correspondent Huìyuǎn, both of whom nevertheless knew no Sanskrit. See T2145.55.75a28-b1, 75b13-18, 76a29-b2.

13 This opinion is most clearly voiced in Miyaji Kakue’s “智度論の本文批評に於ける一觀點”. The literature cited in §1, note 3 provides further references.

14 See Conze, The Prajñāpāramitā Literature, 35–6, 93–4; Demiéville, “Traité”; Lamotte, Der Verfasser des Upadeśa und seine Quellen, 5–21. Richard Robinson goes a little further than the others. While rejecting some Japanese scholars’ attribution of the whole text to Kumārajīva, the Wunderkind born of an Indian father and a Kuchean mother and known for his conversion from the Sarvāstivāda to the Mahāyāna, Robinson indicates that some parts of the Dà zhìdù lùn may have been authored by Kumārajīva (Early Mādhyamika in India and China, 38).

15 We should also avoid assigning a mere passive role to the scribes. As Rafal Felbur writes, “the process of turning the text from a rough draft into a refined literary artifact could well have continued without Kumārajīva’s involvement, and […] it is highly likely that he would not have been in any position to make suggestions at this stage in the process. This final product was then twice removed from ‘the original’: first by its oral translation and its mediation through public discussion, and second by the written transformation of the resulting translation by the scribes and polishers” (Felbur, “Kumārajīva”, 9).

16 The Aśokāvadāna may have been fashioned into its present form by the Buddhist Sanskrit community of Mathurā. See Strong, The Legend of King Aśoka, 26–8. It seems that the composer(s) of the Mahāprajñāpāramitopadeśa drew on the Aśokāvadāna as a repository for stories. In juàn 3 of the Dà zhìdù lùn, for example, we find the story transmitted in the Aśokāvadāna about the elder Mahākāśyapa entering a meditative trance; in juàn 11, the tale of a young bhikkhu whose breath smelled like perfume; in juàn 20 that of Aśoka giving his brother Vītaśoka a seven-day reprieve; and in juàn 12 and juàn 32 Aśoka’s gift of dirt.

17 There are conflicting accounts about the date of the Āyù wáng zhuàn (T2042). Confusion has been created by the fact that there are several different translations of the Aśoka legends. Most scholars follow Zhìshēng’s 智昇 Kāiyuán shìjiào lù 開元釋教錄 (730 CE), the most important Buddhist scriptural catalogue compiled in East Asia, which attributes a translation entitled Āyù wáng zhuàn, probably the one now under this title, to Ān Fǎqīn 安法欽 and dates it to the Western Jin dynasty 西晉 (266–316 CE) (T2154.55.623a17-21). Recently, Antonello Palumbo has suggested that the Āyù wáng zhuàn should be dated to a period later than Kumārajīva (Palumbo, “Models of Buddhist Kingship in Early Medieval China”, 311). Po-chien Lin 林伯謙, while insisting on the attribution to Ān Fǎqīn, argues that the Āyù wáng zhuàn is perhaps a selective compilation of already translated Aśoka stories and therefore belongs to the category of chāo jīng 抄經 (compiled scriptures). But, as Sēngyòu makes clear (T2145.55.37c1-9), and as Lin himself emphasizes, chāo jīng should not be conflated with spurious scriptures. None of the ancient catalogues characterizes Ān Fǎqīn’s Āyù wáng zhuàn as a chāo jīng, let alone as a spurious scripture (Lin, “《付法藏因緣傳》之譯者及其真偽辨”, 122). For our purposes suffice it to say that there is no evidence that the disciple’s tale in the Āyù wáng zhuàn does not belong to the second-century Aśokāvadāna. The most plausible hypothesis is thus that the composer(s) of the Mahāprajñāpāramitopadeśa adapted the disciple’s tale they found in the Aśokāvadāna, just as they took several other stories that have indeed been preserved in the extant Sanskrit text of this legend.

18 T2042.50.123b3-20. The story is translated in Przyluski, La légende de l’empereur Açoka, 380–1. In a Chinese text entitled Āyù wáng jīng 阿育王經, purportedly produced by 僧伽婆羅 *Saṅghabhara and dated to the Liang dynasty (502–557 CE), we have another version of the story: when the second demon dragged at the man’s arm, the first demon prevents the arm from being ripped off by pulling it in the opposite direction. The two demons spent the whole night performing this dragging and pulling on every part of the man’s body. In the end the man survived, bruised but intact (T2043.50.165b11-c2). A translation of this variant of the story is available in The Biographical Scripture of King Aśoka, 164–5. Interestingly, that the protagonist’s original body was eaten by the first demon, whom he in fact helped, seems always to be a problem for later readers. Chinese commentators changed the plot in having the second demon (they say he is the younger or the bigger one) bite off and immediately swallow the man’s limbs and the first demon (the older or the smaller one) feel guilty and thus replace them by corresponding parts of the corpse (see e.g. Āmítuó jīng shū 阿彌陀經疏, T1757.37.316a10-23).

19 The disciple’s tale and the neighboring stories in the Āyù wáng zhuàn aim to convey a simple, ascetic ideal. In its associative horizon there clearly lies the practice of monks meditating on the body. In this technique they contemplate, using anatomical analysis, the various parts of their body as well as corpses in different stages of dismemberment, which leads them to reflect on the mortality and impurity of their own body. Patrick Olivelle gives a brief introduction to the corpse meditation and its root in Indian ascetic traditions in his “Deconstruction of the Body in Indian Asceticism”. Indeed, ‘seeing one’s body as a corpse’ is a recurring motif in Upagupta tales, see Strong, The Legend and Cult of Upagupta, 83–5.

20 For the therapeutic character of the no-self doctrine, see Ganeri, “A Return to the Self”.

21 This is a translation of *Mañjuśrīparipṛcchā (Questions of Mañjuśrī) produced by 僧伽婆羅 *Saṅghabhara in 518 CE.

22 See Miàofǎ liánhuá jīng wénjù 妙法蓮華經文句, T1718.34.16b29-c16; 阿彌陀經疏, T1757.37.316a10-23; Dà fāngguǎng fó huáyánjīng suíshū yǎnyì chāo 大方廣佛華嚴經隨疏演義鈔, T1736.36.658a16-29. As further evidence of the longevity and cultural importance of our story, we may note that it has most recently been ‘reembodied’ in an opera 歌仔戲 titled Revata Encountered Demons 離婆多遇鬼 by the Taiwanese Dà Ài Technology Company. The ‘body-swapping’ episode (2014) is available on YouTube: https://www.youtube.com/watch?v=qz-OYutz3XU (last accessed 9 January 2021); the ‘Is this me?’ episode: https://www.youtube.com/watch?v=7w8F-nfRxbA.

23 In Chinese Buddhist literature it is customary to refer to the story translated here as “Two demons fought over a corpse” (二鬼爭屍). See e.g. 妙法蓮華經文句 (T1718.34.16c3-12), 大方廣佛華嚴經隨疏演義鈔 (T1736.36.658a16-29), and Zhǐguān fŭxíng zhuàn hóngjué 止觀輔行傳弘決 (T1912.46.375a13-b7).

24 See the version by Raymond Gressieux and Anita Ganeri (in: Ganeri, The Concealed Art of the Soul, 214–5), as well as the version by Chodron, in The Treatise on the Great Virtue of Wisdom of Nāgārjuna, Vol. II, 585–6 (this has not been formally published).

25 Lamotte, Le traité de la grande vertu de sagesse de Nāgārjuna (Mahāprajñāpāramitāśāstra), tome II, 738–40. We must note that Lamotte copied, almost word-for-word, Chavannes’s translation of a parable transmitted in the Zhòngjīng zhuànzá pìyù 眾經撰雜譬喻 (Parables Selected from Various Sūtras), cf. Chavannes, Cinq cents contes et apologues extraits du Tripiṭaka chinois, tome II, 72–4. The Chinese text of this parable (T208.4.531c25-532a17) corresponds for the most part to Kumārajīva’s rendition of the body-swapping story in the Dà zhìdù lùn.

26 Let us, in the space available, give a few examples, all taken from the first few lines of our narrator’s speech, to show why Lamotte’s Traité should be treated with caution. (1) When his opponent finishes with his three arguments against the no-self doctrine, our narrator begins his reply with the remark: “此俱有難!” — “All of these [sc. all your arguments against the no-self doctrine] have refutations!” Lamotte, however, partly confused about the grammatical structure of this sentence and partly confused as to how the argument develops, translates it as “This difficulty is common to us”. (2) Our narrator goes on to talk about the relation between one’s self and the body (身). Lamotte’s mistranslation of “身” as “personne” (“他身” therefore as “la personne d’autrui”) makes the whole debate between the narrator and his opponent crumble and is also at variance with the (correct) translation of the same word in the traveller’s story as “corps” (“他身” accordingly as “le corps d’un autre”), which he copies from Chavannes. (3) “五陰相續” (skandha-saṃtāna “the continuity of the five aggregates”), which our narrator then proposes as the origin of the view that one is one’s body, is translated by Lamotte as “relation to the five aggregates”. This shows that he fails to recognize this important Buddhist concept. (4) A further distortion is created by Lamotte’s rendering of “彼我” (the distinction of other and I) as “the self of a third person”, which turns our narrator’s thesis—the existence of the spirit (神) is a premise for “彼我”—into pure nonsense.

27 Stefano Zacchetti, who was working on a full English translation of the Dà zhìdù lùn, tragically passed away in 2020, before his project could be brought to fruition.

28 That the sentence below “自於身生我” is paraphrased later as “自身中生計我心” shows that “生計我” is probably an elliptical form of “生計我心”. While the word ‘心’ (citta) means ‘mind’ or ‘consciousness’, the verb ‘計’ has ‘reckon’ as its basic meaning. We notice that in Kumārajīva’s rendition of the Pañcaviṃśatisāhasrikā Prajñāpāramitā Sūtra 摩訶般若波羅蜜經, on which the Mahāprajñāpāramitopadeśa comments, and which has Sanskrit parallels preserved, he uses “校計” to translate pratyavekṣate (T223.8.273c24-26). This strongly suggests that the word ‘計’ in our text means ‘conceive’ or ‘think’.

29 There are three arguments against the non-existence of an I. We have only translated the first argument, to which the traveller’s story is intended as an answer.

30 “All of these” refers to all the three arguments against the non-existence of an I.

31 The five aggregates (‘五眾’, also called ‘五蘊’ or ‘五陰’), which are believed to be constitutive of a human being, include matter (rūpa), feeling (vedanā), labelling (saṃjñā), dispositions (saṃskāra), and discriminative consciousness (vijñāna). For a detailed discussion of the five aggregates, see Ganeri, Attention, Not Self, 76–81.

32 Satkāyadṛṣṭi has twenty possible permutations in relation to the five aggregates. 1–5: the self is identical with each of the five aggregates, 6–10: a self possesses each of the five aggregates, 11–15: a self exists within each of the five aggregates, and 16–20: each of the five aggregates exists within a self.

33 Ignorance (avidyā) is one of the twelve causes and conditions (pratītyasamutpāda).

34 Vāsanā (mental tendency, trace; lit. ‘perfume’) comes to be conceptualized as the ‘habitus’ which one’s actions (karma) imprint in one’s mind and whose accumulation predisposes one to act in a certain way. ‘習/習氣’ is discussed in detail in juàn 84 of the Dà zhìdù lùn.

35 The questioner has various descriptions of ‘神’, which is the word for ‘soul’ or ‘spirit’ in Classical Chinese. He first takes it as present in everything and everyone. Later he describes it as an entity and equates it roughly with the subtle body (細身 liṅga-śarīra), which resides in the gross body (麁身 sthūla-śarīra), entering at birth and leaving at death. The subtle body, he says, “enters the five gatis in each rebirth constantly”. From the refutation of his second and third arguments against the no-self doctrine we also learn that he seems to characterize ‘神’ as “constant” (常) and “unconfined” (自在), claims that it “acts by itself” (自作), and takes it as “matter” (色 rūpa) and “person” (人) (T1509.25.149a5-150a17).

36 In Buddhist literature, the hare’s horn (兔角) and the horse’s horn (馬角) are classical examples of things that do not exist, but for which we have concepts. Although in Chinese texts there is no indication of whether ‘兔角’ and ‘馬角’ are singular or plural, the two terms have often been rendered as plurals in translations, for example in Lamotte’s Traité. For some commentators assume that the phrase ‘兔角龜毛’, which at times denotes deep-rooted delusions, refers to the circumstances where the hare’s two ears are wrongly perceived as two horns and strands of algae that grow on the body of a turtle are seen as its fur. We translate the two words as singular mainly because the Sanskrit term for ‘兔角’, śaśaśṛṅga, is singular. Another reason is that the point of ‘兔角’ or ‘馬角’ may be that it is purely fictional, rather than based on an optical error, as shown by the Chinese idiom ‘兔角牛翼’ (‘牛翼’ means ‘the wings of cattle’), and in this case the horned horse is probably a unicorn and the horned hare, like al-mi‘rāj in Muslim iconography, a minor unicorn.

37 There are ten ‘spheres of totality’ (kṛtsnāyatana), namely the totality of earth, water, fire, wind, space, blue, yellow, red, white, and consciousness. For the ten kṛtsnāyatanas see Wynne, The Origin of Buddhist Meditation, 26–30.

38 In Indian tradition, the five elements of the material world are earth, water, fire, wind, and ‘sky’ or ‘space’ (ākāśa).

39 There are four types of viparyāsa: taking pain as pleasure, taking impermanence as permanence, taking non-self as self, and taking the impure as pure.

40 We suggest that the replier is actually speaking of a viparyāsa of non-Buddhists here. It seems that the previous sentence “顛倒故, 於他身中亦計我” is repeated here in a paraphrased form both for the sake of emphasis and in order to further the discussion, in which the replier reinforces his conclusion with the example of the traveller. The ellipsis of “because of viparyāsa” may have been motivated primarily by stylistic considerations.

41 The word ‘鬼’ can refer to either ‘demon’ or ‘ghost’. This conforms to what we know about the precedents for our story. In the tale transmitted in the Āyù wáng zhuàn 阿育王傳, it is two yakṣas (夜叉) that fought over a corpse (T2042.50.123b3-20). The Āyù wáng jīng 阿育王經 mentions two rākṣasas (羅剎) (T2043.50.165b11-c2). Yakṣas and rākṣasas, mythological beings in Buddhism that consume raw flesh, can be placed somewhere in between demons and ghosts.

42 It is difficult to determine whether the word ‘前’ in “到前國土” describes a spatial or temporal relation, namely, whether it means ‘front’ or ‘previous’. Both “[the man] reached the land ahead” (or Dharmamitra’s more liberal rendering “he reached the neighboring country”) and “[the man] reached the land from where he came” make sense here (although the second option seems more natural in Chinese). What we can directly rule out is Chavannes’s suggestion, which Lamotte follows: “étant arrivé au royaume dont il a été question plus haut”, because no land is mentioned previously in the story. However, the precedent for our story in the Āyù wáng zhuàn provides an important clue to the meaning of “前”. In that tale, Upagupta’s disciple, who was on his way from the master’s dwelling to his family, encountered two yakṣas fighting over a corpse, and the subsequent replacement of his body by parts of the corpse led to his return to Upagupta and to his final deliverance. It seems reasonable to assume that the Dà zhìdù lùn story retains its basic structure: departure–transition–return.

43 Or: “he saw that there were Buddhist stūpas and many monks”.

44 The narrator plays on the double meaning of ‘人’ (rén): person and human being. In Classical Chinese, the customary expression for asking someone’s identity is ‘which person are you?’ (汝是何人?). Although under normal circumstances, its meaning is not importantly different from that of ‘who are you?’, we translate this sentence here literally not only in order that the man’s reply makes more sense. ‘汝是何人?’ also has an implication that is absent in ‘who are you?’: the man appears in the eyes of the monks as a human being, not as something else, say, a ghost. And his reply “我亦不自知是人非人” should be understood as conveying that he himself does not know whether after the replacement of his body by a corpse he is still a human being or now a being that is not human (非人 amānuṣya), let alone, a fortiori, which person he is.

45 Buddhism understands the human body as composed of earth, water, fire, and wind.

46 The arhat is the one who severs the kleśas and attains an awakening by following the teachings of another.

47 Mūlamadhyamakakārikā, 18.2. See Westerhoff, Nāgārjuna’s Madhyamaka, 153–64.