Early Tamil Epigraphy: From the Earliest Times to the Sixth Century A.D.

Abstract

Salomon reviews Early Tamil Epigraphy: From the Earliest Times to the Sixth Century A. D. by Iravatham Mahadevan.

Early Tamil Epigraphy: From the Earliest Times to the Sixth Century A.D. By IRAVATHAM MAHADEVAN. Harvard Oriental Series, vol. 62. Chennai: CRE-A and Cambridge, Mass.: DEPARTMENT OF SANSKRIT AND INDIAN STUDIES, HARVARD UNIVERSITY, 2003. Pp. xxxix + 719.

Early Tamil Epigraphy represents the culmination of Iravatham Mahadevan's lifework on the early Tamil and Vatteluttu cave inscriptions dating from approximately the second century B.C. to the sixth century A.D. Mahadevan has long been acknowledged as the most prominent expert in this field, so it comes as no surprise that he now presents us with a comprehensive and definitive presentation of this important corpus. Mahadevan's claim that "the cave inscriptions of Tamil Nadu, which have generally been regarded as obscure, can now be read" (p. xi) is fully justified, as his edition gives convincing interpretations for nearly every word in the corpus. The two main obstacles to their interpretation in the past, namely "the lack of reliable texts and unfamiliarity with the orthographic rules governing the inscriptions" (ibid.), have now been removed by Mahadevan's energetic and comprehensive fieldwork and by his authoritative analysis of the orthographic systems, respectively.

The majority of the early Tamil inscriptions are either short dedicatory inscriptions in caves or name labels on pottery, coins, rings, and seals. They are written in a locally adapted variant of the north Indian Brahmi script, known as Tamil-Brahmi, which has several interesting varieties and systemic peculiarities. Despite their brevity and stereotyped contents, the old Tamil inscriptions have major significance for the study of the history, culture, language, and especially paleography of early south India. Mahadevan describes the goal of his "Commentary on Inscriptions" (pp. 539-639) as "to situate the Early Tamil inscriptions in the mainstream of Indian epigraphy" (p. 541), and in this regard not only this section in particular but the book as a whole succeeds admirably. For example, as one of many instances of the wider importance of the early Tamil inscriptions for related fields of study, the Jambai inscription "has finally settled the question" (p. 120) of the identity of the Satiyaputras mentioned in Asoka's second Rock Edict. With regard to literary studies, the analyses by Mahadevan and others have provided historical corroboration for various persons mentioned in the Tamil Cankam literature (see, e.g., pp. 85 and 117). For the history of religions in south India, Mahadevan has succeeded in demonstrating that, contrary to what was thought by earlier investigators, "the Tamil-Brahmi inscriptions ... do not reveal any internal evidence for associating the Buddhist faith with the Tamil caves" (p. 126), but rather that they are of Jaina affiliation. This conclusion further leads him to draw important conclusions about the early implantation of Jainism in Tamil Nadu, for which he finds linguistic evidence for links with Karnataka, rather than with Bengal and Orissa as held by some (pp. 108-9, 127-28).

Early Tamil Epigraphy is divided into three main parts: "Early Tamil Inscriptions," presenting a general introduction to the material; "Studies in Early Tamil Epigraphy," with detailed discussions of paleography, orthography, and grammar; and the "Corpus of Early Tamil Inscriptions," a definitive edition of the 110 known cave inscriptions. Although (as explained on p. xi) the cave inscriptions are the main subject of this study, a catalogue of selected specimens from among the hundreds of pottery inscriptions as well as a complete listing of those on coins, seals, and rings is also provided (pp. 60-67).

Throughout, the documentation is thorough, precise, and exhaustive. For example, reproductions of the inscriptions in the corpus are provided both in the form of tracings made directly from the stone according to the editor's interpretation (as explained on p. 82), and of neutral estampages or photographs (pp. 483-537). This dual presentation has the advantage of showing, in the case of the tracings, the expert editor's visual "reading" of the letters which are often indistinct on the rough stone surfaces, and in the estampages, an objective reproduction which permits checking of the readings by other specialists. The readings of all of the cave inscriptions were improved vastly by direct observation of the originals in the course of two series of site visits by Mahadevan and his assistants, first between 1962 and 1966, and again between 1991 and 1996. These field studies enabled the editor to make numerous and substantial improvements on previous readings which had been arrived at by other scholars, or in some cases by himself, on the basis of estampages but "without visiting the caves themselves" (p. 5). Mahadevan's vivid descriptions of the dangers and adventures involved in the fieldwork (e.g., pp. 17 n. 3, 19, and 20 n. 3) allow the reader to share in the enthusiasm, excitement, and adventure involved.

The inscriptions are analyzed in complete detail in various sections of this book with regard to their linguistic, paleographie, cultural, and historical significance. For example, chapter 7, "Grammar," provides an exhaustive linguistic study of the inscriptional corpus, which is further supplemented (not without some duplication) in the notes in the corpus and again in the "Commentary on Inscriptions" that follows it. Detailed comparisons between the phonological and grammatical features of the inscriptions and those of Tamil literature, especially the prescriptions provided in the Tolkappiyam (pp. 261-64), are provided, leading to the conclusion that, contrary to what had previously been believed by some scholars, "starting from accurately copied texts ... it can be demonstrated that the language of the cave inscriptions ... is Old Tamil, not materially different from the language of later Tamil inscriptions or even literary texts" (p. 103). Among other topics that will be of interest to scholars of Dravidian linguistics are Mahadevan's discussions of historical issues such as the question of voicing of medial consonants in early Tamil, for which he finds little evidence in the early inscriptions, concluding that "there was no voicing of consonants in Old Tamil" (p. 251).

Further linguistic data is conveniently provided in the seven appendices, including etymological indices of Dravidian words, Indo-Aryan words, and "doubtful items," as well as an "index to the grammatical morphemes" (appendices iv-vn). The index of doubtful items (pp. 678-79) illustrates the author's typically cautious and nondogmatic approaches to problematic terms for which plausible etymologies can be provided from both Dravidian and Indo-Aryan sources. This measured approach is also amply attested in the commentary section, for instance in connection with the difficult term tara-ani-i, tentatively rendered as 'drip ledge(?)', for which two plausible but uncertain etymologies are proposed with appropriate words of caution (p. 574).

Despite their considerable linguistic and cultural value, the greatest significance of the Tamil-Brahmi inscriptions lies in their paleographie features. Among the several important points established by Mahadevan in this connection is that the early Tamil inscriptions "have finally set at rest the controversy regarding the origin of the Vatteluttu script and proved conclusively its derivation from Tamil-Brahmi" (p. 211). (It is this conclusion that provides the rationale for including in the corpus the early [ca. fifth-sixth centuries A.D.] inscriptions in Vatteluttu). Furthermore, although TamilBrahmi is not the direct parent of the standard and modern Tamil scripts (p. 213), it has had an important indirect influence on their characteristic system of vowel notation, and it is here that the most important point of all regarding the Tamil-Brahmi script lies. From the earliest period, around the second century B.C., Tamil-Brahmi effectively discarded the inherent vowel system that was a fundamental feature of northern Brahmi (p. 227) and its other Indian derivatives, and developed instead several variant systems of vowel notation. The problem of the principles and chronology of these systems has occupied Mahadevan, among other, for decades, and the presentation of this issue in this volume (pp. 226-45) represents the final and authoritative synthesis of his work. He concludes (pp. 227-30) that there were four distinct orthographic systems used among the Tarnil-Brahmi (TB) and related scripts with regard to the notation of vowelless consonants, as follows:

TB-I: An unmarked consonant represents the consonant alone, with no following vowel. A single diacritic sign, visually equivalent to the sign which marks long a in standard Brahmi, indicates in TB-I either a or a, the distinction to be made by the reader "from the linguistic context" (p. 227).

TB-II: An unmarked consonant represents either the vowelless consonant, as in TB-I, or the consonant plus a as in standard Brahmi, with the correct reading to be determined according to the reader's judgment or intuition. The diacritic sign which in TB-I indicated either a or a indicates only a in TB-II, again as in standard Brahmi.

TB-III: Essentially the same as TB-II, except that a consonant which is to be read as vowelless may be marked, though often inconsistently or sporadically, by a pulli or dot. This is essentially the system which, after regularization of the use of pulli, persisted in Vatteluttu and modern Tamil scripts.

The fourth system is attested only among the Bhattiprolu stupa inscriptions, which are however in Prakrit, not in Tamil. It is essentially like the TB-I system, but with the addition of a distinct vowel '' marker to clarify the difference between post-vocalic a and a.

According to Mahadevan's analysis, these orthographic systems developed in chronological order but overlapped in their period of usage; thus TB-I and TB-II were both in use in the first centuries B.C. and A.D., while TB-II and III overlapped from the second to fifth centuries A.D. (fig. 6.4, p. 232). In fact, there seems to be some question as to exactly how distinct these four systems really were, especially in view of the fact that, for example, "both TB-I and TB-II notations occur in contemporary inscriptions at the same site or even within the same inscriptions" (p. 234). Similarly, "since the use of the pulli was optional in practice, the question arises whether there is any real difference between the TB-II and TB-III systems" (p. 230). It is no doubt with these issues in mind that Mahadevan concludes that TB-I and II were not stable or fully developed systems, but rather "transitional, shortlived experiments" in adapting the Brahmi script, which originally developed in connection with Middle Indo-Aryan dialects, to the phonological structure of a Dravidian language (p. 231).

Now that the Tamil-Brahmi inscriptions have been definitively interpreted by Mahadevan, they provide a valuable corpus of data for the study of script typology and the history of writing systems, in that they embody a system, or rather a group of related systems, in a state of flux with regard to their systemic features. The historical and linguistic circumstances underlying this situation are typical of those which promote systemic modifications of scripts, these being prone to occur when a pre-existing script is adapted for use with another language, especially a language belonging to a different family than that of the parent script (see Salomon, "Typological Observations on the Indie Script Group and its Relationship to Other Alphasyllabaries," Studies in the Linguistic Sciences 30 [2000]: 94). Moreover, as in the case of Tamil-Brahmi, such changes frequently involve modified or "improved" methods of representation of vowels, as also, for example, in the classic case of the alphabetization of the Phoenician consonantary when it was adapted for writing Greek. And again as in the case of Tamil Brahmi, such modifications sometimes specifically involve the mode of notation, whether positively by means of a distinct sign or negatively by default, of a particular vowel, which is typically either the statistically most common vowel in the target language or a neutral central vowel, or both (ibid., 98). In this regard, the TB-II system, in which an unmarked consonant may be read either as vowelless or with the neutral/inherent vowel a according to the reader's judgment, is typologically similar to certain non-Indie scripts such as Ethiopian and Meroitic, where vowelless consonants are written with the sign for a weak or neutral vowel which may be suppressed in pronunciation at the reader's discretion (ibid., 93-95, 101). Thus this kind of approximative notation appears to be a "natural" development, insofar as it produces a reasonably efficient graphic system, the ambiguities involved evidently not presenting any serious difficulties for the native speaker of the language being represented.

Nonetheless, in Tamil, unlike Ethiopian and Meroitic, this ambiguous system was not allowed to continue, and the pulli, which was originally used only sporadically and inconsistently in TB-III to mark some vowelless consonants, eventually became standardized and mandatory as a vowel suppressor, as it still is in modern Tamil script. According to Mahadevan, this was "clearly a response to the felt need to obviate the confusion in having to read the unmarked consonantal symbol either as basic [i.e., vowelless] or with the inherent -a as in the earlier system" (p. 230). Mahadevan is no doubt correct, at least in part, to attribute this regularization to a "felt need" for a less ambiguous script, but this explanation should not be assumed automatically, since a comparison with the similarly structured non-Indie systems mentioned above, and more broadly with writing systems in general, shows that consistency, precision, and completeness in the representation of the spoken language is by no means always the prime determinant of a script's development. Many writing systems-among modern scripts, Arabic and Hebrew are prominent examples-seem to be content, so to speak, with partial or approximative representations of the languages that they represent. Such systems have a logic of their own in that, though theoretically imprecise, they tend to be highly economical in terms of speed of writing and recognition and provide little if any difficulty for the native speaker (though they can be very troublesome for the foreign reader). Thus if we are to seek a specific reason why the Tamil scripts, like most Indie scripts but unlike scripts in many other parts of the world, did develop in the direction of a higher degree of precision, this is probably to be connected with specific cultural values; presumably it was the high degree of linguistic awareness and concern with precise linguistic analysis prevailing in ancient India that produced a sense of dissatisfaction with such highly functional but theoretically imperfect systems as those of TB-I and II.

Mahadevan's formulation of the Tamil-Brahmi system also raises interesting issues with regard to script types and terminology. Without going into these issues in detail, Mahadevan appropriately, if somewhat vaguely, characterizes the Tamil-Brahmi scripts as "quasi-alphabetic" (p. 164). He discusses in some detail certain orthographic phenomena in Tamil-Brahmi which superficially look like features of alphabetic writing, in particular "analytical writing" whereby word stems ending in consonants and suffixes beginning with vowels are not joined into syllabic units as they would be in other Indian scripts but rather kept separate, as for example in the proper name Katalan, written syllabically as kata-l-an rather than ka-ta-la-n (p. 243). Similarly, in "pause" spelling, even within a single morpheme a CV syllable is sometimes written with separate C and V signs; thus, for example, a-ti-t-an-am instead of expected a-ti-ta-na-m (p. 244). Although this looks like partial or incipient alphabetization, Mahadevan shows that it only occurs in limited and defined situations, namely "in lieu of doubling of consonants" or "indicating that the following vowel is a diphthong" (ibid.). He therefore cautiously concludes that such features do "not appear to be alphabetic writing" (ibid.).

Still, it remains to provide a suitable designation for a script such as Tamil-Brahmi; "quasialphabetic" is not wrong, but it is not as precise a term as ideally one would like to see. Even more sophisticated modern typologies which go beyond the traditional tripartite classification of logographic, syllabographic, and alphabetic signs, such as the six-type system presented in Daniels and Bright (The World's Writing Systems [New York: Oxford Univ. Press, 1996], 4), do not provide an adequate terminology for the graphic repertoire of Tamil-Brahmi. The closest is their abugida, that is, the Indie/ Ethiopie script type, in which "each character denotes a consonant accompanied by a specific vowel, and the other vowels are denoted by a consistent modification of the consonant symbols" (ibid.). But this term does not suffice for Tamil-Brahmi, especially for TB-I, in which some of the consonantal characters are not "accompanied by a specific vowel," but rather, as shown by Mahadevan, are vowelless, representing only the consonantal phoneme. How one wishes to describe such a system is of course very much a matter of opinion; I would here propose, by way of an amplification of Daniels and Bright's terminology, "partially alphabetized abugida" However this may be, the case of Tamil-Brahmi, like some of the other systems discussed above, shows how even Daniels and Bright's expanded typology does not cover all writing systems in actual use. I do not mean this as a criticism or refutation of their system, which is a vast improvement on previous attempts. Rather, I only wish to emphasize the desirability of understanding this or any sophisticated typology of writing systems as no more than a set of ideal archetypes rather than of rigid categories, and of recognizing that many writing systems, in actual practice, partake of and combine in various ways features of two or more such ideal archetypes.

In this regard, what is particularly interesting about the Tamil-Brahmi system, especially the TB-I variety, is how, though not an alphabet in the full sense of the term, it incorporates among its basic features one of the fundamental characteristics of true alphabets, namely a set of graphs representing vowelless consonants; that is to say, it contains graphs that represent phonemes rather than syllables, which is normally thought to be a defining characteristic of "true" alphabetic scripts. Thus in Tamil-Brahmi we find a system that seems to be taking steps towards full alphabetization at certain points in its history, but that instead of turning into a true alphabet develops into a modified variety of the abugida or Indic alphasyllabic type that remains in use in modern Tamil, where the vowel-canceling pulli obviates the need for the complex and sometimes cumbersome consonantal conjuncts used in most other Indie scripts. Whether the "failure" to develop into a true alphabet in the Tamil country is to be attributed to this being "too radical a departure from all other Indian systems of writing" as Mahadevan (p. 230) would have it, or perhaps rather to other indeterminate factors, is anybody's guess. But in any case, this interesting side-path in the history of script development can serve as the final nail, if one is still needed, in the coffin of obsolete teleological views of the alphabet as the necessary and inevitable culmination of the development of writing systems.

Finally, although I am essentially in agreement, terminological quibbles aside, with Mahadevan's characterization of Tamil-Brahmi as "quasi-alphabetic," I would hesitate to endorse his conclusions about the social effects of the modification of the original Brahmi orthography into the Tamil-Brahmi system, to the effect that "the enormous importance of such a simple, easy-to-learn script in the spread and democratization of literacy can hardly be overestimated" (p. 164). Though often stated or assumed, the alleged relationship between the perceived "simplicity" of a writing system, high levels of literacy, and a democratic society is not well borne out by history and experience. The example of modern Japan, where an extraordinarily complex and difficult writing system coexists with democracy and a very high literacy rate, suffices to at least cast doubt on such alleged causal linkages. Mahadevan does support his claims with concrete evidence in referring to the relatively large number of inscribed potsherds in Tamil-Brahmi, which he takes to imply widespread literacy among the common people (pp. 160-61). But in and of itself, this seems a somewhat fragile basis for such an overarching conclusion, especially one which is a priori questionable on general and comparative grounds. Therefore, while Mahadevan's theory is not impossible, it can hardly be considered proven on the basis of the available evidence.

Such minor issues aside, there can be no question that Early Tamil Epigraphy is a masterwork that will stand the test of time as one of the landmarks of Indian epigraphic studies. The author is to be heartily congratulated for his lifelong labors that have now produced such a magnificent result.