Critical editions
What is a Critical Edition of a Classical Tamil Text?
A critical edition is an editor’s reconstruction of a text as he supposes it to have been at a particular time in its transmission (...). Although it is a hypothesis, it is made on the basis of all evidence for the wording of the text that the editor can consult (ideally all surviving evidence) and by an editor who has striven to understand as far as possible the ideas of the author(s) as well as the relationships between the sources that make up that evidence, and it is equipped with an apparatus that reports all of that evidence that is relevant to the constitution of the text (in some cases this means all the evidence). Such editions, as yet all too rare, are an invaluable tool for all who are interested - from any perspective - in texts and their transmissions.
Goodall 2003: XXIV
The basic idea of a critical edition is easy to summarise: trying to reconstitute a text that is as close as possible to the original wording, and doing it on the basis of all the remaining witnesses. Since this is a scholarly process and as such can be argued (and is open to error), it is necessary to lay open the working procedure and to defend its results. The first is done by giving the reader access to all the material that has been at the editor’s disposal when coming to a decision, that is, by giving a critical apparatus that contains all the significant variants, significant in the sense that they either represent veritable alternative readings or, if they are clearly erroneous, have taught the editor something about the process of transmission and the development of readings. The second is done by annotation: if there are serious doubts as to which reading to put into the text, if it is unavoidable to emend or if the text has been printed before and the new edition deviates from a reading adopted by its predecessors, an explanation is in order.
In several respects, a scholar of Classical Tamil is working under especially difficult conditions. Every editor working with handwritten material is faced with the problem of having to transcribe an idiosyncratic and often highly individual writing style into a printable standard script, filling in lacunae and recognising and coming to terms with features like regional deviations from standard spelling. In addition to this in the case of Tamil, since the form of script used in manuscripts is ambiguous – marking neither puḷḷi-s, nor the difference between e/ē, o/ō and ā/ra, employing scriptio continua and sandhi –, a Tamil text as it is transmitted does not make sense by itself. Reading is an act of interpretation: separating the words, restoring the ligatures, vocalising. Many graphemes allow for more than one or even several interpretations, and the same is true for word- and metrical borders. Moreover, for the time being our ideas as to standards are at best vague: Which sandhi rules should actually be adopted? How flexible is our metre? Which morphological variations can be ruled out? Is our understanding of syntax sufficient to choose between variants? Which value can be attributed to arguments of context?
Moreover, a Tamil editor is confronted with a very different kind of problem, namely one of a theoretical nature. The tradition of critical editing and its idea of an original wording have been developed in domains where the textual transmission can be reconstructed over long periods of time. Manuscripts in the best of cases reach back into or near to the era of text production, and in many cases they reach back at least several hundred years. In the case of Sanskrit, to bring a point of comparison, often the oldest manuscripts can be shown to go back as far as the 9th century CE, in some cases even earlier. What can be the meaning of the term “original wording” when we deal with a kind of tradition where the oldest testimonies are no older than 200 or perhaps 300 years, while we are talking about texts that are supposedly nearly 2000 years old? And a related question, though from a different background: what can be the meaning of the term “original wording” in the case of highly formulaic poetry? The answer to the first question we can give for the moment is that even if the period spanned by our manuscripts may only be 300 years, at the utmost, it is possible to show a development of the text and its transmission. It may be possible to locate errors and their sources. Thus the original text is, as ever, an unattainable ideal, but nevertheless useful as a heuristic principle. The answer to the second question is that there are indeed a number of variants which can be described as formulaic, for example the variation between two adjectives connected with a system.
But there are several other types of variants which are decidedly not of that kind ‒ regional, lexical, morphological, and syntactical ones, to leave aside for the moment mere spelling- and sandhi variants. Should it not be possible to argue for correctness, or at least for precedence, in such cases? Here the answer, unwelcome as it may be, is that with a reasonably well-documented text it is necessary not only to classify variants, but to consider carefully their sources. The first important distinction to be made is whether a variant comes from a palm-leaf manuscript or from a paper one. While differences between palm-leaf manuscripts can be seen as actual transmission variants, disregarding the mere scribal errors, Tamil paper manuscripts generate their own variants, because they are the place where the under-marked palm-leaf notation is disambiguated into modern and fully marked script. Most paper scribes ‒ except for those who do a mere copying job without adding splits and diacritics ‒ are already at the beginning of a process of editing: splitting poems, lines, metrical feet, eventually sandhi, adding puḷḷi-s, closed kāl-s and double kompu-s, and perhaps even starting to compare, to note variants, to correct and to emend. They have to be regarded as precursors of the early editors.
Thus it seems justified to refer to palm-leaf variants as primary variants, to paper variants as secondary variants – unless, of course, the paper manuscript is a case of what may be called “emergency copying”, that is, simply transferring the text of a disintegrating palm-leaf onto paper, as a rule easily recognisable by the absence of modern diacritics.. In some cases we possess several sets of primary variants, those of each of the distinguishable transmission strands plus one from the old commentary. In the case of primary variants it is extremely difficult to argue for precedence, and it is necessary to do so in notes. There is no justification for the editorial standard process of just silently choosing one’s preference. In the case of secondary variants it may make sense to argue by age and by majority. Since paper variants as a rule are older than print variants they may be counted as older witnesses, unless they are clearly marked as corrections. Isolated variants in print sources usually are either corrections or mistakes on the part of the editor, unless we have reason to believe that the editor had additional material at his disposal.
A problem apart is posed by the quotations. Prima facie those deviating readings could be conceived as primary variants, but unfortunately, for the time being, it is impossible to go back to the primary sources still available for them. Editions, especially later editions, have undergone the usual normalisation process, often even biased by the standard editions of the poetry cited by them available by that time. It is possible to demonstrate a good stretch of that development by comparing the older editions with the more recent ones, and by “old” I mean here editions from the era when the Caṅkam works had not yet been printed. For the Tolkāppiyam Poruḷatikāram, one of the most important sources for quotations, this means going back to the edition by Tāmōtaram Piḷḷai of 1885. There we still find readings that look like copied manuscript evidence, while in most later editions, such as the otherwise admirable Kaṇēcaiyar 1943 and 1948, the editing process has done much to distort the original evidence.