Introduction

i.1The Exposition of Karma (Karmavibhaṅga) opens in Prince Jeta’s grove, where the Buddha announces to the brahmin youth Śuka Taudeyaputra that he will deliver this exposition on karma. The ensuing teaching provides a detailed analysis of the complex workings of karma. This is enlivened by many examples drawn from the rich store of Buddhist narrative literature, especially the Buddha’s past-life stories. It begins with a section in which the Buddha poses and answers a series of questions about how conspicuous differences in life circumstances such as longevity, happiness, illness, and appearance have been determined by past deeds. A second section follows, in which questions relating to specific causes for rebirth in various worlds are answered. A third section contains a series of miscellaneous questions and answers that examine the specific outcomes of deeds when certain factors are either present or absent in their performance. Two final sections focus more broadly on virtuous and nonvirtuous deeds and their respective positive and negative consequences. In each of these contexts, the relationship between actions and their results is illustrated by examples and morality tales from Buddhist narrative literature.

Karma

i.2The central theme of the Karmavibhaṅga is the concept of cause and effect, the complex system of positive and negative results that, in Indian religious thought, are attributed to karma (“action”) itself. In the Buddhist context, the term karma designates both morally good (kuśala) and bad (akuśala) actions of body, speech, and mind. Once committed, all such deeds “ripen” (vipāka) into their corresponding pleasant and unpleasant (or neutral) results, called “karmic fruition” (karma-phala). From this standpoint, the entire universe and everything in it is the result of individuals’ actions. According to the Karmavibhaṅga, certain unpleasant features of one’s environment are the direct outcome of the ten nonvirtuous courses of action. The botanical or agricultural metaphor employed in the Buddhist description of the karmic process of the individual is perhaps noteworthy: through the ripening of karma one reaps or harvests the fruits of one’s actions. Although the historical Buddha was not the first teacher in ancient India to teach the concept of karmic cause and effect, it has been argued that he advanced and redefined the existing notions of karmically relevant actions as consisting primarily in mental intention (Skt. cetanā; Tib. sems pa). This is summarized in the frequently cited passage from the Aṅguttara Nikāya of the Pāli canon: “By action I mean intention, monks. Having formed a (moral) intention, one carries out an action with body, speech, or mind.”

The textual references in the Karmavibhaṅga

i.3The Karmavibhaṅga is rich in references to sūtras and citations from Buddhist literature. Its longest illustrative story is a version of the popular narrative from the Maitrakanya­kāvadāna (Divyāvadāna no. 38) of the voyage of the sea merchant’s son Maitrakanyaka (called Maitrāyajña in the Karma­vibhaṅga), who undertakes a sea voyage to make his fortune, disregarding his mother’s pleas for him to remain on shore and instead physically mistreating her. Because of his disobedience and abuse, he is shipwrecked on the shores of a foreign country and ends up suffering the torments of his personal hell. Another popular story is that of the sthavira Mahāmaudgalyāyana, who is refused alms by a family and subsequently reveals to a stranger their karmic relationship. Both stories are widely known among Tibetan Buddhists from orally transmitted anecdotes of Tibetan lamas. However, many of the other stories and text titles referenced in the Karmavibhaṅga are either completely unknown to us or differ from their better-known versions and other extant texts that bear identical or similar titles.

The Karmavibhaṅga’s Significance and Geographical Distribution in the Buddhist World

i.4Lokesh Chandra, writing about the Javanese Buddhist monument the Borobudur, notes the wide-ranging influence of the Karmavibhaṅga: “It was a popular text from the island of Java to the sands of Central Asia and as far as the sprawling land of China, that is, wherever the doctrine of Buddha held sway.”

i.5A further measure of the work’s widespread popularity is the diverse range of languages in which we find extant versions or fragments of the work: Sanskrit, Pāli, Khotanese, Kuchean, Sogdian, and Chinese. Indeed, it was translated into Chinese five times over eight centuries. Thus, in a variety of cultural contexts, the work served as an important source for the central Buddhist doctrine that humans are responsible for the consequences of their actions.

i.6The design of the Borobudur on the island of Java in Indonesia is thought to include pictorial representations drawn from the Karmavibhaṅga. According to Lokesh Chandra, the monument is a physical model of the Buddhist path to awakening in terms of the four sambhāras or accumulations of merit (puṇya), wisdom (jñāna), tranquility (śamatha), and special insight (vidarśanā) according to the Lalitavistara, while skillfully integrating and harmonizing other textual traditions. The lowest or most basic level of religious merit (puṇya), which must be accomplished before one can ascend to the higher levels of the path, is represented by Borobudur’s so-called hidden base, which features reliefs depicting stories from the Karmavibhaṅga that illustrate the law of karma.

The title of the text

i.7The exact original title of Toh 338 cannot be established beyond a doubt. Sylvain Lévi, the first to edit and translate the text, referred to it as the Mahākarma­vibhaṅga (MKV). However, the adjective mahā- (“great”) only occurs in the title given to one of the two surviving nearly complete manuscripts (called MS[A] by Kudo Noriyuki), and only in an appendix to the text. The second of the two nearly complete manuscripts (called MS[B] in Kudo’s edition) bears the title Karma­vibhaṅga­sūtraṃ. There are similar variants in the Tibetan translations of the text preserved in the different Kangyur collections.

i.8For the sake of simplicity, we here follow Kudo and use the title Karmavibhaṅga instead of Mahākarma­vibhaṅga or Karma­vibhaṅga­sūtra to refer to the text translated here (Toh 338), with the caveat that different versions of this text with either the same or a different title are extant. The Karmavibhaṅga belongs to a group of texts which has been labeled the Karmavibhaṅga- or Śukasūtra class.

Genre

i.9Although one manuscript (MS[B]) contains the word sūtra in its title, there is insufficient evidence from the extant Sanskrit manuscripts to determine whether the Karmavibhaṅga actually belonged to the scriptural category of sūtra or not. As indicated by the example of the Cakravartisūtra ‍—a text the Karmavibhaṅga quotes four times‍—texts that were designated as sūtras may nevertheless have belonged to the Abhidharma Piṭaka of one of the early Buddhist schools. Indeed, from the point of view of style, the actual “sūtra-portion” of the Sanskrit version as edited by Lévi seemingly ends after presenting a mere list of eighty karmic categories. Subsequently, something more akin to a commentary on those categories is inserted, bracketed by the list and the title (Karma­vibhaṅgasūtraṃ samāptam; given in the colophon of MS[B]) that formally marks the end of the text. But there is no “classical” sūtra ending such as a statement that the assembled audience was delighted and rejoiced in the Blessed One’s words. Indeed, the Karmavibhaṅga’s diction is rather characteristic of a treatise or commentary (Sanskrit śāstra): the different actions and their karmic results are presented in the form of a (hypothetical) dialogue in which replies are given to questions about the expected outcomes of specific types of action. This seamless inclusion of what reads like a commentary as well as the diction of the sūtra, which appears to be more in line with a commentarial treatise, is unusual for the sūtra genre. Equally unusual is the absence of a formulaic, sūtra-typical closure in the Sanskrit version of both MS[A] and MS[B].

i.10The Tibetan tradition on the other hand regarded the Karmavibhaṅga as belonging to the sūtra category (mdo sde), and the Tibetan version possesses the characteristic sūtra frame. All the editions of the Tibetan Kangyur available through the website Resources for Kanjur and Tanjur Studies classify it under the sūtra category. The Degé Kangyur contains the Karmavibhaṅga in a subsection called Collection of Sūtras Belonging to the Hīnayāna (theg dman gyi mdo mang), in the vicinity of such celebrated Buddhist classics as the Udānavarga and the Karmaśataka.

Extant versions of the text

i.11A good deal of excellent scholarly work has been done on the Karmavibhaṅga and its related texts. In what follows we collate and summarize some general information about the extant versions of the Karmavibhaṅga and the existing scholarship.

Sanskrit versions

i.12The Sanskrit text of the Karmavibhaṅga was first edited and published together with a French translation by Sylvain Lévi (1932). Lévi used handwritten copies of the original manuscripts. His edition and translation of the text remains the most comprehensive study, bringing together in one place most of the extant versions and fragments of this important text. Kudo Noriyuki (2004) has published a transliteration of the original manuscripts together with extensive annotations on the quotations of the Karmavibhaṅga.

Pāli versions

i.13As is the case with so many sūtras, we have little concrete information about the origin, the circumstances, or the age of the text of the Karmavibhaṅga. Perhaps one of the oldest canonical versions of a more detailed discussion of the Buddhist formulation of the doctrine of karmic cause and effect‍—if one accepts that parts of the Pāli canon are among the oldest representatives of Indian Buddhism, that is‍—can be found in two texts (Pāli sutta) of the Majjhima Nikāya of the Pāli canon: the Cūḷakammavibhaṅgasutta (MN 135) and the Mahākammavibhaṅgasutta (MN 136). The phraseology and the “cast of characters” of the Cūḷakammavibhaṅgasutta, which contains altogether fourteen karmic categories that partly overlap with those of the Karmavibhaṅga, bear some resemblance to the Karmavibhaṅga. The Pāli version of the brahmin youth’s name, Śuka, is Subha.

Chinese translations

i.14According to Lokesh Chandra, the Karmavibhaṅga was popular in China. Different recensions of the text were translated into Chinese altogether six times‍—some of them probably from versions transmitted via Central Asia, where the text was equally well known. The earliest translation dates to the third century ᴄᴇ. All the different translations bearing various related titles, and possibly representing different recensions of the text, have been collectively called the Śukasūtra-class after Śuka Taudeyaputra, the protagonist of the frame story. The Chinese translations, in the order of their dates of translation, are as follows:

Taishō 78: Doutiao jing 兜調經 (*Taudeyasūtra?), the earliest translation, was prepared under the Western Jin, 265–316 ᴄᴇ. The name of the translator is unknown (Lévi: Cha).

Taishō 26: Yingwu jing 鸚鵡經 (Śukasūtra), the 170th sūtra of the Madhyamāgama, was translated 397–98 ᴄᴇ by Saṅghadeva (Lévi: Chs).

Taishō 79: Yingwu jing 鸚鵡經 (Śukasūtra), translated 435–43 ᴄᴇ by Guṇabhadra (Lévi: Chb).

Taishō 80: Fo wei Shoujia zhangzhe shuo yebao chabie jing 佛爲首迦長者說業報差別經, translated 582 ᴄᴇ by Fazhi 法智 Gautama Dharmaprājña (Lévi: Chg).

Taishō 81: Fenbie shan e baoying jing 分別善惡報應經, translated 984 ᴄᴇ by Tian Xizai (Lévi: Cht).

Taishō 755: Jingyi youpose suowen jing 淨意優婆塞所問經, translated 982–1017 ᴄᴇ by Shihu (Lévi: Chc).

i.15Of these, Taishō 80 is given as the Chinese translation equivalent of the Tibetan text translated here, the las rnam par ’byed pa (Toh 338).

Central Asian versions

i.16We can infer from the large number of surviving fragments of versions of the Karmavibhaṅga­sūtra from the Buddhist centers of the Central Asian oasis towns along the ancient Silk Road that this text and its cognate versions must also have been very popular among Central Asian Buddhists. We know of an old Khotanese version, a Central Asian fragment in Sanskrit, several fragments of a Kuchean version, and a Sogdian version.

Tibetan translations

i.17Apart from the Sanskrit and the Central Asian recensions of the Karmavibhaṅga, three different Tibetan versions are preserved in different Kangyurs. While the Kangyurs of the Tshalpa (tshal pa) line mainly contain the versions of the text as preserved in Toh 338 and Toh 339, the Kangyurs belonging to the Thempangma (them spangs ma) line contain the Toh 339 version and, instead of the Toh 338 version, another version of the text. The mixed-lineage Lhasa Kangyur includes all three.

i.18A text bearing the title las rnam par ’byed pa zhes bya ba (Karma­vibhaṅganāma) in the Tengyur (Toh 3959) is an independent work attributed to the authorship of Atiśa Dīpaṅkaraśrījñāna (982–1054) and has no direct or explicit relation to the Karmavibhaṅga or the Tibetan versions. Apart from the Nepalese Sanskrit commentary contained in Lévi’s 1932 edition of the Karmavibhaṅga, we are aware of only one (combined) commentary on the las rnam par ’byed pa (Toh 338) and the (here so called) las rnam par ’gyur ba’i mdo (Toh 339) by Choné Lama Drakpa Shedrup (co ne bla ma grags pa bshad sgrub, 1675–1748). No canonical commentary on the Karmavibhaṅga is known to us.

i.19None of the other known versions is an exact match of Toh 338. In other words, we do not possess, and do not know whether there ever existed, a complete Indic source text of the las rnam par ’byed pa. The relationships of the different Tibetan versions of the Karmavibhaṅga as well as their relationships to the other extent versions in other languages await further research.

Notes on the English translation

i.20We have based our English translation on the Tibetan text (Toh 338) of the Degé Kangyur as well as the Comparative Edition (dpe bsdur ma) and prioritized the diction and register of the Tibetan translation. However, we have also perused the Sanskrit editions made by Lévi and Kudo in parallel with the Tibetan text and have chosen to translate the corresponding Sanskrit passage instead of the Tibetan in cases where the Tibetan translation was ambiguous or unintelligible. Our preferences are recorded in the notes.