Introduction

In the present chapter I study the Karajakāya-sutta in the light of its parallels, based on a translation of its Madhyama-āgama counterpart. The Karajakāya-sutta,[n.1] Title adopted from Be and Ce. found among the Tens of the Aṅguttara-nikāya, begins by highlighting the point that the results of intentionally performed deeds will inevitably have to be experienced.[n.2] AN 10.208 at AN V 299,11: "I do not say, monks, that without being experienced there is a coming to an end of intentional deeds that have been undertaken and accumulated", nāhaṃ, bhikkhave, sañcetanikānaṃ kammānaṃ katānaṃ upacitānaṃ appaṭisaṃviditvā vyantibhāvaṃ vadāmi (Be reads appaṭisaṃveditvā and byantībhāvaṃ, Ce vyantībhāvaṃ). Similar proposals on the definite relationship between karma and its fruit can be found elsewhere in the Pāli canon.[n.3] Statements with similar implications can be found, e.g., in Dhp 127, Ud 5.4 at Ud 51,16 and Sn 666; for further references cf. Dayal 1932/1970: 190, Mc- Dermott 1980: 176 and Krishan 1997: 66–69. The statement in AN 10.208 is quoted in Kv 466,11 (12.2) in the context of a discussion of karma and its fruit; cf. also McDermott 1975: 427. Van Zeyst 1965: 654 notes that the opening proposal in AN 10.208 does not leave much room for the later notion of ahosi kamma, listed at Paṭis II 78,1 and explained in Vism 601,6 to stand for karma that does not have a result, e.g., karma that was bound to ripen in the present life but did not get an opportunity to do so.

Such statements need to be contextualized with other passages which clarify that this relationship does not operate in a rigidly predetermined manner.[n.4] That the early Buddhist karmic theory is not predeterministic has been pointed out repeatedly; cf., e.g., Jayatilleke 1968, Gómez 1975: 82, Ñāṇaponika 1975: 91, Story 1975: 74, Fujita 1982: 151, Siderits 1987: 153, Jayawardhana 1988: 408, de Silva 1991: 273, Siridhamma 1998: 62–67, Halbfass 2000: 102, Harvey 2000/2005: 23, Hershock 2005: 6–7, Nelson 2005: 4, Harvey 2007: 59 and Pāsādika 2007a: 319. This is reflected in a simile about a piece of salt, found in a discourse in the Aṅguttara-nikāya and in its Madhyama-āgama parallel. According to this simile, just as a piece of salt will have a different effect on the drinkability of water if it is thrown into a small cup of water or into a large river, so too the effect of a particular deed depends on the overall moral development of the performer of this deed.[n.5] AN 3.99 at AN I 250,1 and MĀ 11 at T I 433a21. According to the Mahākammavibhaṅga-sutta and its parallels, this is the case to such an extent that even someone who commits a particularly bad deed may nevertheless be reborn in heaven if actions undertaken during his or her life have otherwise been of a wholesome nature.[n.6] MN 136 at MN III 214,20, MĀ 171 at T I 708b16 and D 4094 ju 267a7 or Q 5595 thu 10b2; cf. also the *Mahāprajñāpāramitā-(upadeśa-)śāstra, T 1509 at T XXV 238b17, translated in Lamotte 1944/1970: 1535. Conversely, one who has performed a remarkably wholesome deed may be reborn in hell due to having otherwise committed predominantly bad deeds.

A complementary perspective is provided in a discourse in the Saṃyutta-nikāya and its parallels in the two Chinese Saṃyuktaāgama translations. These discourses point out that, in addition to the fruition of former deeds, a range of other factors could be having their conditioning impact on the nature of present moment's experience, such as bodily disorders, change of climate, or external violence.[n.7] SN 36.21 at SN IV 230,13, SĀ 977 at T II 252c21 and SĀ2 211 at T II 452b14. Keown 1996: 340 sums up: "the Buddhist theory of karma does not rule out the possibility of accidents and adventitious misfortune". That is, in early Buddhism the inevitability of karmic retribution is seen to operate within a network of various causes and conditions and therefore does not exercise its influence in a mono-causal manner.

What makes the Karajakāya-sutta worthy of further attention is that it continues with another statement, which runs: "Yet, monks, I do not say that there is a making an end of dukkha without having experienced [the fruits of] intentional deeds that have been undertaken and accumulated."[n.8] AN 10.208 at AN V 299,14: na tvevāhaṃ, bhikkhave, sañcetanikānaṃ kammānaṃ katānaṃ upacitānaṃ appaṭisaṃviditvā dukkhass' antakiriyaṃ vadāmi (Be: appaṭisaṃveditvā).

This statement is surprising, since elsewhere the discourses associate such a position with Jain thought. Thus the Devadaha-sutta and its Madhyama-āgama parallel report that contemporary Jains undertook ascetic practices in the belief that the karmic fruition of formerly undertaken bad deeds needs to be expiated. According to their reasoning, if in addition no new deeds are undertaken, all dukkha can be eradicated and liberation will be won.[n.9] MN 101 at MN II 214,8: "by eradicating former deeds through asceticism and by not doing new deeds there will be no flowing on in future, from no flowing on in future [comes] the destruction of deeds, from the destruction of deeds [comes] the destruction of dukkha, from the destruction of dukkha ... all dukkha will be exhausted", purāṇānaṃ kammānaṃ tapasā vyantibhāvā, navānaṃ kammānam akaraṇā āyatiṃ anavassavo, āyatiṃ anavassavā kammakkhayo, kammakkhayā dukkhakkhayo, dukkhakkhayā ... sabbaṃ dukkhaṃ nijjiṇṇaṃ bhavissati. MĀ 19 at T I 442c3: "if those former deeds are eradicated through ascetic practice and no new ones are created, then all deeds [will be] eradicated, [if] all deeds have been eradicated, then the eradication of dukkha is attained, [if] the eradication of dukkha is attained, then the end of dukkha is attained", 若其故業因苦行滅, 不造新者, 則諸業盡, 諸業盡已, 則得苦盡, 得 苦盡已, 則得苦邊. This seems to be a fairly accurate record of Jain thought (for a discussion of some other Pāli discourses that report Jain theories about karma cf. Jain 1966). The Devadaha-sutta and its parallel continue by presenting a criticism of the assumptions underlying such a belief, with the Buddha interrogating the Jains if they had any knowledge of their former bad deeds or were aware of the resultant karma being diminished. Since the Jains have to deny both, the Buddha concludes that their belief is ill founded and their ascetic practices fruitless. Another discourse sets a clear contrast to the Jain approach by proclaiming that the goal of living the celibate life (brahmacariya) in the early Buddhist monastic tradition is not to alter or eradicate karmic retribution, but to develop insight into the four noble truths.[n.10] AN 9.13 at AN IV 382,3.

The distinctive nature of the early Buddhist conception regarding the relation between karma and liberation can be seen from the tale of the former brigand Aṅgulimāla. According to this tale, even a serial murderer can reach full awakening within the same lifetime, too short a time span to allow for the karmic retribution of recurrent killing to be fully exhausted.[n.11] That he had not yet exhausted the karmic retribution of his killings is reflected in MN 86 at MN II 104,14, EĀ 38.6 at T II 721a29 and T 119 at T II 511c27, which reckon his experience of being attacked and beaten up when going to beg alms, an event that took place after he had become an arahant, to be in retribution for his former bad deeds. Though intentionally performed murder is inevitably followed by karmic retribution, once full awakening has been attained, this karmic retribution can come to fruition only during the remainder of that single lifetime.[n.12] This principle is succinctly summed up in Th 81: "whatever evil done by me previously in other births, it is to be experienced here and now, as there exists no other occasion [for it to ripen]", yaṃ mayā pakataṃ pāpaṃ, pubbe aññāsu jātisu, idh' eva taṃ vedaniyaṃ, vatthu aññaṃ na vijjati. Th-a I 186,15 explains that the speaker was an arahant. Applying this perspective to the stanza found in MN 86 at MN II 104,23 (= Th 872) and Dhp 173 then suggests that their reference to closing off evil deeds by what is wholesome may refer to the same principle, in that through having attained awakening the scope for the ripening of karmic retribution is confined to the present lifetime. As explained by Main 2005: 2, "arahats, by means of their attainment, have cut off the causes for rebirth and will only experience the results of past action until their death. In this short span of time, much of their store of kamma will not have the opportunity to come to fruition". Within that last lifetime, however, karma will still bear its fruit, in fact even the Buddha is reckoned as having been subject to the experience of karmic retribution; cf., e.g., Walters 1990, Cutler 1997, Hara 1997: 250–253 and Guang Xing 2002.

In sum, the above quoted statement from the Karajakāya-sutta does not fit well with what can be gathered about the conception of karma and its relationship to liberation elsewhere in the early Buddhist discourses.[n.13] The idea that Nirvāṇa can be attained through good karma, advanced in Khp 7,19+27 (8.9+13) or in Mil 341,23, is a later development; cf. McDermott 1973, McDermott 1977: 466, McDermott 1984/2003: 118–122 and Schmithausen 1986: 207. Instead, this statement appears to correspond to the Jain position on this issue. In Jain thought, karma is perceived as a material substance that, as a result of a person's activity, adheres to the self and will fall away again once its effect has been experienced.[n.14] Cf., e.g., von Glasenapp 1915: 19–20, Tatia 1951: 220–260, Mehta 1957: 13– 30, Schubring 1962: 172–185, Kalghatgi 1965, Jaini 1979/2001: 111–127, Dundas 1992/2002: 97–102, Johnson 1995, Bronkhorst 2000: 119, Halbfass 2000: 75–85 and Jain 2005: 248–259. To transcend this process requires the expiation of karmic retribution through asceticism. Thus the Dasaveyāliya-sutta (Daśavaikālika-sūtra), one of the four Mūlasūtras of the Jain canon, proclaims that liberation will be attained only when one's former bad deeds have been experienced, emphatically asserting that it is impossible to reach liberation without karmic retribution being either experienced or else expiated through asceticism.[n.15] and the Tibetan version, would then imply that the development of the brahmavihāras has the potential of leading to non-returning or full liberation if – following the suggestion in Mp V 78,7 (cf. also the similar progression of ideas in the Yogācārabhūmi, T 1579 at T XXX 808b24) – such brahmavihāra practice is undertaken by those who have already reached the insight of a stream-enterer or once-returner, something to which AN 10.208 at AN V 300,14 seems to be implicitly referring by qualifying the monk who reaches non-return as endowed with wisdom. This statement corresponds closely to the proposal made at the beginning of the Karajakāya-sutta that, to make an end of dukkha, karmic retribution has to be experienced.

The Karajakāya-sutta has parallels in a discourse in the Madhyama-āgama, translated below, and in Śamathadeva's compendium of discourse quotations in the Abhidharmakośabhāṣya, extant in Tibetan.[n.16] [18] D 4094 ju 236b2 to 238b5 or Q 5595 tu 270a3 to 272b5; translated in Martini 2012. Both of these parallel versions begin their exposition by affirming the inevitability of karmic retribution, but differ from the Karajakāya-sutta in that they do not make a statement about the need of experiencing karmic retribution in order to be able to make an end of dukkha.[n.17] [19] MĀ 15 at T I 437b27 and D 4094 ju 236b3 or Q 5595 tu 270a5, which instead point out that unintentionally performed deeds do not entail karmic retribution.

---

Translation

Discourse on Intention[n.18] The translated text is MĀ 15 at T I 437b24 to 438b11.

Thus have I heard. At one time the Buddha was dwelling at Sāvatthī, staying in Jeta's Grove, Anāthapiṇḍika's Park.[n.19] AN 10.208 at AN V 299,11 does not give any information on the Buddha's whereabouts, but opens directly with the statement on the inevitability of karmic retribution. At that time, the Blessed One told the monks:

"If [someone] performs deeds intentionally, I say that he will inevitably have to experience their fruits, either experiencing them in this life or experiencing them in a later life.[n.20] AN 10.208 at AN V 299,13 and D 4094 ju 236b2 or Q 5595 tu 270a5 speak of three time periods for experiencing karmic retribution; on which cf. von Hinüber 1971 and Anālayo 2011a: 779f note 118. If [someone] performs deeds unintentionally, I say that he will not necessarily have to experience their fruits.

"Herein, three are the [types] of intentionally performed bodily deeds that are unwholesome, that result in the experience of dukkha and that have dukkha as their fruit; four are the [types] of verbal deeds and three are the [types] of mental deeds that are unwholesome, that result in the experience of dukkha and that have dukkha as their fruit.[n.21] AN 10.208 has neither a summary statement nor a detailed exposition of the ten unwholesome actions.

"What are the three [types] of intentionally performed bodily deeds that are unwholesome, that result in the experience of dukkha and that have dukkha as their fruit? Killing living beings is reckoned the first. [Someone is] supremely evil and blood-thirsty,[n.22] MĀ 15 at T I 437c3: 飲血, literally "drinking blood", which perhaps renders an expression similar to the Pāli lohitapāṇī, "bloody-handed", mistaking the second member of this compound to be related to pāna or pānīya, "drink". having the wish to injure and being without compassion for living beings, including insects.[n.23] MĀ 15 at T I 437c3: 蜫蟲 (according to Hirakawa 1997: 1037f, 蜫 renders kunta and 蟲 can render kṛmi or kīta, hence my rendering as "insects"). Since a detailed treatment of the ten unwholesome actions is not given in AN 10.208, a comparison can only be made with the description of killing in AN 10.206 at AN V 292,15, which, however, just refers to beings in general, sabbapāṇabhūtesu, without any further specification; on this type of difference cf. also Anālayo 2011a: 190. The corresponding section in D 4094 ju 236b5 or Q 5595 tu 270a8 mentions the killing of ants, grog sbu.

"Taking what is not given is reckoned the second. Out of attachment he takes the possessions of others with the intention of stealing.

"Sexual misconduct is reckoned the third. He has intercourse with a woman that is protected by her father, or protected by her mother, or protected by both parents, or protected by her sister, or protected by her brother, or protected by the parents-in-law, or protected by relatives, or protected by the clan; or a woman protected by threat of corporal punishment,[n.24] MĀ 15 at T I 437c7: 有鞭罰恐怖, literally "having whip-punishment fear", which I assume to be rendering an original similar to AN 10.206 at AN V 292,20: saparidaṇḍā, literally "with a stick around", an idiomatic expression indicating that a woman is protected by threat of corporal punishment (cf. Vin III 139,35); with a counterpart in D 4094 ju 236b7or Q 5595 tu 270b3: chad pa dang bcas pa. For a survey of different types of women who are reckoned to be under protection cf. also Collins 2011: 51. even one who has been garlanded in token of betrothal.[n.25] MĀ 15 at T I 437c8: 及有名假賃至華鬘, literally: "even [one who] is conceptually designated [as] engaged, having obtained a flower garland", which I take to be similar in meaning to AN 10.206 at AN V 292,20: antamaso mālāguṇaparikkhittā (Be: mālāguḷaparikkhittā, Ce: mālāgulaparikkhittā); on this expression cf. also Silk 2007a: 7. D 4094 ju 236b7 or Q 5595 tu 270b3 qualifies such a girl as yongs su bzung ba la phreng ba ʼphangs te.

"These are reckoned the three types of intentionally performed bodily deeds that are unwholesome, that result in the experience of dukkha, and that have dukkha as their fruit.

"What are the four [types] of intentionally performed verbal deeds that are unwholesome, that result in the experience of dukkha and that have dukkha as their fruit? Speaking falsehood is reckoned the first. On being questioned in an assembly, or among family members, or in the king's palace thus: 'say what you know', he claims to know what he does not know, or claims not to know what he knows; claims to have seen what he has not seen, or claims not to have seen what he saw; he knowingly speaks falsehood either for his own sake or for the sake of others, or for the sake of wealth.

"Divisive speech is reckoned the second. Wishing to divide others, he tells those what he has heard from these, out of a wish to harm these, [or else] tells these what he has heard from those, out of a wish to harm those. Wishing to divide those who are united and further to divide those who are already divided, he forms factions, delights in [the forming of] factions and praises [the forming of] factions.

"Harsh speech is reckoned the third. He employs a type of speech that is rough and rude in tone, which sounds offensive and grates on the ear, that beings neither enjoy nor desire, which causes others suffering and vexation, and which does not lead to calmness, speaking such type of speech.

"Frivolous talk is reckoned the fourth. He speaks at the wrong time, speaks what is not true, what is not meaningful, what is contrary to the Dharma, what does not [lead] to appeasement, and also commends issues that do not [lead] to appeasement. Disregarding the [proper] timing, he does not teach or admonish properly.

"These are reckoned the four types of intentionally performed verbal deeds that are unwholesome, that result in the experience of dukkha, and that have dukkha as their fruit.

"What are the three [types] of intentionally performed mental deeds that are unwholesome, that result in the experience of dukkha and that have dukkha as their fruit? Covetousness is reckoned the first. On seeing another endowed with wealth and all the necessities of life, he constantly has the wish and desire: 'May I get it!'

"Irritation and ill will are reckoned the second. With a mind [full of] dislike and irritation,[n.26] MĀ 15 at T I 437c25: 嫉恚. Hirakawa 1997: 367 indicates that 嫉, besides its main meaning of "jealousy", can also render pratighāta, which would fit the present context better. he has the thought: 'May those beings be killed, be bound, be arrested, be removed,[n.27] MĀ 15 at T I 437c26: 應收, 應免. Hirakawa 1997: 569 and 156 gives grahaṇa as a Sanskrit equivalent for 收, and pari-√hṛ for 免, hence my rendering as "arrested" and "removed". and be banished', having the wish that others experience infinite dukkha.

"Wrong view is reckoned the third. He has a view that is distorted, a view like this, declaring thus: 'There is no [efficacy in] giving, there is no [efficacy in] offerings [during a sacrifice], there is no [efficacy in] reciting hymns [during a sacrifice],[n.28] MĀ 15 at T I 437c28: 呪說, which together with the preceding 齋 appear to correspond to the reference to what is "offered" and "sacrificed", yiṭṭha and huta, in AN 10.206 at AN V 293,30. While the two Pāli terms are similar in meaning, the rendering in MĀ 15 seems to reflect two aspects of a sacrifice: the performance of offerings and the recitation of hymns. there are no wholesome and evil deeds, there is no result of wholesome and evil deeds, there is neither this world nor another world, there is no [obligation towards one's] father and mother,[n.29] The description of wrong view in AN 10.206 at AN V 293,32 adds a denial of the existence of spontaneously arisen beings, n' atthi sattā opapātikā; cf. also above note 22 page 296. Spontaneously arisen beings are also mentioned in D 4094 ju 237b6 or Q 5595 tu 271b4: sems can rdzus te byung ba rnams med do; on this difference in general cf. also Anālayo 2011a: 659f. in the world there are no worthy men who have reached a wholesome attainment,[n.30] MĀ 15 at T I 438a1: 真人, which judging from its usage elsewhere in the Madhyama-āgama (cf., e.g., MĀ 85 at T I 561a20) would render a term equivalent to the Pāli sappurisa, though in its general usage 真人 can also render "arahant". AN 10.206 at AN V 294,1 instead speaks of "recluses and brahmins", samaṇa-brāhmaṇā, and D 4094 ju 237b7 or Q 5595 tu 271b5 of arahants, dgra bcom pa. who are well gone and have progressed well, who by their own knowledge and experience abide in having themselves realized this world and the other world'.

"These are reckoned the three types of intentionally performed mental deeds that are unwholesome, that result in the experience of dukkha and that have dukkha as their fruit.

"The learned noble disciple leaves behind unwholesome bodily deeds and develops wholesome bodily deeds, leaves behind unwholesome verbal and mental deeds and develops wholesome verbal and mental deeds. That learned noble disciple, being endowed with diligence and virtue in this way, having accomplished purity of bodily deeds and purity of verbal and mental deeds, being free from ill will and contention, discarding sloth-and-torpor, being without restlessness[n.31] Adopting the variant 掉 instead of 調. or conceit, removing doubt and overcoming arrogance, with right mindfulness and right comprehension, being without bewilderment, dwells having pervaded one direction with a mind endowed with benevolence, and in the same way the second, third and fourth directions, [all] of the four directions, above and below, completely and everywhere. Being without mental shackles, resentment, ill will, or contention, with a mind endowed with benevolence that is supremely vast and great, boundless and well developed, he dwells having pervaded the entire world.

Then he reflects like this: 'Formerly my mind was narrow and not well developed; now my mind has become boundless and well developed.'

"When the mind of the learned noble disciple has in this way become boundless and well developed, if because of [associating with] bad friends he formerly dwelt in negligence and performed unwholesome deeds, those [deeds] cannot lead him along, cannot defile him and will not come back to meet him.

"Suppose there is a small boy or girl,[n.32] AN 10.208 at AN V 300,2 refers only to a boy, kumāro. Since all versions later on mention a man and a woman, the presentation in MĀ 15 fits the context better. D 4094 ju 238a6 or Q 5595 tu 272a5 also mentions both a boy and a girl, khyeʼu dang bu mo. who since birth is able to dwell in the liberation of the mind through benevolence. Later on, would [he or she] still perform unwholesome deeds by body, speech or mind?" The monks answered: "Certainly not, Blessed One."

"Why? Not performing evil deeds themselves, how could evil deeds arise? Therefore, a man or woman, at home [or] gone forth, should constantly make an effort to develop liberation of the mind through benevolence. If that man or woman, at home [or] gone forth, develops liberation of the mind through benevolence, [since] when going towards the other world [he or she] will not take this body along, [he or she] will proceed [just] in accordance with [the developed quality of their] mind.

"Monks, you should reflect like this: 'Formerly I was negligent and performed unwholesome deeds. Let the fruits of these be experienced entirely now, not in a later world'.[n.33] This reflection of the monks is not found in AN 10.208, although a similar reflection occurs in D 4094 ju 238b1 or Q 5595 tu 272b1. Limitations of space do not allow a full discussion of this significant difference, which seems to be another case of textual loss on the side of AN 10.208. Briefly stated, in MĀ 15 and the Tibetan version the present passage marks a shift from the theme of the rebirth prospects of a man or woman in general to practice undertaken by monks, where the reflection that the fruits of evilness shall be experienced entirely now represents an aspiration to develop insight in such a way that non-return or full awakening is realized within the same lifetime. According to the commentarial explanation at Mp V 77,17, the earlier passage on the man or woman developing the brahmavihāras implies that any 'limiting deeds' related to rebirth in the sensual realm have no scope to influence the next rebirth. That the issue at stake is the next rebirth finds support in other passages that take up the effect of the brahmavihāras on 'limiting deeds'. In DN 13 at DN I 251,7 such a reference (which is not found in the description of the brahmavihāra practice in the parallel version DĀ 26 at T I 106c17) occurs in the context of a teaching on how to reach union with Brahmā, as is the case in MN 99 at MN II 207,25 (where such a statement is not found in the description of the brahmavihāra practice in the parallel MĀ 152 at T I 669c10). The same type of statement in SN 42.8 at SN IV 322,13 and its parallels SĀ 916 at T II 232b5 and SĀ² 131 at T II 425b29 is addressed to a Jain householder who had proposed that unwholesome deeds inevitably lead to rebirth in hell. That is, all these passages are related to the nature of the next rebirth. This makes it highly probable that the effect of brahmavihāra practice on limiting deeds, described in AN 10.208 and its parallels, also refers to the next rebirth. Thus the present passage need not be understood as entailing a complete elimination of karma, pace Maithrimurthi 1999: 76 (cf. also Enomoto 1989a: 46–48 and Wiltshire 1990: 268), but would only intend the effect of karmic retribution on the next rebirth. The subsequent passage on the monks' aspiration, found only in MĀ If liberation of the mind through benevolence has become boundless and well developed like this, certainly non-returning will be attained, or else that which is even higher.

"It is the same with dwelling having pervaded the entire world with a mind endowed with compassion, sympathetic joy or equanimity, without mental shackles, resentment, ill will, or contention, [with a mind that is] supremely vast and great, boundless and well developed.

Then he reflects like this: 'Formerly my mind was narrow and not well developed; now my mind has become boundless and well developed.

"When the mind of the learned noble disciple has in this way become boundless and well developed, if because of [associating with] bad friends he formerly dwelt in negligence and performed unwholesome deeds, those [deeds] cannot lead him along, cannot defile him and will not come back to meet him.

"Suppose there is a small boy or girl, who since birth is able to dwell in the liberation of the mind through equanimity. Later on, would [he or she] still perform unwholesome deeds by body, speech or mind?" The monks answered: "Certainly not, Blessed One."

"Why? Not performing evil deeds themselves, how could evil deeds arise? Therefore a man or woman, at home [or] gone forth, should constantly make an effort to develop liberation of the mind through equanimity. If that man or woman, at home [or] gone forth, develops liberation of the mind through equanimity, [since] when going towards the other world [he or she] will not take this body along, [he or she] will proceed [just] in accordance with [the developed quality of their] mind.

"Monks, you should reflect like this: 'Formerly I was negligent and performed unwholesome deeds. Let the fruits of these be experienced entirely now, not in a later world.' If liberation of the mind through equanimity has become boundless and well developed like this, certainly non-returning will be attained, or else that which is even higher."

The Buddha spoke like this. The monks, having listened to what the Buddha said, were delighted and received it respectfully.

---

Study

As the above translated Madhyama-āgama discourse shows, the Karajakāya-sutta's proclamation that it is impossible to make an end of dukkha without experiencing karmic retribution – absent from the two parallel versions – comes at a point in the text where there is anyway a substantial difference between the Pāli discourse on the one side and the Chinese and Tibetan versions on the other side. The Pāli discourse at this point continues by referring to a noble disciple who practises the brahmavihāras. This passage gives an impression of abruptness, as it begins with: "that noble disciple, monks, in this way free from covetousness, ill will and delusion".[n.34] AN 10.208 at AN V 299,16: sa kho so, bhikkhave, ariyasāvako, evaṃ vigatābhijjho vigatavyāpādo asammūḷho. Woodward 1936/1955: 193 note 1 comments that "all this ... is introduced without apparent reason thus suddenly". Yet, the preceding part has not introduced any noble disciple, nor has it described in what way such a noble disciple has become free from covetousness, etc. Thus the way this sentence is formulated gives the impression that a loss of text has taken place, creating a lacuna that can be filled with the help of the parallel versions.

In the Chinese and Tibetan counterparts, the initial statement on karmic retribution is illustrated through a detailed exposition of the ten unwholesome actions, ranging from killing to wrong view. Then these two versions turn to the noble disciple, who abstains from these ten unwholesome actions and develops the brahmavihāras. It is at this point that the above-mentioned passage in the Karajakāya-sutta seems to fit in, with its reference to "that noble disciple" who is "in this way free from covetousness, ill will and delusion" – the last three of the unwholesome actions – and thus able to engage in the practice of the brahmavihāras.

The impression that the Karajakāya-sutta has lost an exposition of the ten unwholesome actions is further supported by its placement in the Tens of the Aṅguttara-nikāya. In its present form, the Pāli discourse does not have any aspect related to the number ten. This suggests that at an earlier time, when discourses were allocated to different sections in the four Nikāyas, the Karajakāya-sutta may still have had an exposition on the ten types of unwholesome actions.

Now, precisely at the point where the Karajakāya-sutta affirms that to make an end of dukkha requires experiencing karmic retribution, a rather substantial loss of text appears to have taken place. This, together with the absence of such a statement in the parallel versions, makes it quite probable that this statement is also the outcome of some error during transmission.[n.35] [37] My conclusions differ from those of Vetter 1988: 90, who seems to take the statement on the need to experience karmic retribution for being able to make an end of dukkha as original and then interprets the reference to the brahmavihāras in AN 10.208 as "an attempt to react to this idea", where "cultivating the four immeasurables is recommended as a means to overcome the power of former deeds".

Although the situation so far seems still relatively straightforward, the picture gets somewhat complicated by the fact that a detailed exposition of the ten unwholesome actions, such as what would have been found in an earlier version of the Karajakāya-sutta, occurs in the two discourses that precede the Karajakāya-sutta in the Aṅguttara-nikāya, the first and the second Sañcetanika-sutta.[n.36] [38] AN 10.206 at AN V 292,1 and AN 10.207 at AN V 297,14, title adopted from Be and Ce. The opening section of these two discourses also agrees with the Karajakāya-sutta, that is, the statement that making an end of dukkha requires experiencing the fruition of former deeds occurs not only in the Karajakāya-sutta, but also in these two discourses.

The parallelism between these discourses can best be illustrated with a survey of their main themes, given in table 20. After the introductory statement on karmic retribution and the need to experience it in order to make an end of dukkha, the two Sañcetanika-suttas (AN 10.206 and AN 10.207) give a detailed exposition of the ten unwholesome actions, similar to the Chinese and Tibetan discourses. Subsequently, the two Sañcetanika-suttas turn to the bad rebirth that is to be expected from undertaking the ten unwholesome actions, a prospect that the first Sañcetanika-sutta (AN 10.206) illustrates with the simile of a die which, when thrown up in the air, will fall down and stay where it has fallen. The remainder of the two Sañcetanika-suttas then repeats the same pattern for the opposite case of undertaking the ten wholesome deeds, which results in a good rebirth.

Table 20: Main Themes in the Exposition

Sañcetanika-suttas	Chinese & Tibetan	Karajakāya-sutta
karmic retribution	karmic retribution	karmic retribution
↓	|	↓
make end of dukkha	|	make end of dukkha
↓	↓	|
10 unwholesome acts	10 unwholesome acts	|
↓	↓	|
exposition of these 10	exposition of these 10	|
|	↓	↓
|	brahmavihāra	brahmavihāra
|	↓	↓
|	fruits of brahmavihāra	fruits of brahmavihāra
↓
bad rebirth
↓
repetition of above for
10 wholesome acts

The parallelism between the Sañcetanika-suttas and the Karajakāya-sutta in regard to their beginning parts, and between the Sañcetanika-suttas and the Chinese and Tibetan discourses in regard to the detailed treatment of the ten unwholesome actions, makes it probable that all these discourses were interrelated during oral transmission, in fact the Chinese and Tibetan versions are parallels to all three Pāli discourses.[n.37] Akanuma 1929/1990: 341 also lists SĀ 1047, SĀ 1048 and EĀ 48.1 as parallels to AN 10.206 and AN 10.207. Yet, SĀ 1047 at T II 274a6 takes up unwholesome deeds, mental states and views, 惡業, 惡心, 惡見, making it a closer parallel to AN 3.116 at AN I 270,9, which examines sīlavipatti, cittavipatti and diṭṭhivipatti, and illustrates their bad consequence with the simile of the die found also in SĀ 1047 at T II 274a21, as well as in AN 10.206 at AN V 294,15. The other two parallels mentioned by Akanuma, SĀ 1048 at T II 274a26 and EĀ 48.1 at T II 785c25, offer an examination of the karmic retribution for the ten unwholesome actions that has more in common with MN 135 at MN III 203,16 than with AN 10.206, even though MN 135 does not base its exposition on the ten unwholesome actions. Thus MĀ 15 and the discourse quotation in Śamathadeva's compendium seem to be the two chief parallels to AN 10.206, AN 10.207 and AN 10.208. Thus the way the Karajakāya-sutta and the Sañcetanika-suttas have been preserved in the Pāli canon could be the result of a garbling of what originally was a single discourse.

The phrasing of the statement shared by these three consecutive discourses might then be due to a relatively common type of transmission error, where either a negation is lost or else an originally positively worded phrase is negated.[n.38] Lack of space does not allow an exhaustive survey of this phenomenon, instead of which three examples drawn from the Majjhima-nikāya will have to suffice: 1) In a description of what appears to be an arrival at the final goal, MN 29 at MN I 196,29 reads samayavimokkha, where from the context one would rather expect asamayavimokkha, a reading found in fact in a repetition of the same passage at MN I 197,27, and throughout in Be and Ce, as well as in the commentary Ps II 232,3, whereas Se reads samayavimokkha. 2) The advice of a doctor in regard to a wound in MN 105 at MN II 257,4 reads alañca te antarāyāya, whereas Be and Ce read analañca te antarāyāya, and Se alañca te anantarāyāya (the editions also disagree as to whether this wound still contains poison and whether the doctor is aware of that). In this case Ee finds support in a parallel version in SHT IV 500 folio 3 V4, Sander 1980: 220, which reads ala[ṃ]te-t(r)-ānta[r](ā)yāya. 3) As a heading for an exposition of how clinging to a sense of self leads to agitation, MN 138 at MN III 227,26 speaks of anupādā paritassanā, a reading found not only in other Pāli editions but also in its Chinese parallel MĀ 164 at T I 695c19: 不受而恐怖, even though this is a clear misfit and the context would require "agitation due to clinging". In fact the reading upādā paritassanā is found in a similar treatment in SN 22.7 at SN III 16,3; cf. also Bodhi in Ñāṇamoli 1995/2005: 1350 note 1253. In sum, the above examples suggest transmission errors that involve: 1) loss of a negation in the PTS edition, 2) addition of a negation in the Asian editions, 3) addition of a negation in the Pāli editions and in the Madhyama-āgama parallel. These examples thus appear to testify to a tendency for negations to become lost or else be added during textual transmission. That is, the sentence in question might earlier have read paṭisaṃviditvā instead of appaṭisaṃviditvā, in which case the statement would have been: "Yet, monks, I do not say that there is a making of an end of dukkha through having experienced [the fruits of] intentional deeds that have been undertaken and accumulated."

Though this obviously remains hypothetical, as I am not aware of any variant that would support the suggested emendation, nevertheless, such a statement would better accord with early Buddhist teachings than the reading appaṭisaṃviditvā. By setting a contrast to the position adopted by the Jains, the suggested emendation paṭisaṃviditvā would highlight the point that, even though karma will definitely bear its fruit, nevertheless liberation is not to be won through expiating karmic fruition. The assumption that the text may have read paṭisaṃviditvā would also better suit the introductory phrase "yet" or "however", tveva (tu + eva), which would then introduce a real contrast to the preceding statement on the inevitability of karmic retribution.[n.39] The only recurrence of the pattern na tvevāhaṃ ... dukkhassa antakiriyaṃ vadāmi that I have been able to locate is SN 22.99 at SN III 149,27, where the point is also a refutation of a tenet held by contemporary recluses, preceded by a statement of a general principle (here on the nature of saṃsāra). This instance has a positively worded condition: sandhāvataṃ saṃsarataṃ, being in that respect similar to my suggested emendation. If the parallelism with SN 22.99 should be a valid indication, then in the passage under discussion one would expect a positively worded condition, i.e., paṭisaṃviditvā, instead of appaṭisaṃviditvā. The suggested emendation paṭisaṃviditvā would also fit the subsequent exposition in the Karajakāya-sutta, which culminates with the attainment of non-return or full awakening, clearly showing that, from the perspective of the remainder of the discourse, to make an end of dukkha does not require exhausting karmic retribution.

In sum, assuming the occurrence of a relatively common transmission error would fit better with the remainder of the Karajakāya-sutta and would place the teachings given in this discourse fully within the range of early Buddhist doctrine. In fact, although the Buddhist and Jain traditions obviously influenced each other in various ways, one would not expect the canonical texts of one tradition to uphold a position of the other tradition with which they elsewhere explicitly disagree, especially when it comes to such vital matters as the relationship between karma and liberation.

Yet another example where a Pāli discourse appears to have assembled Jain thought can be found in the first discourse of the third chapter of the Udāna. The udāna itself – the 'inspired utterance' that forms the key aspect of each discourse in this textual collection – begins by describing the aloofness "of a monk who leaves behind all karma, shaking off the dust of what formerly had been done".[n.40] [42] Ud 3.1 at Ud 21,11: sabbakammajahassa bhikkhuno, dhunamānassa purekataṃ rajaṃ (Se reads bhūnamānassa purekkhataṃ). According to the prose narration that precedes this stanza, the Buddha spoke this inspired utterance on seeing a monk seated in meditation who, "with mindfulness and clear comprehension was enduring without complaint painful, fierce, sharp and severe feelings that were the result of former deeds".[n.41] [43] Ud 3.1 at Ud 21,4: purāṇakammavipākajaṃ dukkhaṃ tippaṃ kharaṃ kaṭukaṃ vedanaṃ adhivāsento sato sampajāno avihaññamāno (Be and Ce read tibbaṃ, Se does not have kharaṃ). According to Ud-a 165,2, this monk was already an arahant, so that from the perspective of the commentators the monk's endurance of pain due to former deeds was not something that brought him closer to liberation.

What is striking about this stanza is the idea of "leaving behind all karma", sabbakammajaha, by "shaking off", dhunati, the dust of former deeds. The expression sabbakammajaha does not seem to recur in other Pāli discourses,[n.42] [44] A related expression would be sabbakammavihāyīnaṃ or sabbakammavihāyinaṃ found in Be, Ce and Se of AN 4.28, though Ee at AN II 29,1 reads sammā kammaviyākataṃ instead. though the term sabbakammakkhaya, "the destruction of all deeds", occurs as a term representing awakening.[n.43] Cf., e.g., SN 5.8 at SN I 134,4, AN 4.23 at AN II 24,19 and It 4.13 at It 123,3. The verb dhunati in the four Pāli Nikāyas and in verses found in the Sutta-nipāta and the Thera- and Therīgāthā tends to have a figurative sense as a "shaking off" or "casting off" of evil states.

Examples are descriptions where the disciples of the Buddha "cast off evilness",[n.44] [46] DN 30 at DN III 179,8: dhunanti pāpaṃ (Se reads panudi-pāpassa). or else a monk "casts off evil things", comparable to a tree shaking off its leaves in the wind.[n.45] [47] Th 2 and Th 1006: dhunāti pāpake dhamme. Similar in implication is the allegorical injunction to "shake off Māra's army", just as an elephant would knock over a hut made of reeds.[n.46] [48] SN 6.14 at SN I 156,35 and 157,20: dhunātha maccuno senaṃ, found again at Th 256; cf. also the similar expression dhunāma maccuno senaṃ at Th 1147 and 1149 (Se reads dhunāmu). A stanza specifies that the monks "cast off the three roots of evilness",[n.47] [49] Thī 276: tīṇi pāpassa mūlāni dhunanti. and in another stanza someone requests to have his "doubt cast off", in the sense of having the doubt dispelled by an explanation.[n.48] [50] Sn 682: dhunātha me saṃsayaṃ. In sum, in these contexts dhunati conveys a figurative sense of casting or shaking off detrimental mental qualities.

Thus the relationship established in the above quoted Udāna stanza between dhunati and the dust of former deeds as an illustration of leaving behind all karma appears to be unique in early Buddhist texts. Gombrich (1994: 1078ff) comments that the above stanza "sounds ... unequivocally Jain", giving the impression that it "was of non-Buddhist origin". In fact in Jain scriptures, as noted by Enomoto (1989a: 45), "the elimination of karman is expressed as 'shaking off (dhū-)', and the word raja (rajas, raya) ... implies karman".

Examples of such usage can be found in the Dasaveyāliya-sutta, which presents shaking off the dust of karma as the way to liberation,[n.49] Lalwani 1973: 53,27 and 54,8 (chapter 4 part 12 verses 20–21): dhunāti karmarajaḥ. explaining that the evilness from the past is to be shaken off through austerity.[n.50] Lalwani 1973: 198,9 and 203,8 (chapter 9 part 4 stanza 8 (here counted as 4) and chapter 10 stanza 7): tapasā dhunoti purāṇapāpakaṃ; for further examples cf. Yamazaki 2003: 498–499. In such passages, the idea of "shaking off" has a literal sense, as the point at stake is to free the soul from the material particles of karma that have been accumulated. Hence the above-quoted Udāna passage would be another case where a Pāli passage attributes Jain thought to the Buddha.[n.51] Enomoto 1989a: 45 points out that a similar image can be found in the Kauṣītaki Upaniṣad 1.4; cf. Radhakrishnan 1953: 757,4: tat-sukṛta-duṣkṛte dhunute vā ... sa eṣa visukṛto viduṣkṛto brahma-vidvān brahmaivābhipraiti, translated ibid.: "there he shakes off his good deeds and his evil deeds ... thus one, freed from good and freed from evil, the knower of Brahman, goes on to Brahman".

Turning to the parallel versions, instead of sabbakammajahassa the Chinese and Tibetan counterparts speak of "leaving behind all sensuality", indicating that they were based on a reading that had kāma instead of karma.[n.52] T 212 at T IV 765a6: 比丘盡諸愛, 捨愛去貢高, T 213 at T IV 796b20: 苾 芻諸愛盡, 捨愛去貢高, Beckh 1911: 131 (32.3): dge slong ʼdod pa kun spangs shing, mdun gnas pa yi rdul bsal ba, Zongtse 1990: 378 (32.3): dge slong ʼdod pa kun spangs shing, mdun na gnas paʼi rdul bsal ba. Enomoto1989a: 44 note 2 comments: "the Tibetan and Chinese versions reflect the reading like *sarvakāmajahasya". In the case of the Sanskrit version, the beginning of the stanza has not been preserved, reading only ... hasya bhikkhuno, dhunvānasya puraskṛtaṃ rajaḥ, Bernhard 1965: 432 (32.3). This gives the whole stanza a different thrust. Once the point is leaving behind sensuality, the reference to dust would also fall into place, as other discourses associate the image of dust with sensuality and the life of a householder.[n.53] Dust as representing the household life is part of the standard pericope for going forth, which qualifies living at home as a path of dust, rajapatha/塵勞 之處; cf., e.g., MN 27 at MN I 179,12 and its parallel MĀ 146 at T I 657a5. In Sn 974 the image of dust represents passion, rāga, in regard to the objects of the five senses. Notably Ud-a 166,34, in its gloss on purekataṃ rajaṃ in Ud 21,11, speaks of rāgaraja, the "dust of passion". Though the qualification of such dust as something "formerly done" remains puzzling, a reference to sensuality would also concord with the figurative usage of dhunati in other discourses, since to shake off the dust of kāma would be a mode of shaking off detrimental mental qualities.[n.54] Enomoto1989a: 44 note 2, however, seems to take the reading kamma/karma to be the original one, as he comments that "karman is replaced by kāma". In fact, the qualification purekata would suit karma better than kāma (ibid. note 3).

Regarding the explanation given in the prose section of this Udāna discourse, the corresponding prose section in the Chinese parallel has no reference to experiencing pain that results from former deeds; in fact it does not even mention a monk seated in meditation.[n.55] T 212 at T IV 765a8. In general, prose sections in the Udāna tend to be later than the verses,[n.56] Cf. above note 66 page 392. making it possible that the description of a monk seated in meditation and enduring pain came into being at a time when the stanza already read sabbakammajaha. In sum, just as in the case of the Karajakāya-sutta, in the present instance, too, the attribution of Jain thought to the Buddha could be due to an error in textual transmission.[n.57] This would support the suggestion by Bronkhorst 1998a: 12 that the presence of such passages "in the Buddhist canon can be most easily explained on the assumption that they are not part of earliest Buddhism" (emphasis removed); for a reply to Bronkhorst 1998a cf. Enomoto 2002/2003: 243f.

Another case of affinity with Jain thought can be found in a discourse in the Aṅguttara-nikāya, which presents the training in morality, concentration and wisdom as three modes of "annihilation", nijjarā, as a result of which one does not perform new deeds and gets rid of former deeds by experiencing them.[n.58] AN 3.74 at AN I 221,23: so navañ ca kammaṃ na karoti purāṇañ ca kammaṃ phussa phussa vyantikaroti (Be, Ce and Se read vyantīkaroti/byantīkaroti), with a counterpart in SĀ 563 at T II 147c27: 業更不造, 宿業漸已斷. Enomoto (1989a: 52) comments that "the word nijjarā (nirjarā) is a technical term of Jain practice and indicates the annihilation of karman". In this case, however, the use of Jain terminology seems to be intentional. The discourse begins with a Jain follower presenting his beliefs in a somewhat challenging manner, hence the use of Jain expressions in the reply serves a purpose, since in this way a teaching could be given in terms understandable and acceptable to the visitor.[n.59] The expressions used in AN 3.74 recur in AN 4.195 at AN II 197,32, with a counterpart in MĀ 12 at T I 434b23; cf. also Bronkhorst 1993/2000: 29 note 8. In this case, too, the context is an encounter with a Jain, which would explain the choice of terminology. According to Ruegg 1989: 143, cases like AN 3.74 and AN 4.195 are "evidently to be explained by the fact that ... [the] auditor was a Nirgrantha and that the teaching was thus intended as an introductory salvific device"; cf. also Gombrich 1994: 1095.

Yet another instance relevant to the present context could be the recommendation to constrain the mind forcefully, given in the Vitakkasanthāna-sutta and its Madhyama-āgama parallel as the last of five methods for dealing with unwholesome thoughts.[n.60] MN 20 at MN I 121,1 and MĀ 101 at T I 588c17. At first sight, this recommendation seems to contrast with other discourses that include the same practice among exercises that had not been able to lead the bodhisattva to awakening.[n.61] [63] E.g., MN 36 at MN I 242,26; cf. also the more briefly formulated counterparts in the Mahāvastu, Senart 1890: 124,3, and the Saṅghabhedavastu, Gnoli 1977: 100,3. Bronkhorst (1993/2000: xll) takes this instruction to be a case where "Jaina practices had come to be accepted by at least some Buddhists".[n.62] Cf. also King 1980/1992: 10 and Bronkhorst 1999: 86.

Closer inspection of the Vitakkasanthāna-sutta and its parallel reveals, however, that this exercise is not presented as something that on its own results in awakening,[n.63] This can be seen more clearly in MĀ 101, which does not have a counterpart to the reference in MN 20 at MN I 122,4 to cutting off craving, etc.; for a discussion cf. Anālayo 2005: 8–9. but rather as a last resort in case all other attempts to deal with unwholesome thoughts have failed. Even though forcefully restraining the mind is not a method that will result in awakening, it does serve to stop unwholesome thoughts and thereby prevent their spilling over into unwholesome actions.[n.64] Gombrich 1994: 1080 comments that it is not "at all strange that a technique which, used by itself ... turned out not to lead to Enlightenment, could be recommended ... for overcoming a particular difficulty". Thus the instructions given in the Vitakkasanthāna-sutta do not conflict with other Pāli passages that consider forceful restraint of the mind to be on its own incapable of leading to awakening.

In sum, besides the case of the Karajakāya- and Sañcetanika-suttas, the above discussed Udāna discourse seems to be another case where a Pāli discourse takes up a position that fits Jain thought better than the teachings of early Buddhism. In both cases, the formulation found in the Pāli version is not supported by the parallel versions, so that both cases may well be the outcome of errors that occurred at some point during their transmission. These two cases thus reveal the potential of comparative studies based on parallel versions transmitted by different schools for properly assessing the doctrinal legacy of the early Buddhist discourses.