According to widespread opinion ancient Indian drama is one of the most valuable creations of Indian kāvya. It is therefore not a matter of chance that classical Indian literature has maintained a sustained position in Europe’s horizon for the first time through the translation of a drama. In 1789 Sir William Jones’ original translation of Kālidāsa’s famous Śakuntalā was published, which two years later had already been rendered into German by Georg Forster. The educated circle in Germany then also got to know Indian drama through this secondary translation. Herder and Goethe may be mentioned here as representatives of this group who took to Kālidāsa’s piece with virtual exuberant enthusiasm. The first academic work, in the strict sense, began with Horace Hayman Wilson’s Select Specimens of the Theatre of the Hindus in the years 1826/27. Since then, and to the present day, Indian drama has been the subject of numerous indological and other publications and has thus enjoyed a relatively continuous attention of indologists for the period of about 170 years. This attention corresponds to the esteem which the forms of theatre enjoy in traditional Indian histrionics, according to which drama is the best literary genre of all.

The starting point of my work is the play Nāgānanda (“Joy of the Snakes”) composed by King Harṣadeva in the first half of the seventh century CE, which may not only be regarded as one of the classical works of this literary form, but also one that is often consulted by the Indian poeticians themselves. In his as yet unreplaced standard work Geschichte der indischen Litteratur (“History of Indian Literature”) Moriz Winternitz in 1922 describes the Nāgānanda “as one of the most interesting works of Indian literature”, and also in Kindlers Literatur Lexikon (“Kindler’s Encyclopedia of Literature”) the Nāgānanda is in any case still recorded as “one of the most interesting Indian dramas”. So it is not surprising that remarks about the Nāgānanda in general treatises on Indian theatre are practically never omitted and since 1864—the publication of the editio princeps—at least 58 different editions and translations of this work have, to my knowledge, been published. It was all the more surprising for me in the course of my concern with this work to come across a series of both general and specific problems which in no way may be seen as being clarified.

The occasion for my concern with this accepted well-mastered work was the discovery made by Professor M.Hahn of Germany in the beginning of the 1980s of the hitherto oldest known manuscript of the Nāgānanda belonging to the National Archives, Kathmandu. The manuscript is dated 1155 or 1156 and is, thereby, a few centuries older than all the known manuscript testimonies from India. In the work for my Masters degree I showed that the Nāgānanda has survived in at least five recensions, among which the version represented by the manuscript from Nepal in many ways clearly contained the best text in the original languages. Besides this, of particular importance is the translation of the Nāgānanda contained in the Tibetan Tanjur, which was completed in the second half of the thirteenth century, and whose Sanskrit original may well have been even a few centuries earlier. The Tibetan translation represents an independent fifth recension which, however, indicates a special affinity to the Nepali version. This situation proved to be especially favourable for me because, in this way, both the supposedly best versions could often be mutually corrected. Finally, mention must be made of a very useful Sanskrit commentary possibly composed in the thirteenth or fourteenth century by the South Indian Śivarāma, who not only preserved the South Indian version of the wording fairly well, but who also offers invaluable help for the linguistic and literal interpretations of the play.

Since I had particularly analyzed only the fifth of the six acts of the play for my Masters, I had intended at the outset to comment on the entire text of the play in detail, from a philological perspective and, over and above that, to explore the technique and peculiarity of the Tibetan translation. For this purpose I first collected and discussed all the textual variants. However, during my work it became clear that such a method would lead to a dissertation of gigantic and unwieldly proportions and, indeed, would not have made pleasant reading for a putative reader because such a work would be ideal rather for reference purposes, or as a mere documentation of the theses themselves. My occupation with the Nāgānanda, however, unintentionally led to formulations of questions, answers to which could be of general significance for Indian drama. Individual text-critical or interpretative problems of the Nāgānanda thus made up the starting point of each chapter, which I then had to expand into questions which in some respects go beyond the Nāgānanda.

The first chapter of my thesis deals with text-critical and interpretative problems of the initial benediction, which in the Nepali tradition was an unfortunate victim of a page loss. This is why the Tibetan translation was especially useful, also because it contains an additional stanza at the beginning. When in 1970 Professor Hahn compared this benediction with the three introductory stanzas of the play Lokānanda (“Joy for the World”) by Candragomin, he had already come to the convincing conclusion that one benediction must have served as a model for the other, and that the Nāgānanda must have also had a three-stanza Nāndī, as it is technically called. Here I tried to show that none of the views—presented in the literature even up to 1992—which speaks for the priority of the Lokānanda, independent of the dating of the two authors, stands up to critical examination. In this context I also dealt with the question about the source which Candragomin might have used for his Lokānanda, and came to a conclusion which also differs from the current view. As regards dating Candragomin to the fifth century, I agree with Professor Hahn in all points of the argument presented in his various publications. His dating, however, has not remained unchallenged, which is why I again dealt with the other standpoints and tried to criticize them with further considerations.

So, two of the three benedictory stanzas of the Nāgānanda make up the beginning of another “hymn” (*Mārajitstotra) of three stanzas which are preserved in the Tibetan Tanjur. Here I tried to show (1) that the composer of the hymn took both the stanzas from the play, (2) that the hymn was probably not by the author of the Nāgānanda, and (3) that the hymn was presumably translated into Tibetan from another Sanskrit version of the hymn. Furthermore, I think that I have given a plausible reason why the first Nāgānanda-stanza which, in fact, is only in the Tibetan translation, was not included in the hymn, so that its absence there does not speak against the authenticity of the stanza belonging to the Nāgānanda.

Then follows a detailed discussion of the Tibetan translation of the three benedictory stanzas of the Nāgānanda, in comparison with the different Indian versions. Here I tried, on the one hand, to correct or to clarify a series of hitherto unclear sections and, on the other, to present further arguments for the originality of the first benedictory stanza. With this stanza Harṣadeva apparently adopts—as indeed other authors do—rules laid down in the oldest Indian textbook for the performing arts, the Nāṭyaśāstra. Moreover, with this stanza Harṣa seems to want to intentionally surpass the Nāndī of his ideal, Candragomin, on a specific point and is not just a mere imitation.

The following section of chapter 1 I think contains perhaps the most important implications for Indian drama. It concerns the form of the first stage direction which in all classical dramas directs the entry of the stage director, the sūtradhāra, and his debut has been preserved in two different forms. The length and complexity of this section, albeit important, make it almost impossible to say anything more about it in the context of this summary.

The implications for Indian drama mentioned above concern, among other things, the various meanings of the term nāndī and the correlation, assumed over and over again in the literature on the subject, that is supposed to exist between the form of the first stage direction and the practice of the religious ceremony of the “consecration of the stage” (pūrvaraṅga) which, according to the instructions of the textbooks, should precede the actual play. Besides, it is also assumed that the form of the entry direction in the South Indian tradition represents an older period of drama technique. On the basis of these mutually different assumptions regarding the stage direction and the consecration, conclusions, in part far-reaching, have been drawn concerning the development of the theatre technique and the relative chronology and authorship of a few classical works which are usually difficult to date. When one now critically examines the dramas and the relevant observations of the Indian poeticians one will find, however, that none of the assumptions given are sound. The stage consecration, on the one hand, and the text of the plays, on the other, belong, I believe, to two completely different traditions, so that it is not possible to draw conclusions for the form of the one from the form of the other.

In contrast to this, I have tried to explain the content of the different wording of the two forms of the stage direction, and to trace them back, against what appears plausible at first sight, to a uniform practice of reciting the entry benediction. However, no chronological conclusions are possible with such an assumption anymore. Moreover, it was necessary in this context to examine and, where possible, to explain individual places of other texts. Out of this section I would like to refer here briefly to my investigation about the two terms for the stage director: sūtradhāra and sthāpaka, about whom also far-reaching and, in my opinion unconvincing, conclusions have been drawn. Furthermore, these conclusions concern, on the one hand, a supposed development of drama technique with chronological implications and, on the other hand, the apparently proven priority, in India, of the puppet show over drama. One of the things I have tried to show here is how the function of the sthāpaka can be understood on the basis of the reconstructed practice of the theatre contained in the texts of the plays. Furthermore, I have tried to justify why I think that the term sūtradhāra probably did not originally mean “a thread bearer”, i.e., a puppeteer, but represents rather a learned person who “carries” (dhāra) the “text” (sūtra) in his memory. In this context I have also suggested a different explanation for the original meaning of the term pūrvarańga.

The chapters that follow the above briefly deal with individual problems which were necessary to consider. In the reference to Indra’s festival at which the Nāgānanda is being performed, according to what is said in the prologue, I see a hint at the “prevailing mood” (rasa) going through the piece. This also compares well with Harṣadeva’s two other plays which are said to have been staged on the occasion of the spring festival (ch. 2). A further reference in the prologue, according to which the Nāgānanda has been interpreted as being based on a Vidyādharajātaka, does not lead to the conclusion, contrary to a view often repeated, either of an independent work now lost to posterity, nor that Harṣa at all drew from a collection of Jātakas. The term is presumably another expression for Vidyādharakathā, as the legend seems to be called in the Bṛhatkathā, a now lost collection of narratives in which the legend was probably contained. This kind of linguistic usage can be reconstructed from the adaptations of the narrative collection that have survived (ch. 3).

Further chapters of my work deal with the expressions for the interludes (ch. 4), as well as for the hero and heroine of the play (ch. 5). Here the documentation and evaluation of the ancient hand-written tradition become very significant because, in this way and in conjunction with the statements of the Indian theoreticians, criteria for an assessment of other textual traditions in India could possibly come to light. In the chapter following this I have tried to show that the term nepathya in the stage directions can not mean “curtain”, but functions merely as statement of direction (ch. 6).

In chapter 7 I have given the genesis of an interpolation in the Central Indian version of the Nāgānanda. Chapter 8 deals with the concluding praise of the Nāgānanda. In it I have tried to reconstruct a terminological development which seems to indicate a change in the actual method of performance. Furthermore, I have tried to improve the wording and the understanding of this stanza. The famous dramatist Bhavabhūti apparently used this praise as a model just as Harṣa, on his part, seems to have done by apparently alluding to Candragomin. The additional concluding stanza of the South Indian tradition is not authentic, but rather an apparently widely spread verse used by scribes.

The last two chapters of my work deal, further, with huge complexes: the Prakrit of the Nāgānanda version from Nepal (ch. 9) and the metre of Harṣadeva’s three plays (ch. 10). Indeed, one of the characteristics of classical Indian drama is its multi-lingualism. Depending on the character, each person speaks Sanskrit, or different Middle Indo-Aryan languages for which the general term Prakrit is used. The problem which an editor faces when dealing with the different Prakrits of a play is that, in contrast to Sanskrit, he is usually confronted with a relatively poor Prakrit tradition of less strictly normed languages. That is why I first dealt with the question whether and to what extent the traditional Prakrit grammarians should at all be consulted when publishing an Indian drama. It is clear that extreme positions are not proper here, but complicated interrelations between poets and grammarians need to be taken into consideration. That is why I think that it is absolutely imperative in this context, at the outset, to critically examine each textual tradition on its own and in doubtful cases to carefully consult the grammarians of the different schools, and of the tradition of related texts. In the other parts of this chapter I dealt with detailed questions concerning the Prakrit of the Nāgānanda version from Nepal, where I tried throughout to restore the relationship to the grammarians and others, on the basis of the tradition based on the oldest possible textual evidences, such as inscriptions or fragments of dramas found in Central Asia. In the process it became further evident that the grammarians, in a few cases, have not yet been reliably interpreted. The most important result of my investigation here was that the Prakrit-forms of the Nāgānanda version from Nepal, in comparison with the sound-system handed down in other dramas of the classical and later period, seem to be relatively ‘archaic’. This situation can be explained in various ways. I assume here for specific reasons that the age and geographic origin of the respective textual evidences are of decisive significance. Consequently, this means that for the time being we have to desist from mutually comparing the Prakrit of individual authors for the purpose of more general questions, e.g., chronological considerations, and that, instead, we should direct our attention completely at the textual tradition in question. A further unhappy conclusion is that not even the transmitted Prakrit by itself can be sufficient to date a specific author.

In the final chapter which deals with the metre of Harṣadeva’s three dramas, I tried at first to examine the places in which metrical and related philological problems are coupled and, wherever possible, I have given explanations and suggested improvements. It is especially important for an assessment of the Nāgānanda version from Nepal to note that even the stanzas which are metrically corrupted in the manuscripts, in contrast to the metrically correct Indian versions, could point to the original wording which can be conjectured. The second part of this chapter deals with the use of the caesuras contained in Harṣa’s plays. It is noteworthy that these caesuras quite often do not conform to the conventional norms. So I have tried to show here that they are regular exceptions which have not yet been observed and which, with one exception, can be divided into three clearly definable groups, which indicate their regularity. A huge corpus of texts has yet to be investigated from this point of view to see if these definable exceptions could indeed be based on a metrical principle of composition.

In the appendix I have discussed the authorship and the dating of a few plays extant in South India, among them the so-called, and famous, ‘Trivandrum Plays’ (or, ‘Bhāsa Plays’).