Introduction

In the present appendix I study the Pāli Udāna collection with a view to discerning its gradual development. The topics I will be taking up are:

• a comparison of the prose sections of the Udāna collection with one of its Chinese parallels,
• a survey of the thematic continuity in the Udāna,
• an identification of the presence of concatenation among the 'inspired utterances' collected in the Udāna,
• an examination of the sequence of the Udāna narrations.

The Pāli Udāna collects a series of 'inspired utterances' by the Buddha in eight chapters (vagga) with ten discourses each. Throughout the Pāli Udāna collection, a prose narration provides information about what according to tradition led to the inspired utterance in question.

The Pāli Udāna collection has a Sanskrit counterpart in the Udāna(varga), [n.1] Regarding the title of this collection, Bernhard 1969 suggests that the Sanskrit collection would have had the same title as its Pāli counterpart, i.e. just Udāna, without the additional "varga" qualification that is commonly used when referring to the Sanskrit work; cf. also the discussion in Dhammajoti 1995: 39 and Willemen 1978: xxv. a work that collects over a thousand inspired utterances in thirty-three chapters.[n.2] Edited by Bernhard 1965a, on this edition cf. esp. Schmithausen 1970. For earlier publication of Udāna fragments cf. Pischel 1908, de La Vallée Poussin 1912 (cf. also the study in Lévi 1912), Chakravarti 1930, Pauly 1960: 213– 258, and Pauly 1961: 333–410. Subsequent to Bernhard's edition, Nakatani 1987 appeared, in addition to which several fragments have been published in the SHT series: SHT VI 1274, 1293, 1470, SHT VII 1680, 1715, 1786, 1795, SHT VIII 1993, SHT IX 2007, 2039, 2046, 2074R, 2077, 2078, 2079V, 2116, 2129, 2198, 2328c+d, 2483, 2484, 2489, 2496, 2499, 2500, 2505, 2510, 2528, 2547a+b, 2548, 2559, 2561, 2563, 2564, 2574, 2599, 2612, 2642, 2652, 2655, 2675, 2701, 2791, 2796, 2842, 2849, 2851, 2855, 2859, 2863, 2873, 2963, and 3013, SHT X 4395 a b iR, 4408, 5463+5471, 5483, 5494, 5500, 5504, 5509h, 5516, 5518, 5530, 5541, 5558, 5562, 5564, 5567, 5576, 5591, 5599, 5612, 5613, 5614, 5615, 5632, 5639, 5658, 5714, 5716, 5723, and 5794. Fragments from the Stein and Hoernle collections can be found in Wille 2006a, 2006b, and 2009, Fukita 2009, Hirabayashi 2009, Karashima 2009, Kudo 2009, Melzer 2009, Nagashima 2009, and Ye 2009. Unlike the Pāli version, the Sanskrit collection does not employ prose sections to introduce or comment on its inspired utterances. The same is the case for the Udāna(varga) preserved in Tibetan translation.[n.3] Edited by Beckh 1911 and Zongtse 1990. Prose passages can, however, be found in a commentary on the Udāna collection attributed to Prajñāvarman, preserved in Tibetan, edited by Balk 1984.

In the Chinese canon two translations of an Udāna collection can be found.[n.4] These are the 出曜經, T 212 at T IV 609c20 to 776a13, and the 法集要頌經, T 213 at T IV 777a8 to 799c4, the latter of which has been translated by Willemen 1978. Besides these two Udāna collections, the Chinese canon has also preserved two Dharmapada translations: the 法句經, T 210 at T IV 559a7 to 575b9, a collection entirely in verse, of which the (probably more original) first twenty-six chapters have been translated by Dhammajoti 1995; and the 法句譬喻經, T 211 at T IV 575b19 to 609b20, a selection of Dharmapada stanzas combined with prose narrations, which has been translated by Willemen 1999. Besides these there exists also an apparently 'apocryphal' work, the 佛說法句經, T 2901 at T LXXXV 1432b6 to 1435c4, together with a commentary, the 法句經疏, T 2902 at T LXXXV 1435c10 to 1445a5. One of these also contains only stanzas, whereas the other version (T 212/出曜經) accompanies its inspired utterances with prose sections.[n.5] According to Willemen 1978: xxi, these prose sections are probably "a translation of an Indian text, Dharmatrāta's commentary" (who is held to have compiled the Udāna(varga); cf., e.g., the *Mahāvibhāṣa, T 1545 at T XXVII 1b18. According to a preface to this version, this Udāna collection was rendered into Chinese by Zhú Fóniàn (竺佛念) based on an original read out to him by a Kashmiri monk, the translation being completed in the year 399. [n.6] T IV 609c9; cf. also Willemen 1973: 217. This Chinese version (T 212) has nearly as many inspired utterances as the Sanskrit Udāna(varga), [n.7] Willemen 1978: xxi points out that T 212 sometimes follows a particular inspired utterance with a prose summary of similar stanzas, where the Sanskrit Udāna(varga) instead has a whole set of inspired utterances. This may be one of the factors responsible for the differing count of inspired utterances in T 212 and in the Sanskrit version; cf. also the surveys in Mizuno 1981: 62 or Mukherjee 1939: 750 (my ignorance of Japanese has prevented me from making full use of Mizuno's study or other studies published in Japanese). which together with the prose sections that accompany them cover thirty fascicles in the Taishō edition.

---

Study

The Pāli Udāna Collection in Comparative Perspective

Of the various extant parallel versions,[n.8] Besides versions listed above, several inspired utterances have also been preserved in Tocharian; cf. Sieg and Siegling 1931, Lévi 1933: 41–56, Thomas 1971 and 1979, and Tamai 2009. In addition, some inspired utterances can also be found in Uighur fragments; cf. von Gabain 1954: 23–24 and 38–44. Parts of a commentary on the Udāna, the Udānālaṃkāra attributed to Dharmasoma, are also extant in Tocharian; cf. Lévi 1933: 72–77 and Sieg and Siegling 1933. for the purpose of examining the Pāli Udāna collection this Chinese version (T 212/出 曜經) is of central relevance, as it is the only canonical version that stands a chance of having not only counterparts to the inspired utterances themselves,[n.9] My correlation of Pāli udānas with their counterparts in T 212 is based on the concordances in Bernhard 1965b: 217–263, Mizuno 1981: 394–469, and Willemen 1974. but also parallels to the prose narrations found in the Pāli Udāna collection. In what follows I survey parallelisms and differences between the prose materials in these two versions, following the sequence in which the discourses are located in the Pāli collection.

The first chapter of the Pāli Udāna collection opens with a set of three inspired utterances spoken by the Buddha when, soon after his awakening, he reviewed the law of dependent arising. The Chinese parallel associates its version of these three inspired utterances with the same background narration.[n.10] Ud 1.1 at Ud 1,1 and T 212 at T IV 775c17.

In regard to the remaining inspired utterances found in this first chapter of the Pāli Udāna collection, however, the parallel stanzas that have been identified in the Chinese version are instead accompanied by a word commentary.[n.11] The inspired utterances Ud 1.5, 1.7, 1.8, 1.9, and 1.10 have counterparts in T 212 at T IV 774b17, 775b10, 771a2, 770a10, and 735a17 (in the case of Ud 1.5, the Chinese counterpart only shares the second half of the stanza). T 212 accompanies each of the respective stanzas with a word commentary only. For a comparative study of Ud 1.7 and other parallels cf. Lamotte 1968. This is in fact a general feature, in that often the prose sections in the Chinese version offer merely a word commentary on key expressions found in the inspired utterance in question, without providing a narrative setting for the stanzas. This is similar in kind to the Pāli commentary on the Udāna collection, the Paramatthadīpanī.

Just to provide one example, the report of how the non-Buddhist ascetic Bāhiya Dārucīriya meets the Buddha on the streets of Sāvatthī and receives an enigmatic instruction that causes his instant full awakening (Ud 1.10) is not found at all in the Chinese version,[n.12] Ud 1.10 at Ud 6,25. not even in relation to another inspired utterance.[n.13] A case of distantly related narrations can be found in relation to Ud 1.6 at Ud 4,11, which describes how Mahākassapa goes begging among the poor after having been sick, discarding the offer of some devas who want to supply him with food. T 212 at T IV 657b16 reports that, after receiving gross food due to his preference for begging at the houses of the poor, Mahākassapa becomes sick and is thereupon visited by the Buddha. Although the two narrations share the motif of Mahākassapa being sick and both refer to his practice of begging for alms, they nevertheless differ considerably and do not seem to be parallel accounts of the same event. Besides, the respective inspired utterances are not related to each other. In the case of Ud 1.8 at Ud 5,11, although the narration does not recur at all in T 212, parallel stanzas with a similar narration can be found in SĀ 1072 at T II 278b11 and SĀ2 11 at T II 376b21; cf. also Lamotte 1968: 446.

The second chapter of the Pāli collection introduces its first inspired utterance by describing how the nāga king Mucalinda encircled the body of the recently awakened Buddha and spread its hood over him (Ud 2.1). Mucalinda remained like this for seven days in order to protect the Buddha, who was seated in meditation, from a downpour. [n.14] Ud 2.1 at Ud 10,6; on this tale cf. also Bareau 1963: 101–105 and Bloss 1973: 49–50. The Chinese counterpart describes instead how the recently awakened Buddha approached the palace of a nāga king to deliver the inspired utterance in question, with neither prose nor stanza referring to any encircling of the Buddha's body by the nāga king or to a downpour.[n.15] T 212 at T IV 755b7. In this case parallel versions of an inspired utterance are accompanied by narrations that have only some elements in common, but differ in a number of other respects.

Another inspired utterance in this chapter of the Pāli version has as its audience a group of mourning lay disciples (2.7), one of whom had lost his only son.[n.16] Ud 2.7 at Ud 14,21. A tale of a brahmin who has lost his only son can be found in T 212 at T IV 649c8. Besides being related to a different inspired utterance, however, this tale is rather a parallel to MN 87 at MN II 106,2, MĀ 216 at T I 800c22, T 91 at T I 915a8, and EĀ 13.3 at T II 571b29. According to the Chinese parallel, however, the occasion for this inspired utterance was instead a teaching given by the Buddha to his monk disciples, in which he described a class of devas that pass away from their realm due to excessive merrymaking.[n.17] T 212 at T IV 651c19. This description is in several respects similar to a passage found in the Brahmajāla-sutta and its parallels, which depicts the khiddāpadosikā devā; cf. DN 1 at DN I 19,11, DĀ 21 at T I 90c15, and T 21 at T I 266c9, a discourse quotation in the *Śāriputrābhidharma, T 1548 at T XXVIII 657b12, a Tibetan discourse parallel in Weller 1934: 28,23 (§76), and a discourse quotation in D 4094 ju 146b3 or Q 5595 tu 168b3. This particular type of deva features also elsewhere; cf., e.g., DN 24 at DN III 31,5 and its parallel DĀ 15 at T I 69b28, or DN 20 at DN II 260,15 (for the parallel versions cf. Skilling 1997: 532 and 548 (§23)). In this case the narrations that accompany parallel versions of an inspired utterance differ quite radically from each other.

With the remaining discourses in this chapter of the Udāna, the Chinese text accompanies its version of the respective inspired utterance by a word commentary only.[n.18] This is the case for Ud 2.2, 2.3, 2.4, 2.5, 2.6, 2.8, 2.9, and 2.10 and their counterparts in T 212 at T IV 757a16, 753b1, 758b15, 758a5, 758a16, 652a18, 758a25, and 757c20. In the case of Ud 2.2, the narration recurs in SĀ 413 at T II 110a3, although without being related to any stanza; cf. Lamotte 1968: 446. Noteworthy stories like the tale of a wanderer who tries to get oil for his pregnant girlfriend and then suffers from having drunk too much of it at a free dispensary (Ud 2.6),[n.19] Ud 2.6 at Ud 13,20. A distantly related tale about drinking too much oil occurs in the Mūlasarvāstivāda Vinaya, where a monk passes away after drinking a bowlful of ghee; cf. Dutt 1984b: 118,10. or the account of how the Buddha's blessing helps Suppavāsa to give birth after seven years of pregnancy (Ud 2.8),[n.20] Ud 2.8 at Ud 15,7. do not occur at all in the Chinese text.

Coming to the third chapter, here the second discourse presents the well-known story of Nanda (Ud 3.2), whom the Buddha takes up to heaven in order to show him the heavenly nymphs he will win if he continues to live the life of a monk. The Chinese collection does contain a version of this story, although it relates this to different stanzas.[n.21] Ud 3.2 at Ud 21,16, and T 212 at T IV 739c14, whose stanza is similar to Dhp 344. A version of the Nanda tale in the Ekottarika-āgama, EĀ 18.7 at T II 591c5, relates this tale to still other stanzas (cf. T II 591c1, 592b9, and 592c5), whereas a version in the Mūlasarvāstivāda Vinaya has no stanzas at all; cf. T 1451 at T XXIV 252b1. Most of the other inspired utterances found in the third chapter of the Pāli Udāna collection are accompanied in the Chinese parallel by a word commentary, so that the narrative prose that introduces these in the Pāli version remains without a counterpart.[n.22] A prose parallelism can be found between Ud 3.8 at Ud 30,18 and T 212 at T IV 628c27, which report how monks deliberate that on going begging they will be able to get access to pleasant sensory experiences, whereupon they are rebuked by the Buddha. However, this narration introduces different inspired utterances in the two versions. The description in Ud 3.7 at Ud 29,8 of Sakka disguising himself in order to be able to offer alms to Mahākassapa does not occur in T 212 at all (which only has a counterpart to the respective inspired utterance at T IV 764c13). A similar tale can, however, be found in the Mahīśāsaka Vinaya, T 1421 at T XXII 53b16, where it forms the occasion for the Buddha to lay down a rule permitting the acceptance of food offered by devas.

A discourse in the fourth chapter of the Pāli Udāna collection records an occasion when the Buddha went to live alone (Ud 4.5), his only company being an elephant that similarly had left its herd in order to be undisturbed. The Chinese version accompanies the corresponding inspired utterance with a similar story.[n.23] Ud 4.5 at Ud 41,6 locates this event at Kosambī and explains that the Buddha wished to retire into solitude because he had been too crowded in by disciples and visitors, ākiṇṇo viharati bhikkhūhi bhikkhunīhi upāsakehi upāsikāhi … ākiṇṇo dukkhaṃ na phāsu viharati. T 212 at T IV 762c3 instead reports that the monks had been in sore disputation among themselves, to the extent that they were unwilling to accept the Buddha's intervention, and it was due to this that he decided to leave them and be on his own. Another occurrence of this stanza in T 212 at T IV 698a19 is preceded by a reference to the litigating Kosambī monks (T IV 698a11), thereby confirming that the Buddha's desire for solitude was related to the notorious Kosambī incident, recorded in several discourses and Vinayas; cf., e.g., MN 128 at MN III 152,24, MĀ 72 at T I 532c11, EĀ 24.8 at T II 626b12, the Dharmaguptaka Vinaya, T 1428 at T XXII 879b24, the Mahāsāṅghika Vinaya, T 1425 at T XXII 333c27, the Mahīśāsaka Vinaya, T 1421 at T XXII 158c7, the Mūlasarvāstivāda Vinaya, Dutt 1984b: 176,3, and the Theravāda Vinaya, Vin I 337,2.

According to the Pāli version, another inspired utterance in this chapter was spoken by the Buddha on seeing Sāriputta seated in meditation (Ud 4.7).[n.24] Ud 4.7 at Ud 43,10. The Chinese text instead associates this inspired utterance with the story of Soṇa Koḷivisa. After excessive striving he wants to disrobe, but is then taught the simile of the lute by the Buddha and eventually becomes an arahant.[n.25] T 212 at T IV 638c2, with counterparts in AN 6.55 at AN III 374,10 (= Vin I 182,5), MĀ 123 at T I 612a2, SĀ 254 at T II 62b24, EĀ 23.3 at T II 612a22, in the Dharmaguptaka Vinaya, T 1428 at T XXII 844b7, in the Mahīśāsaka Vinaya, T 1421 at T XXII 146a22, and in the Mūlasarvāstivāda Vinaya, Gnoli 1978: 142,9; cf. also Waldschmidt 1968. In the case of another tale, according to which a yakkha hit Sāriputta on the head when the latter was seated in meditation (Ud 4.4), the Chinese parallel only provides a word commentary on the corresponding inspired utterance itself. The tale of Sāriputta being hit on the head does not occur anywhere else in the Chinese collection.[n.26] Ud 4.4 at Ud 39,18 and T 212 at T IV 763b25. The same tale is, however, found in SĀ 1330 at T II 367b5, SĀ2 329 at T II 485a24, and EĀ 48.6 at T II 793a13; cf. also Lamotte 1968: 446.

The fifth chapter of the Pāli Udāna collection contains several remarkable narrations, such as the story of Queen Mallikā (Ud 5.1), who with frank honesty tells her husband that she holds nothing dearer than herself.[n.27] Ud 5.1 at Ud 47,2, where only the stanza has a parallel in T 212 at T IV 653b6. Another tale involves the leper Suppabuddha (Ud 5.3) who mistakes a congregation that surrounds the Buddha for a food distribution and approaches in the hope of begging for something to eat, but instead of receiving food attains stream-entry during the discourse given by the Buddha.[n.28] Ud 5.3 at Ud 48,17; the stanza parallel occurs in T 212 at T IV 743c29.

Neither story is found in the Chinese version at all. The same is the case for the account of Soṇa Koṭikanna of Avantī (Ud 5.6), a disciple of Mahākaccāna, who after some delay finally goes forth and then visits the Buddha,[n.29] Ud 5.6 at Ud 57,2 (= Vin I 194,18), with its stanza parallel in T 212 at T IV 742b23; for a study of the parallel versions cf. Lévi 1915. and for the announcement made by Devadatta in the presence of Ānanda that he will carry out the uposatha observance on his own (Ud 5.8).[n.30] Ud 5.8 at Ud 60,15, where the stanza has a counterpart in T 212 at T IV 744c11. An account of Devadatta's activities can be found in T 212 at T IV 695c18 to 696c11, although without any reference to an announcement made by Devadatta to Ānanda regarding his schismatic intentions. For comparative studies of the Devadatta narrative cf. Bareau 1991 and Mukherjee 1966.

The first inspired utterance in the sixth chapter of the Pāli collection revolves around the Buddha's rejection of the volitional formation related to existence (Ud 6.1).[n.31] Ud 6.1 at Ud 64,29: bhavasaṅkhāram avassajji (Be: avassaji) muni; cf. also above p. 410 note 66. The Pāli prose introduction to this inspired utterance indicates that the background to the stanza in question was that the Buddha had given up of his life force after Ānanda had failed to ask him to live longer. According to the Chinese version's word commentary, however, the implication of this line is that at the time of his awakening the Buddha had rejected any volitional formation in regard to the three realms of existence.[n.32] T 212 at T IV 735c6: 至真等正覺 … 捨其三有不造其行.

The prose introduction to the next inspired utterance in the Pāli version (Ud 6.2) relates how Pasenadi worships spies disguised as ascetics and then asks the Buddha about the degree of spiritual development of these ascetics. The narration that accompanies the Chinese version of the inspired utterance in question is instead about a newly ordained monk who finds it difficult to handle the coarse food he gets on his alms round.[n.33] Ud 6.2 at Ud 65,2 and T 212 at T IV 690b23. T 212 at T IV 691b13 continues by linking this event to several other stories, among them also the attempt by the Buddha's foster mother to offer a robe to the Buddha, paralleling the Dakkhiṇāvibhaṅga-sutta and its parallels, cf. MN 142 at MN III 253,10, SHT III 979, Waldschmidt 1971: 241, MĀ 180 at T I 721c28, T 84 at T I 903c4, D 4094 ju 254a3 or Q 5595 tu 289b3, T 202 at T IV 434a6, T 203 at T IV 470a15, T 1421 at T XXII 185b21, a Tocharian version, Xianlin 1998: 168, and an Uighur version, Geng and Klimkeit 1988: 192; for further parallels cf. Jaini 1988/2001: 461 and Lamotte 1958/ 1988: 703 and for a comparative study Anālayo 2011b: 811f. Another inspired utterance in the Chinese version is, however, accompanied by a similar narration.[n.34] T 212 at T IV 748c5, with the difference that here Pasenadi only finds out that the ascetics are fake after he has asked the Buddha about them, whereas in Ud 6.2 he knows this all along. The story of how king Pasenadi asks the Buddha's opinion about some fake ascetics also occurs in a discourse in the Saṃyutta-nikāya and its parallels in the two Saṃyukta-āgamas, which relate it to an inspired utterance that differs from the one associated with this event in the Pāli Udāna collection.[n.35] SN 3.11 at SN I 79,16, SĀ 1148 at T II 306a13, and SĀ2 71 at T II 399b18; cf. also Lamotte 1968: 446. In this case even a Pāli parallel found in the Saṃyutta-nikāya disagrees with the Pāli Udāna collection on the relation between a particular event and the stanza(s) spoken by the Buddha on that occasion.

The prose narration introducing the ninth inspired utterance in this chapter of the Pāli Udāna collection depicts the Buddha seated at night, watching how numbers of moths pass away due to being attracted and then scorched by the lamps that are kept burning (Ud 6.9). In the Chinese version, the depiction of moths being scorched recurs, but here this is a simile used in the word commentary to illustrate the implication of the stanza.[n.36] Ud 6.9 at Ud 72,6 and T 212 at T IV 748a26: 如蛾投火, 不顧後慮. Whereas in the Chinese version the image of moths that fly into a flame is an innocent metaphor, in the Pāli account the Buddha keeps watching the passing away of these moths without intervening by putting out the lamps.[n.37] [38] As already mentioned above p. 404f, this presentation is surprising. It stands in contrast to the importance accorded to avoiding harming other living beings among ancient Indian wanderers and recluses, something also reflected in Vinaya rules that, for example, prohibit using water that contains living beings. The respective pācittiya/pātayantika rules, which proscribe pouring out or using water that contains living beings, can be found in the Dharmaguptaka Vinaya, T 1428 at T XXII 646c18 and 677c2, in the Kāśyapīya Vinaya, T 1460 at T XXIV 662b19 and c21, in the Mahīśāsaka Vinaya, T 1421 at T XXII 45a6 (as already pointed out by Pachow 1955: 143, the Mahīśāsaka Vinaya has this regulation as a single rule), in the Mahāsāṅghika Vinaya, T 1425 at T XXII 345a14 and 372c23, in the Mūlasarvāstivāda Vinaya, T 1442 at T XXIII 789b21 and 828c5, in the Sarvāstivāda Vinaya, T 1435 at T XXIII 79c15 and 97b18, and in the Theravāda Vinaya at Vin IV 49,3 and Vin IV 125,20. A description of how this observance was carried out in seventh-century India can be found in Yìjìng's (義淨) travel records, T 2125 at T LIV 208a13, translated in Takakusu 1966: 30. In view of the concern with not harming even minute living beings, reflected in these regulations, one would expect a Buddha to have more compassion for the insects that keep flying into the flame, instead of just coming out with an inspired utterance on seeing their predicament.

Nine of the ten counterparts to inspired utterances from the seventh chapter of the Udāna collection are accompanied in the Chinese version only by a word commentary,[n.38] [39] This is the case for Ud 7.1, 7.2, 7.3, 7.4, 7.6, 7.7, 7.8, 7.9, and 7.10 and their counterparts in T 212 at T IV 739a21, 734a8, 739a7, 632c28, 752a28, 752b7, 699a9, 707c18, and 737a14. In the case of Ud 7.7 there is nevertheless a parallelism, as the Pāli version's prose at Ud 77,12 describes the Buddha seated in meditation, something also recorded in the Chinese version, T 212 at T IV 752b12: 如來坐禪寂然入定. so that the great majority of the narrations delivered in the Pāli account are without parallel. In the case of the fifth discourse in the Pāli collection (Ud 7.5), which reports how the Buddha eulogizes the monk Lakuṇṭabhaddiya, the parallel stanza in the Chinese version does have a narration. This involves the tale of a king, however, and thus differs substantially from the Pāli version.[n.39] [40] Ud 7.5 at Ud 76,10 and T 212 at T IV 740b4. The stanza found in Ud 7.5 is quoted in SN 41.5 at SN IV 291,19 and its parallel SĀ 566 at T II 149b8.

A prose passage nearly identical to the Pāli Udāna version's description of how the Buddha eulogized the monk Lakuṇṭabhaddiya can be found in the Saṃyutta-nikāya. However, the Saṃyutta-nikāya discourse reports that the Buddha spoke different stanzas on this occasion.[n.40] [41] SN 21.6 at SN II 279,12; cf. also SĀ 1063 at T II 276b11 and SĀ2 2 at T II 374b4. Here again a Pāli parallel found in the Saṃyutta-nikāya disagrees with the Pāli Udāna collection on the relationship between a particular narration and the stanza(s) the Buddha spoke in relation to it.

Several remarkable narrations found in the final chapter of the Pāli Udāna collection do not recur at all in the Chinese version. One such case is the report of how Dabba Mallaputta passed away by rising into the air and performing a self-cremation (Ud 8.9), although this tale occurs in the two Saṃyukta-āgama collections.[n.41] [42] Ud 8.9 at Ud 92,25 and again Ud 8.10 at Ud 93,18, where the respective stanzas have counterparts in T 212 at T IV 734b11 and T IV 757c1; the story of Dabba's spectacular passing away is then found in SĀ 1076 at T II 280b25 and SĀ2 15 at T II 378a28; cf. in more detail above p. 389ff. The story of how the Buddha's attendant Nāgasamāla disagreed with the Buddha on which road they should take, left the Buddha, and thereupon met with misfortune (Ud 8.7) is not found in the Chinese Udāna collection. Similarly, the tale of Visākhā mourning the death of a grandchild (Ud 8.8) is without a counterpart in the Chinese parallel.[n.42] [43] Ud 8.7 at Ud 90,16 and Ud 8.8 at Ud 91,13, where only the inspired utterances have a counterpart in T 212 at T IV 742b5 and T IV 650c1. A narration that has some elements in common with Ud 8.7 can be found in T 212 at T IV 730a21. The Buddha is on a road in the company of a monk and tells the latter that they should take a by-road to avoid Devadatta. The monk then enquires if the Buddha is afraid, which leads to an exposition by the Buddha on his aloofness from any fear. Besides differing substantially from the narration given in Ud 8.7, this tale is also related to a different inspired utterance.

As a narrative background to the fifth inspired utterance of this chapter, the Pāli collection presents the famous account of the Buddha's last meal with Cunda. The parallel stanza in the Chinese version is accompanied by a narration that describes how Mahāmoggallāna takes his younger brother up to heaven in order to convince him of the need to be generous.[n.43] [44] Ud 8.5 at Ud 81,11, where the respective inspired utterance at Ud 88,21 has a counterpart in T 212 at T IV 741c13. On the Buddha's last meal cf. above p. 411 note 68.

In the case of the sixth discourse in this chapter, however, the Pāli collection and its Chinese counterpart agree that the inspired utterance in question was spoken on an occasion when the Buddha had been invited to a meal by two brahmins who were fortifying Pāṭaligāma.[n.44] [45] Ud 8.6 at Ud 85,24 and T 212 at T IV 707c4. A noteworthy difference between the two versions is that the Chinese account does not report any prediction made by the Buddha regarding the future of Pāṭaligāma. According to Ud 8.6 at Ud 88,27 the Buddha predicted: pāṭaliputtassa … tayo antarāyā bhavissanti: aggito vā udakato vā mithubhedato (notably this prediction changes from Pāṭaligāma, used earlier and also employed in the title to this chapter of the Udāna, to Pāṭaliputta). This prediction recurs also in DN 16 at DN II 88,1 and in Vin I 229,11, which lists the three dangers as aggito vā udakato vā abbhantarato vā mithubhedā (cf. also Kv 314,6, which quotes this passage in support of the possibility of predicting the future); with a similar reading found in fragment S 360 162 V4, Waldschmidt 1950:12: agnito-py-udakato- (py-a)[bhya]nta[ra]to-pi [m]i[th]obh[edāt]. The three causes recur in the discourse parallels to DN 16, DĀ 2 at T I 12c10: 一者大水, 二者大火, 三者中人 與外人謀, T 5 at T I 162c17: 一者大火, 二者大水, 三者中人與外人謀, and T 6 at T I 177c28: 一者大火, 二者大水, 三者中人與外人謀. Whereas the Tibetan version of the Mūlasarvāstivāda Vinaya similarly lists three causes, Waldschmidt 1951: 149: me dang chu dang nang ʼkhrug gis so, the Chinese version, T 1448 at T XXIV 22b12, mentions only two causes, water and fire, 水火. In relation to DN 16, Pande 1957: 102 comments that this prophecy "could hardly have been added much before the Village of Pāṭaligāma actually attained such eminence as to become the capital of Magadha".

In sum, of the narrations that in the whole of the Pāli Udāna collection provide the background to a particular inspired utterance, only three cases are associated with the parallel stanzas in the Chinese counterpart. These are the description of the Buddha being seated in meditation soon after his awakening and reviewing dependent arising (Ud 1.1–3); the report of how the Buddha goes to live by himself, his only company being an elephant (Ud 4.5); and the Buddha being invited to a meal by two brahmins who are fortifying Pāṭaligāma (Ud 8.6).

Besides this lack of correspondence in relation to parallel stanzas, in the great majority of cases narrations found in the Pāli version do not recur in the Chinese version at all, not even in relation to another inspired utterance. In this way a comparative study of the Pāli Udāna collection in the light of the only extant canonical parallel that also has prose narration brings to light that, although most of the inspired utterances in the Pāli version have a counterpart in the Chinese version, only a rather small percentage of prose narrations are shared by the two texts.

This indicates that the relationship between an inspired utterance and its prose introduction in the Pāli Udāna collection in the majority of cases is the result of textual developments specific to the Pāli version. Such a conclusion concords with the position taken by a range of scholars, mostly based on examining the Pāli Udāna collection on its own, in that the inspired utterances in this collection seem to present an earlier textual layer, to which the majority of prose narrations would have been added only later.[n.45] [46] Cf., e.g., Seidenstücker 1920: xvi, Winternitz 1920/1968: 67, Woodward 1935: v, Pande 1957: 72, Lamotte 1968: 465, Nakamura 1980/1999: 43, Norman 1983: 61, Abeynayake 1984: 66, Ireland 1990: 7, and von Hinüber 1996/ 1997: 46.

Thematic Continuity in the Pāli Udāna Collection

Further inspection of the Pāli Udāna collection shows that several of the inspired utterances within a chapter are related to one another by the sharing of a common thematic term.[n.46] [47] This characteristic was first noticed by Seidenstücker 1913: 93. The degree of recurrence of such a thematic term can be seen from the following survey, which lists each occurrence for the eight chapters of the Pāli collection.

In the first chapter, the term "brahmin" is present in each inspired utterance and thus forms the uniting theme for this chapter:

1.1b brāhmaṇassa
1.2b brāhmaṇassa
1.3b brāhmaṇassa
1.4a brahmaṇo
1.5d brāhmaṇā
1.6d brāhmaṇan
1.7b brāhmaṇo
1.8d brāhmaṇan
1.9d brāhmaṇo
1.10h brāhmaṇo

In the second chapter, the common theme is "happiness", which occurs in nearly all of the inspired utterances:

2.1a sukho & 2.1c sukhaṃ & 2.1e sukhā & 2.1h sukhan
2.2a -sukhaṃ & 2.2b sukhaṃ & 2.2c -sukhass'
2.3a sukha-[n.47] [48] In this particular case, I adjusted the count of stanza parts in accordance with the Asian editions, as the first part of this inspired utterance (sukhakāmāni bhūtāni) is missing in Ee, which thus does not have the first instance of sukha-, as a result of which the other instances occur at 2.3b, 2.3c, 2.3d, 2.3f, and 2.3g respectively. Here and elsewhere, some obvious misprints in Ee have not been taken into account as variant readings. & 2.3c sukham & 2.3d sukhaṃ & 2.3e sukha- &

2.3g sukham & 2.3h sukhan 2.4a sukha-
2.5a sukhaṃ
2.6a sukhino
2.8c sukhassa
2.9b sukhaṃ
2.10c sukhiṃ

In the third chapter, the thematic term found in each inspired utterance is the "monk":

3.1a bhikkhuno
3.2d bhikkhū
3.3d bhikkhū
3.4c bhikkhu
3.5c bhikkhu
3.6d bhikkhū
3.7a bhikkhuno
3.8a bhikkhuno
3.9d bhikkhū
3.10g bhikkhuno

In the fourth chapter, each inspired utterance mentions the "mind":

4.1d -citto
4.2e -cittassa
4.3c cittaṃ
4.4a cittaṃ & 4.4e cittaṃ
4.5c cittaṃ cittena
4.6e -citte
4.7a -cetaso
4.8d -citto
4.9f -cittassa
4.10a -cittassa

In the fifth chapter, several stanzas refer to "evil":

5.3d pāpāni
5.4c pāpakaṃ[n.48] [49] I have adjusted the count of stanza parts, as the first part of this inspired utterance (sace bhāyatha dukkhassa) is missing in Ee, which has pāpakaṃ at 5.4b and d. & 5.4e pāpakaṃ
5.6c pāpe & 5.6d pāpe
5.8b pāpena &5.8c pāpaṃ pāpena & 5.8d pāpam

In the sixth chapter, several stanzas mention "vision", "views" or "seeing":

6.4d -dassino
6.6c diṭṭhīsu[n.49] [50] Ce reads diṭṭhisu.
6.8 diṭṭhī[n.50] [51] This inspired utterance is not in verse, diṭṭhī is found at Ud 72,1; Be and Ce read diṭṭhiṃ.
6.9d diṭṭhe
6.10 -diṭṭhī[n.51] [52] This inspired utterance is not in verse, -diṭṭhī is found at Ud 73,13; Ce reads diṭṭhi.

In the seventh chapter, manifestations of water or things related to water occur in several stanzas. However, the terms used vary and among them there is only one term that appears more than once:

7.1c oghaṃ
7.3d oghaṃ

In the eighth chapter most stanzas refer in some way to the final goal of liberation, or to those who are liberated, although the terms employed show a range of variations and only one expression occurs more than once:

8.1 anto dukkhassā ti
8.4 anto dukkhassā ti [n.52] [53] Neither instance is in verse; anto dukkhassā is found in the last line.

As the above survey shows, in the case of the first four chapters the usage of a particular term – brāhmaṇa, sukha, bhikkhu, or citta – serves as a thematic element present in nearly all inspired utterances within the respective chapter. In the later four chapters, however, a thematic element becomes considerably less prominent. Whereas in the fifth and sixth chapters a particular term still occurs in several instances – pāpa and diṭṭhi/dassana – the presence of such a thematic term is considerably less prominent than in the earlier chapters. In the case of the final two chapters, such continuity becomes even less discernible.

The chapter headings of the Pāli Udāna collection do not reflect these themes, but are usually taken from the first discourse or a significant discourse in the chapter.[n.53] [54] An exception to this pattern could be the Cūla-vagga (7), where the chapter title might reflect the relative shortness of the chapter, although alternatively it might also be a referent to the protagonist of its first two discourses, a monk of particularly small size. This is particularly remarkable in the case of the first four chapters, where the presence of a thematic term is quite prominent. Nevertheless, the titles of these four chapters are Bodhi-vagga, Mucalinda-vagga, Nanda-vagga, and Meghiya-vagga, entitled after the Buddha's awakening, the nāga king, and the two monk protagonists that feature in the first discourse of the respective chapters. Most of the inspired utterances in these four chapters of the Pāli collection find their counterparts in the Sanskrit Udāna(varga) in the Brāhmaṇa-varga (33 chapter), the Sukha-varga (30), the Bhikṣu-varga (32) and the Citta-varga (31), respectively.[n.54] [55] The Brāhmaṇa-varga has counterparts to the Pāli udānas 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9. The Sukha-varga has counterparts to 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.9, 2.10. The Bhikṣu-varga has counterparts to 3.1, 3.2, 3.3, 3.4, 3.7, 3.8, 3.9, and 3.10. The Citta-varga has counterparts to 4.1, 4.2, 4.3, 4.4, 4.5, and 4.6; cf. also the survey in Mizuno 1981: 8–10.

In this way, whereas the headings of these four chapters in the Pāli Udāna collection do not reflect awareness of the theme that underlies their inspired utterances, in the Sanskrit collection the theme that forms the uniting theme of the inspired utterances collected in a chapter finds explicit expression in that chapter's title.

The reason for this could well be that by the time the Pāli Udāna collection reached its final state, due to the presence of the prose narrations the themes of the inspired utterances themselves were not sufficiently prominent and therefore had no impact on the title given to the chapters of the collection. In the case of the Sanskrit version, however, due to the absence of prose material the theme that unites the inspired utterances contained in a chapter would have naturally been a characteristic evident enough to influence the naming of that chapter.

Concatenation in the Pāli Udāna Collection

Besides being related to each other by sharing a thematic term, several inspired utterances within a chapter are also related to each other through 'concatenation'. Concatenation involves the repetition of a word (or a closely similar word) from a preceding textual line or paragraph in order to facilitate recall in proper sequence.[n.55] [56] I already drew attention to concatenation in the Udāna in Anālayo 2008c; on concatenation in pātimokkha rules cf. von Hinüber 1996/1997: 12. An example for the principle governing such concatenation is the fifth discourse of the first chapter (Ud 1.5), whose inspired utterance reads:

1.5a bāhitvā pāpake dhamme,
1.5b ye caranti sadā satā,
1.5c khīṇasaṃyojanā buddhā,
1.5d te ve lokasmiṃ brāhmaṇā ti.[n.56] [57] Ud 4,8; Be and Se read lokasmi.

The expression dhamme in the reference to having discarded evil states in the first line (Ud 1.5a) relates to the inspired utterance of the preceding discourse, whose third line reads dhammena so brahmavādaṃ vadeyya (Ud 1.4c). The present inspired utterance's reference to destroying the fetters, khīṇasaṃyojanā (Ud 1.5c) shares the expression khīṇa with the inspired utterance of the next discourse, which reads khīṇāsavaṃ vantadosaṃ (Ud 1.6c).

The reference to a brāhmaṇa in the last line (Ud 1.5d) takes up the word that forms the uniting theme for the entire chapter.

Thus, besides the recurrence of the thematic term brāhmaṇa, which is responsible for the present inspired utterance being allocated to the first chapter of the Udāna, the location of this inspired utterance as fifth in this chapter involves the principle of concatenation, as it shares a term with the preceding inspired utterance, dhamme, and another term with the subsequent inspired utterance, khīṇa-. The following survey shows the occurrence of such concatenation between inspired utterances found in adjacent discourses in the Udāna collection.[n.57] [58] My survey only takes into account cases where concatenation can be found between inspired utterances in discourses that follow each other, thereby not covering cases where similar phrases occur at a distance, such as e.g. 3.6b amamo nirāso → 3.9c amamo nirāso.

Concatenation in the first chapter of the Udāna:

1.1a yadā have pātubhavanti dhammā → 1.2a yadā have pātubhavanti dhammā
1.1b ātāpino jhāyato → 1.2b ātāpino jhāyato
1.1c athassa kaṅkhā vapayanti sabbā → 1.2c athassa kaṅkhā vapayanti sabbā
1.2a yadā have pātubhavanti dhammā → 1.3a yadā have pātubhavanti dhammā
1.2b ātāpino jhāyato → 1.3b ātāpino jhāyato
1.4d dhammena → 1.5a dhamme
1.5c khīṇa- → 1.6c khīṇ-
1.6a anañña → 1.7a sakesu
1.9a udakena → 1.10a āpo

Concatenation in the second chapter:

2.1f kāmānaṃ → 2.2a kāma-
2.2a kāma- → 2.3a & 2.3e -kāmāni
2.5c sakiñcanaṃ passa vihaññamānaṃ → 2.6c sakiñcanaṃ passa vihaññamānaṃ
2.5d jano janamhi paṭibandha- → 2.6d jano janamhi paṭibandha-[n.58] [59] Be and Ce read in both instances jano janasmiṃ, a reading found only in the second instance in Se; Ce moreover reads paṭibaddha-.
2.7a piyarūpa- & 2.7f piyarūpaṃ → 2.8b piyarūpena

Concatenation in the third chapter:

3.2b kāmakaṇṭako → 3.3a kāmakaṇṭako
3.2d sukhadukkhesu na vedhati sa → 3.3d sukhadukkhesu na vedhati sa[n.59] [60] Be and Se read vedhatī.
3.3c pabbato → 3.4a & 3.4d pabbato
3.4b suppatiṭṭhito → 3.5a upaṭṭhitā[n.60] [61] Se reads supatiṭṭhito.
3.7a piṇḍapātikassa → 3.8a piṇḍapātikassa
3.7b attabharassa anaññaposino → 3.8b attabharassa anaññaposino
3.7c devā pihayanti tādino → 3.8c devā pihayanti tādino

Concatenation in the fourth chapter:

4.2b micchā- → 4.3c micchā-
4.3b verī … verinaṃ → 4.4d kopa- … kuppati
4.9f santa- → 4.10a upasantasanta-
4.9g vikkhīṇo jātisaṃsāro → 4.10c vikkhīṇo jātisaṃsāro

In the fifth chapter no concatenation is apparent. In the sixth chapter concatenation occurs in a single instance, where the fourth and fifth inspired utterance share part of the stanza:

6.4a imesu kira sajjanti → 6.5a imesu kira sajjanti
6.4b eke samaṇabrāhmaṇā → 6.5b eke samaṇabrāhmaṇā

Concatenation in the seventh chapter similarly involves only a single pair of inspired utterances:

7.6c dhīraṃ → 7.7c muniṃ
7.6e devā → 7.7d sadevako

In the eighth chapter no concatenation is apparent.

From the perspective of concatenation the same pattern can be discerned as with the thematic terms, in that the earlier chapters show a greater degree of concatenation than the later chapters. Now concatenation would have the effect of facilitating recall in proper sequence only if the textual pieces it connects follow each other immediately. That is, concatenation of the inspired utterances would work only as long as these are memorized on their own, without the intervening prose sections. Thus the existence of concatenation between inspired utterances strongly suggests that at an earlier point in time these inspired utterances were being memorized on their own. The prose narrations that at present accompany the inspired utterances would have been added only after the sequential structure of these four chapters was already fairly fixed.

Whereas this pattern is quite evident in the case of the first four chapters of the Udāna collection, the sixth and seventh chapters show concatenation only for a single twin of inspired utterances in each case, and in the fifth and eighth chapters concatenation seems to be altogether absent. That is, in the case of the final four chapters of the Udāna the inspired utterances do not seem to have been held together on their own by concatenation, nor, as shown above, are they interrelated as closely by a thematic term as the earlier four chapters.

The Sequence of the Discourses in the Pāli Udāna

The degree to which in the case of the first four chapters the thematic term of the inspired utterance has played a central role in deciding the location of the discourse can be seen right away with the first four discourses in the first chapter (Ud 1.1-4), which cover the events soon after the Buddha's awakening. The same is the case for the first discourse in the second chapter (Ud 2.1), which describes how the nāga king Mucalinda protected the recently awakened Buddha from rain.

Whereas in the Vinaya these four discourses follow each other in direct succession,[n.61] [62] Vin I 1,1 to 3,30. thereby reflecting the temporal sequence of the events they portray, in the Udāna collection the Mucalinda tale occurs at a considerable distance from the other three. The reason for this would be that the inspired utterance related to Mucalinda does not contain the word brahmaṇa and thus does not fit the pattern that dominates the first chapter, where all inspired utterances contain this word. That is, in this case the connection between the inspired utterances had a stronger influence on the arrangement of the collection than their narrative background.

As in the case of the Mucalinda tale, elsewhere among the first four chapters of the Udāna collection similar or closely related prose narrations tend to be separated from each other. The sixth discourse in the first chapter (Ud 1.6) and the seventh discourse in the third chapter (Ud 3.7) describe how Kassapa goes begging. Again, the second discourse in the second chapter (Ud 2.2) and the eighth discourse in the third chapter (Ud 3.8) feature monks who are engaged in chattering about trivialities. Another instance involves the third discourse in the second chapter (Ud 2.3) and the fourth discourse in the fifth chapter (Ud 5.4), which are concerned with children who are harming an animal. In a similar vein, the fourth discourse in the second chapter (Ud 2.4), the eighth discourse in the fourth chapter (Ud 4.8) and the tenth discourse in the sixth chapter (Ud 6.10) share a narration of other recluses who receive fewer gains than the Buddha and his monks. The same pattern manifests with the seventh discourse in the fourth chapter (Ud 4.7) and the tenth discourse in the same chapter (Ud 4.10), both of which depict Sāriputta seated in meditation.

In each of these cases, it would have been natural to group these discourses together to facilitate recall, had the prose been a determining factor at the time of the shaping of this part of the Udāna collection.

This is, in fact, the case for the later four chapters of the Udāna, where similar narrations tend to be grouped together. A nearly identical narration can be found in the fourth (Ud 6.4), fifth (Ud 6.5), and sixth (Ud 6.6) discourses of the sixth chapter. In the seventh chapter the first two discourses (Ud 7.1 and Ud 7.2) report Sāriputta instructing the same monk. The third and fourth discourses in the same chapter (Ud 7.3 and Ud 7.4) depict the attachment to sensuality of the inhabitants of Sāvatthī in the same terms. The prose of the first four discourses in the eighth chapter (Ud 8.1, Ud 8.2, Ud 8.3, and Ud 8.4) is identical, and the last two discourses in the eighth chapter (Ud 8.9 and Ud 8.10) are based on the same event.

In sum, in the earlier four chapters a thematic term is prominent throughout the inspired utterances, concatenation manifests frequently among them, and similar prose narrations tend to occur in different sections. In the later four chapters a thematic term shared by the inspired utterances is less evident, concatenation is rare, and similar prose narrations tend to occur together.

These findings point to a gradual development of the Udāna collection, where in the case of the earlier four chapters at first only the inspired utterances were assembled, whereas in the case of the later four chapters the inspired utterances would already have been embedded in their respective prose narrations at the time of the formation of these chapters.

Conclusion

The foregoing reveals that the Udāna collection would have taken shape during oral transmission in a gradual manner. Comparison with its Chinese parallel (T 212) indicates that the inspired utterances would at first have been assembled without the prose narration that now accompanies them in the Pāli collection.

Whereas the first half of the collection would have come together in its present order when the inspired utterances in question still stood on their own, in the case of the second half of the collection it seems that the prose narrations were already part of the material that was arranged into chapters. Needless to say, such gradual development does not imply that the inspired utterances in the first four chapters are early and those in the other four chapters are late. The formation of the Udāna collection may at a relatively late stage still have incorporated early material that up to that point had not found its final allocation.

The growth of most of the parallel versions would have been different, in that the Udāna(varga) collections now preserved in Sanskrit, etc., appear to be based entirely on assembling ever more inspired utterances, without supplying these with a prose narration or explanation.[n.62] [63] This process thus incorporated stanzas that in the Pāli tradition are allocated to the Dhammapada; cf. also Bernhard 1969: 881, Nakatani 1984: 139, and von Hinüber 1996/1997: 45. Perhaps such a prose commentary was also known in the respective reciter traditions. In fact without any explanation or narrative background, some of the inspired utterances would be difficult to understand and would also lose much of their effect. Yet only in the case of one of the Udāna collections preserved in Chinese (T 212) did such explanations and narrations become part of the text on which they comment.

The case of the Pāli Udāna collection thus appears to be in some respects the reverse of the case of the Aṭṭhakavagga. The Aṭṭhakavagga in the Pāli canon is entirely in verse,[n.63] [64] Sn 766–975; for a study of the Aṭṭhakavagga cf. Jayawickrama 1948: 232– 239 and 1978: 45–57. and as such has become part of a greater collection that is predominantly in verse, the Sutta-nipāta. The Chinese counterpart to the Aṭṭhaka-vagga, however, presents its stanzas embedded in a prose narration similar to the material found in the Pāli commentary on the Sutta-nipāta. At the same time, the Chinese version has not become part of another collection, but instead stands on its own as a single work.[n.64] [65] T 198 at T IV 174b11 to 189c23; for a translation cf. Bapat 1945 and 1950. These two cases, the Udāna and the Aṭṭhakavagga, seem to testify to the same type of processes, although the final products manifest in the Pāli canon in opposite forms.

In sum, then, the process of formation of the Pāli Udāna collection appears to have been gradual, involving different types of interaction between stanza and prose. In this way, the Udāna collection reveals itself as an intriguing product of oral transmission, with all the complexity that this can involve.