IeT 40

English Summary*

Mārgavibhāga — Distinction of Styles. Critical Edition of the First Chapter of Daṇḍin’s Kāvyādarśa and its Tibetan Rendition Sñan ṅag me loṅ, with a German Translation of the Sanskrit Text. By Dragomir Dimitrov. Marburg 2002. (Indica et Tibetica, 40).


With the availability of more than forty modern editions and reeditions of Daṇḍin’s Kāvyādarśa (“Mirror of Poetic Art”) it could be a moot question what contribution to scholarship yet another work on the text can furnish. Daṇḍin’s treatise is an indispensable prerequisite for any study on Indian poetics and it is for this reason an astonishing fact that so far not a single critical edition of the Sanskrit text exists, based on a thorough study of the textual material and presented in accordance with the principles of textual criticism.1 As for the Tibetan rendition of the text, the Sñan ṅag me loṅ, which is available in an even greater number of publications containing mostly photomechanic reprints of old versions with and without commentaries, the situation is not any better. Hence, it is clear that a critical edition of the Kāvyādarśa and the Sñan ṅag me loṅ has long remained a glaring desideratum. Despite the fact that this desideratum was recognized years ago it was to no avail. The absolute necessity to fill the gap gave the main impetus to prepare the present work which may be considered as an attempt to offer a reliable basis for any further research on the text in particular, and on Indian and Tibetan poetics in general.

Already in Tarkavāgīśa’s editio princeps of the Kāvyādarśa, published in 1863, variae lectiones were reported, which indicated the existence of different traditions of its transmission. The same variants were later recorded in Böhtlingk’s reprint of Tarkavāgīśa’s basic text together with a German translation (1890). The different traditions became even more evident when in 1910 Rangacharya presented a new edition of Daṇḍin’s work on poetics together with Taruṇavācaspati’s commentary and the anonymous Hṛdayaṃgamā, based on South Indian material which sometimes manifests significant differences from the basic text of the editio princeps. The need for a new edition was acknowledged by Belvalkar who, for the preparation of his edition and English translation (1924), relied on the evidence of a greater number of manuscripts, including two old palmleaf manuscripts from the Jaina libraries in Patan and Jaisalmer. However, even in the last edition of Śivanārāyaṇa Śāstrī, accompanied by a good, exhaustive commentary in Hindī (published in three volumes, 1988–90), the basic text is more or less arbitrary, i. e., subjectively constituted and consequently, from the standpoint of textual criticism, of no real value.

Almost a century after Csoma de Körös had listed “Snyan-dñags-kyi-me-long. The mirror of sweet language, (Sans. Kávyadershana)” as found in the Narthang Tanjur,2 van Manen drew attention to the fact that there is a vast Tibetan literature on the Kāvyādarśa, not only various commentaries on the work but also different versions of the basic text.3 It was Banerjee who for the first time in 1939 prepared a modern bilingual Sanskrit-Tibetan edition of the Kāvyādarśa based evidently on material that was found by Sarat Chandra Das in Tashilhunpo in 1882.4 As is now obvious, Banerjee’s edition represents a version of the Tibetan translation which was prepared by Si tu Paṇ chen Chos kyi ’byuṅ gnas (1699–1774) as late as half a century after Daṇḍin’s work had been translated by Śoṅ ston Rdo rje rgyal mtshan (mid 13th–early 14th c.) and the Indian paṇḍita Lakṣmīkara (13th c.). From the time of the first complete translation until Si tu Paṇ chen’s revision the Tibetan text has been repeatedly, and sometimes considerably, changed. In this connection van der Kuijp notes that “the numerous text-critical problems in the transmission of the Tibetan text of Daṇḍin render its careful study imperative”5, while Rajapatirana observes that “... a comparison of the Tibetan version [in Banerjee’s edition] with that found in the Peking edition of the Tibetan Tripiṭaka (No. 5789) clearly indicates the need for a new edition of the Tibetan text ...”6 Although the necessity of a critical edition of the Tibetan translation based on a closer study of the abundant Tibetan material has thus been recognised, little has been done to meet this need until now.

The present book aims at partially fulfilling this desideratum in the field of Indo-Tibetan studies. Here a critical edition of the first of the three chapters of the Kāvyādarśa and its Tibetan translation is being presented. A complete German translation of the first chapter, Sanskrit-German-Tibetan and Tibetan-Sanskrit glossaries as well as various appendices, concordances and indices have also been included. This work, which is a slightly corrected and enlarged version of my MA thesis submitted to the Faculty of Foreign Languages at the University of Marburg (Germany) on September 25, 1999, was begun at the inspiration and under the supervision of my teacher Professor Michael Hahn (Marburg). The primary intention was to consult all the Nepalese manuscripts of the Kāvyādarśa which have become easily available, thanks to the admirable work done within the Nepal-German Manuscript Preservation Project (NGMPP), and to compare their evidence with the Tibetan translation in its canonical and non-canonical versions.

After consulting the Nepalese material the investigation took on a most interesting and exciting turn. One of the manuscripts, an undated palmleaf manuscript written in the Newārī script (my MS A), appeared to be quite old (11–13th c.?) and valuable for its textual evidence, while another palmleaf manuscript neatly written in the Old Bengali script (my MSB; written presumably at the end of the 14th c. or in the first half of the 15th c.)7 was found to contain a short hitherto unknown commentary on the Kāvyādarśa presented in the form of marginal notes. Two other incomplete paper manuscripts in the Devanaāgarī script, apparently of a later date (my MSS E and F) partly preserving the text of the first chapter, also shed more light on the different traditions of transmission of the Sanskrit work. Regrettably, the Nepalese manuscript of Ratnaśrījñāna’s old and admirable commentary (early 10th c.), the so-called Ratnaśrīṭīkā, proved to be only partially preserved, with the commentary to the first and part of the second chapter of the Kāvyādarśa completely missing. Thus, while preparing the present book, it was necessary to rely solely on Thakur/Jha’s edition (1957) of the Ratnaśrīṭīkā, which is not always precise and trustworthy. Nonetheless, it was not difficult to establish the close correspondence of MS A with the basic text which Ratnaśrījñāna commented on and which he must have had at his disposal.

With regard to the Tibetan material, it was realized already at the outset that much more work would have to be accomplished than had been planned originally. Apart from the five canonical versions in the editions of the Tibetan canon available nowadays, Peking (Q), Narthang (N), Ganden (G), Derge (D) and Cone (C), three non-canonical bilingual editions had to be examined: one manuscript copy in the Dbu can script (my Z/ZT) preserving a version prepared by Źa lu Lo tsā ba Chos skyoṅ bzaṅ po (1441–1528); one xylograph copy (my S/ST) containing the revision of Si tu Paṇ chen; and finally one more manuscript copy in the Dbu med script (my Y/YT). Besides this, it was necessary to consider the vast Tibetan commentary tradition. For the purposes of the present edition the commentaries by Dpaṅ Lo tsā ba Blo gros brtan pa (1276–1342), Snar thaṅ Lo tsā ba Dge ’dun dpal (ca. 1370–1430), Rin spuṅs pa Ṅag dbaṅ (1542–1625), Bod mkhas pa Mi pham (1618–1685) and Bstan ’dzin chos kyi ñi ma (1730–1779) were constantly consulted. Especially the study of Dpaṅ Lo tsā ba’s commentary, the so called Dpaṅ Ṭīkā, proved to be of considerable importance. The Tibetan scholar relied heavily on Ratnaśrījñāna’s Sanskrit commentary and this helps us now to interpret both texts in a better way and to correct the errors in the available sources. The mere fact that Dpaṅ Lo tsā ba was a direct student of Śoṅ ston and, according to the tradition, made the first revision of his teacher’s translation, suggests how fundamental it is to study his commentary.

Furthermore, attention was paid to all significant modern editions and commentaries on the Kāvyādarśa, from Tarkavāgīśa’s editio princeps to Śivanārāyaṇa Śāstrī’s commentary. Since the modern Tibetan works on the Sñan ṅag me loṅ cannot surpass the evidence of the available older Tibetan sources and are practically of no particular help, they were not taken into consideration. The critical edition of the first chapter of the Kāvyādarśa and the Sñan ṅag me loṅ presented in this volume is preceded by an introduction (1. Introduction). At the beginning an outline pointing out the more important publications of the Sanskrit work and its Tibetan translation is given and the need for a new critical edition of both texts is emphasised (1.1).

Then a few general questions concerning the work as a whole are discussed (1.2). First, the problems connected with the exact wording of the Sanskrit and the Tibetan titles are dealt with (1.2.1). No support is found for the Sanskrit title Kāvyalakṣaṇa which, according to Thakur/Jha, appeared in the colophon after each of the three chapters in the only manuscript of the Ratnaśrīṭīkā they consulted. Unfortunately, the incomplete Nepalese manuscript of the Ratnaśrīṭīkā does not permit one to verify the trustworthiness of the colophon information of Thakur/Jha’s manuscript, because the relevant folios are missing. Since the word kāvyalakṣaṇa is, rather, one of the technical terms in Sanskrit for poetics, and, on the other hand, no further evidence for the alleged title is found, not even in the Tibetan sources which were directly dependent on Ratnaśrījñāna’s commentary, the well-known title Kāvyādarśa was retained. The Tibetan title Sñan (d)ṅag(s) (kyi/gyi) me loṅ, though transmitted with minor differences and usually referred to in short as Sñan ṅag me loṅ, is an exact equivalent of the Sanskrit title. Second, the subject matter of Daṇḍin’s work is briefly presented and the question of the differently transmitted title of the third chapter is touched upon (1.2.2).

The third section of the introduction discusses Daṇḍin’s dates (1.3). There is hardly anything new here, but an attempt has been made to present as clearly as possible and as fully as thought necessary the history of the discussion. Although the question still remains open, being as it is one of the many cruces of Indian chronology, it seems quite probable that Daṇḍin lived sometime between the middle of the 7th and the beginning of the 9th c.

Then follows a detailed presentation of the history of the transmission of the Tibetan translation (1.4). This section begins with a discussion on the incomplete Tibetan translation of the Kāvyādarśa as found in Sa skya Paṇḍita Kun dga’ rgyal mtshan’s Mkhas pa rnams ’jug pa’i sgo (1.4.1). From this work some of the stanzas of the first chapter are cited along with the corresponding version found in the Ganden, Narthang and Peking editions of the Tanjur. An interesting question is whether Sa skya Paṇḍita himself translated these stanzas or whether this was done by others on his direct orders, while he merely took over this translation into his own work. Thereafter the translation of Śoṅ ston and Lakṣmīkara is discussed (1.4.2). Here the colophons found in the canonical versions of the Ganden, Narthang and Peking editions of the Tanjur, on the one hand, and of the Cone and Derge editions, on the other, have been cited and translated ( A few words are said about Bla ma dam pa Chos kyi rgyal po (1235–1280), known better as ’Phags pa Blo gros rgyal mtshan, and Dpon chen Śākya bzaṅ po, both of whom are known to have ordered the translation of the Kāvyādarśa (1.4.3). More information is submitted about the two translators, Śoṅ ston and Lakṣmīkara (1.4.4). The question of the dating of this and other translations by them is also considered (1.4.5). According to the conclusion reached in the book, the Tibetan translation of the Kāvyādarśa must have been presented in the holy Sa skya monastery around the end of the 1260s, or more probably, at the beginning of the 1270s. Then the impact on the transmission in later centuries of the more significant Tibetan scholars, both editors and commentators, is referred to (1.4.6–9). This section also contains translations of the colophons appended to the bilingual editions of Źa lu Lo tsā ba and Si tu Paṇ chen.

In the second part of the book the evidence of the textual material that has been used for the critical edition of the Sanskrit and the Tibetan texts is discussed in detail (2. Transmission). The essential information about the manuscripts, xylographs and other sources together with the corresponding sigla has been supplied (2.1). The two subsequent sections are devoted to an evaluation of the transmission of the Sanskrit and the Tibetan text respectively (2.2 and 2.3).

For the Sanskrit sources the readings found only in one source are first evaluated (2.2.1) and then the relationship between the different sources is discussed (2.2.2). The rather important question concerning which Sanskrit manuscripts were used by Śoṅ ston and Lakṣmīkara for their translation into Tibetan is also dealt with (2.2.3). It has been shown that it is practically impossible to judge which sources exactly the translators themselves have used, for we have at our disposal only subsequently revised versions of their translation. However, already for the first redaction done by Dpaṅ Lo tsā ba and included now in the Ganden, Peking and Narthang Tanjur editions, it can be concluded with certainty that this text was based on Sanskrit sources belonging to at least two different lines of transmission. The Tibetan translation in this very early version indicates the influence of a basic text such as that commented on in the Ratnaśrīṭīkā or of the commentary itself, and at the same time it reflects a state of the Sanskrit text which is presumed to be in the line of transmission after my hyparchetypus χ- Besides this, conclusions are drawn about the Sanskrit manuscripts used by the Tibetan scholars who prepared the bilingual editions (2.2.4). For the sake of completeness a few remarks about the South Indian tradition of the Kāvyādarśa are also made, relying here entirely on secondary sources (2.2.5). The conclusions of the discussion are shown in the form of a stemma in the end of this section (2.2.6).

A similar procedure was followed for the sources of the Tibetan text, but here the discussion was facilitated by the availability of more internal and external evidence about the relationship between some of the sources. First, the variants of the two canonical redactions are taken into consideration (2.3.1) – the one prepared by Dpaṅ Lo tsā ba in the Ganden, Narthang and Peking Tanjur editions ( and the other by Sñe thaṅ Lo tsā ba in the Derge and Cone Tanjur editions ( Second, the three sources of the bilingual versions are examined (2.3.2). Whereas it is shown that they form a distinct group in themselves, it is possible to further distinguish the versions prepared by Źa lu Lo tsā ba, by Si tu Paṇ chen and by an anonymous scholar who generally followed Źa lu Lo tsā ba’s version but who was also influenced by Si tu Paṇ chen’s redaction and some of the canonical editions. In the case of the manuscript with Źa lu Lo tsā ba’s redaction there are clear indications that in the course of transmission the text has been contaminated by other versions. The relationship of the bilingual non-canonical texts to the canonical versions is discussed separately (2.3.3). In this connection it is important to notice that the Tibetan text already in the first bilingual edition has been constituted on the basis of both canonical redactions and, besides this, the commentaries of Dpaṅ Lo tsā ba and Snar thaṅ Lo tsā ba have apparently also been consulted. With regard to the basic text of the Tibetan translation it was necessary to also examine the evidence of the more important commentaries (2.3.4). Special consideration was given to the commentaries by Dpaṅ Lo tsā ba (, Snar thaṅ Lo tsā ba (, Rin spuṅs pa Ṅag dbaṅ (, Bod mkhas pa Mi pham ( and Bstan ’dzin chos kyi ñi ma ( It was significant to pay attention to the fact that Dpaṅ Lo tsā ba, as a rule, does not use the collective particle dag while translating Sanskrit plurals, whereas Śoṅ ston shows an extreme predilection for dag ( This simple clear-cut characteristic of the translation technique of both scholars often helps to prove with great probability where the original version of Śoṅ ston/Lakṣmīkara’s translation is retained and where, and to what extent, Dpaṅ Lo tsā ba has changed the text. Eventually one can differentiate between eight phases in the transmission of the Tibetan translation (2.3.5). As in the previous section, the foregoing discussion is shown in the form of a stemma – in this case inevitably a more elaborate one (2.3.6).

The third part of the book contains the critical edition itself (3. Edition). The preliminary notes concerning the technique of constitution of the Sanskrit and the Tibetan texts are followed by a list containing all the sigla used (3.1) and the tables of contents (3.2) for the first chapter in Sanskrit (3.2.1) and in Tibetan (3.2.2). Then the critical edition is presented (3.3). The Sanskrit text is printed in the Devanāgarī script and the Tibetan – in the Dbu can script. The Sanskrit and the Tibetan texts are given on facing pages with the critical apparatus supplied below each stanza.

The fourth part of the book includes a new German translation of the first chapter of Daṇḍin’s work (4. Translation). At the beginning some explanatory notes are given (4.1). These notes concern both Otto Böhtlingk’s more than a hundred year old incomplete translation based on Tarkavāgīśa’s commentary and the new translation of the first chapter supplied in this book. After a table of contents in German (4.2) the translation itself is presented (4.3). My primary aim was to prepare a readable scholarly translation which reflects, as far as possible, the Sanskrit original as given now in the critical edition. Wherever reasonable this translation has been done in agreement with Ratnaśrījñāna’s authoritative commentary. Thus, in the majority of cases this will help one to see how the text is explained in the Ratnaśrīṭīkā. In a number of cases besides the translation of a particular stanza, the Sanskrit text itself has also been given in analytical transliteration in order to facilitate a better understanding of the text exemplifying particular phonetic features in the Sanskrit language.

The fifth part of the book contains two glossaries (5. Glossaries). Both are made on the basis of the Sanskrit and the Tibetan texts as given now in the critical edition. All significant variants have also been included and mentioned where appropriate. First, there is the trilingual Sanskrit-German-Tibetan glossary to the first chapter of the Kāvyādarśa (5.1) which is followed by the bilingual Tibetan-Sanskrit glossary to the same chapter of the Sñan ṅag me loṅ (5.2). In the glossaries all Sanskrit and Tibetan words are given in Roman transliteration.

In the sixth part seven appendices are supplied (6. Appendices): 1. The available information about twenty old Sanskrit commentaries on the Kāvyādarśa collected in one place (Appendix 1); 2. A bibliography of the modern editions of the Sanskrit and the Tibetan basic texts together with the commentaries on them (Appendix 2); more than eighty publications are listed, in a number of cases with detailed annotation; 3. A collection of the abbreviations (skuṅ yig) and the contractions (bsdu yig) found in the sources of the Tibetan text (Appendix 3); 4. A list of the variously written transliterations of Sanskrit words in the Tibetan sources (Appendix 4); 5. A list of the scribal errors found in the primary sources of the Sanskrit text (Appendix 5); 6. A list of erroneous readings found in the basic text cited in Snar thaṅ Lo tsā ba’s commentary as transmitted in a Bhutanese manuscript copy (Appendix 6); 7. An attempt to represent the rather complicated history of the Tibetan translation of the Kāvyādarśa in a schematic way (Appendix 7).

The seventh part contains four concordances and two indices to both the Sanskrit and the Tibetan texts (7. Concordances and Indices). The first concordance specifies where the Sanskrit stanzas of the first chapter of the Kāvyādarśa can be found in the Nepalese manuscripts and the Tibetan bilingual editions (7.1). The second concordance helps in locating the Tibetan text of the first chapter of the Sñan ṅag me loṅ in both the canonical and the non-canonical sources used for the present edition (7.2). The third concordance facilitates a speedy location of the Tibetan stanzas of the Sñan ṅag me loṅ cited and explained in the particular versions of the four Tibetan commentaries by Dpaṅ Lo tsā ba, Snar thaṅ Lo tsā ba, Rin spuṅs pa Ṅag dbaṅ and Bod mkhas pa Mi pham (7.3). The fourth concordance gives information about the stanzas of the first chapter of the Kāvyādarśa which Bhoja has cited in the Sarasvatīkaṇṭhābharaṇa and in the Śṛṅgāraprakāśa (7.4). For easy reference two indices of verses, for the Sanskrit and the Tibetan texts, have been included (7.5 and 7.6). Here both texts are printed in Roman transliteration.

At the end of the book are a bibliography (8. Bibliography), a list of the abbreviations (9. Abbreviations) and this summary (10. English Summary).



* I thank Dr. Jayandra Soni (Marburg) very much for checking the English of this summary.

  1. It will not be out of place to cite what West aptly says in connection with Greek and Latin works, because it applies in an Indian context as well: “Textual criticism is not the be-all and end-all of classical scholarship, which is the study of a civilization. But it is an indispensable part of it. By far the greater part of our knowledge of that civilization comes to us from what the ancients wrote. In almost all cases those writings have survived, if they have survived at all, only in copies many stages removed from the originals, copies of which not a single one is free from error. Often the errors are so great that it is no longer possible to tell what the author meant to say. It follows that anyone who wants to make serious use of ancient texts must pay attention to the uncertainties of the transmission [...] if he is not interested in the authenticity and dependability of details, he may be a true lover of beauty, but he is no serious student of antiquity.” (West 1973, pp. 7–8).
  2. See Csoma de Körös 1839, p. 581.
  3. See van Manen 1923, p. 502.
  4. See Das 1904, pp. 153–154.
  5. Van der Kuijp 1986a, p. 37.
  6. Rajapatirana 1991, Introduction.
  7. For the dating of this manuscript and a palaeographic study with tables of the script see Dimitrov 2002, pp. 27–78.