Early Tamil Epigraphy: From the Earliest Times to the Sixth Century A.D. c25.s15: 5.15 Tamil-Brāhmī legends on Sātavāhana coins

Early Tamil Epigraphy: From the Earliest Times to the Sixth Century A.D. c25.s15

Language: English

c25.s15.t1 5.15.1 Previous studies — 11 parts
c25.s15.t2 5.15.2 Coin-legends — 13 parts
c25.s15.t3 5.15.3 Palaeography of the coin-legends on the reverse — 12 parts
c25.s15.t4 5.15.4 Chronology of Tamil-Brāhmī inscriptions in the light of the coin-legends — 2 parts
c25.s15.t5 5.15.5 Language of the coin-legends on the reverse — 3 parts
c25.s15.t6 5.15.6 Literal translation — 1 part
c25.s15.t7 5.15.7 Reason for the employment of Tamil in the coin-legends — 1 part

5.15 Tamil-Brāhmī legends on Sātavāhana coins

c25.s15.1

Though it has been known for a long time that the silver portrait coins of the Andhra Sātavāhana rulers exhibit bilingual legends in two scripts, it is only recently that it has been established that the legends on the reverse of these coins are in Tamil, written in the Tamil-Brāhmī script. In view of the importance of these coin-legends for the study of Tamil-Brāhmī and the chronology of its evolution, I shall present here the essential arguments concentrating on more recent developments.

5.15.1 Previous studies

c25.s15.t1.1

The literature on the silver portrait coins of the Sātavāhanas is extensive; I shall refer only to select publications relating to the script and the language of the legends on the reverse of the coins. Sircar³ was the first to point out the occurrence of the "Dravidian language and alphabet of about the second century A.D." in the coin-legends. According to him, the legend on the reverse is in "the Dravidian script and the Dravidian Prakrit (resembling Tamil more than Telugu)". Tamil epigraphists who have analysed the palaeographic and linguistic features of the legends have concluded that the language is Tamil and the script akin to that found in the Tamil cave inscriptions (Nagaswamy,⁴

c25.s15.t1.2

³ D.C. Sircar 1963-64b: pp. 247-252.

c25.s15.t1.3

⁴ R. Nagaswamy 1968: pp. 200-202; 1985: pp. 105-114.

c25.s15.t1.5

Figure 5.9. Silver portrait coins of the Sātavāhanas with Prakrit and Tamil legends. Puḷumāvi (1 & 2) and Sātakarṇi (3 & 4).

c25.s15.t1.6

Panneerselvam,¹ Mahadevan² and Krishnan³). Sarma⁴ has argued that the language is Old Telugu and that the coin-legends are in Telugu verse. More recently, Ajay Mitra Shastri⁵ reviewed the situation and came to the conclusion that the language of the coin-legends on the reverse is "neither Tamil nor Telugu but Dravidian from which later both these languages separated". Mention should also be made of the recent paper by Somasundara Rao,⁶ a Telugu scholar, who "rules out the possibility of Telugu having been used in 2nd century A.D." According to him, the language is "Tamil or Dravidi Prakrit (Proto-Dravidian) language".

c25.s15.t1.7

¹ R. Panneerselvam 1969: pp. 281-288.

c25.s15.t1.8

² I. Mahadevan 1985b: pp. 121-128.

c25.s15.t1.9

³ K.G. Krishnan (ed.) 1989: No. 143.

c25.s15.t1.10

⁴ I. K. Sarma 1980: pp. 107-125.

c25.s15.t1.11

⁵ Ajay Mitra Shastri 1991: pp. 45-60.

c25.s15.t1.12

⁶ C. Somasundara Rao 1997: pp. 17-21. Ajay Mitra Shastri lamented in his paper (1991: p. 52, n. 48) that "generally speaking, Telugu-speaking scholars favour Telugu and Tamil scholars, Tamil". Somasundara Rao's paper proves that scholarship can transcend such partisan considerations.

5.15.2 Coin-legends

c25.s15.t2.1

The silver portrait coins with bilingual legends were issued by a succession of Andhra kings from Vāsiṣṭhīputra Puḷumāvi to Vijaya Sātakarṇi, who ruled from about the end of the 1st century to the beginning of the 3rd century A.D. The coins are in an identical format and depict the bust of the king on the obverse and certain dynastic emblems on the reverse. The legend on the obverse giving the name of the king is in Prakrit written in the well-known Southern Brāhmī script of the period. The legend on the reverse is a literal rendering of the Prakrit legend in Tamil written in the Tamil-Brāhmī script of the Tamil cave inscriptions of the same period (Fig. 5.⁹). Typical of the series are the following two examples:⁷

c25.s15.t2.2

⁷ See Nos. 9 & 10 in Table 1.6. The coin-legends are put together from several specimens.

c25.s15.t2.3

On the coins of Puḷumāvi (see coins 1 & 2 in Fig. 5.⁹):⁸

c25.s15.t2.4

⁸ Coin No. 1: Puḷumāvi. National Museum, New Delhi ( Accession No. 79.423 /1). Coin No. 2: Puḷumāvi. Indian Institute of Research in Numismatic Studies, Anjaneri (Acc. No. 134/1-11).

c25.s15.t2.5

c25.s15.t2.6

On the coins of Sātakarṇi (see coins 3 & 4 in Fig. 5.9):⁹

c25.s15.t2.7

⁹ Coin No. 3: Sātakarṇi. National Museum, New Delhi (Acc. No. 74.75/1). Coin No. 4: Sātakarṇi. Indian Institute of Research in Numismatic Studies, Anjaneri (Acc. No. 134/2-6).

c25.s15.t2.8

c25.s15.t2.9

Remarks on the legends on the reverse:

c25.s15.t2.10

(i) The Tamil expression *makaṉ-ku* is replaced by the hybrid *puta-ku* in the issues of Yajña and Vijaya Sātakarṇi.

c25.s15.t2.11

(ii) The Tamil expression *tiru* is replaced by *ciru* (for *ciri* < Pkt. *siri*) in the above two issues.

c25.s15.t2.12

(iii) On the coin of Siva Siri Puḷumāvi, the name is spelt *purumāvi* (R. Nagaswamy 1985).

c25.s15.t2.13

(iv) The word *vāciṭṭi* is spelt [*vāci*]*ṭhi* in the issue of Vijaya Sātakarṇi.

5.15.3 Palaeography of the coin-legends on the reverse

c25.s15.t3.1

The critical palaeographic features of the coin-legends on the reverse, which prove conclusively that the script employed in them is Tamil-Brāhmī are: the letter *c* with the open loop; the tubular form of *m* with the crossbar in the middle; the presence of the characteristic Tamil-Brāhmī letters *ḻ* and *ṟ*; the case-ending *-ku*; and the *puḷḷi* to indicate the basic consonant *ṭ*. These features which are illustrated in the photographs of the coins (in Fig. 5.9) are discussed below.

c25.s15.t3.2

***c:*** This is the later Tamil-Brāhmī form in which the closed loop at the left seen in the earlier forms opens out (see clear examples in coins 2 & 4 in Fig. 5.9). This form is attested in the recently discovered Tamil-Brāhmī cave inscription at Mannarkoil (No. *89*, ca. 2nd century A.D.). The earlier reading of this letter in the coin-legends as a rare form of *h* turned to the left produced strange expressions like *arahaṇa*, *vāhiṭṭi* and *hātakaṇi* not attested elsewhere. Treating the letter as *c* yields meaningful readings like *aracaṉ*, *vāciṭṭi* and *cātakaṇi*, which are the appropriate Tamil renderings of the corresponding Prakrit expressions *rāñō*, *vāsiṭhi* and *sātakaṇi* in the Prakrit legends on the obverse.

c25.s15.t3.3

***m:*** It is noteworthy that the bilingual legends employ the Southern Brāhmī form of *m* in the Prakrit legends on the obverse and the Tamil-Brāhmī form of the same letter in the Tamil legends on the reverse. The letter actually seen on the coins is the later Tamil-Brāhmī form of *m* in which the horizontal 'crossbar' first moves away from the left side of the 'tube' and later projects a little outside to the right (see coins 3 & 4 in Fig. 5.9 for clear examples).¹ This form is attested in the later cave inscriptions from ca. 4th century A.D. (e.g., Nos. *85*, *101*, *108*, *112*, *115*). The evidence of the cave inscriptions indicates that the correct reading of the form seen in the coin-legends is *ma* and not *mā* as read earlier.

c25.s15.t3.4

¹ See especially R. Nagaswamy 1985: Fig. at p. 109.

c25.s15.t3.5

***ḻ:*** The later form of the Tamil-Brāhmī letter *ḻ* with the loop as in the Tirupparankunram cave inscription (No. *55*, ca. 1st century A.D.) occurs in the name *puḻumāvi* in the coin-legend on the reverse (corresponding to the form *puḷumāvi* in the Prakrit legend on the obverse). The earlier readings of this letter as *ḷu* or *ḍu* were merely from the context without taking into account the actual forms seen in the coin-legend on the reverse. In the three issues I have seen (from clear photographs of exceptionally well-preserved forms), the letter is written as *ḻu* in two cases (coins 1 & 2 in Fig. 5.9) and as *ḻa* in one case.² A somewhat similar-looking Brāhmī letter *j* occurs in the Asokan inscriptions at Khalsi,¹ which does not suit the context. On the other hand, treating the letter as the Tamil-Brāhmī form of *ḻ* (as in the Tirupparankunram cave inscription) yields a meaningful reading suited to the context. The occurrence of *ḻ* in this coin-legend is of exceptional importance; it provides us for the first time with the indigenous form of the name of this famous and powerful Andhra monarch confirming his Dravidian lineage; it proves that the script of the coin-legends on the reverse of these coins is Tamil-Brāhmī; it also confirms the date of ca. 1st century a.d. for the Tirupparankunram cave inscription.

c25.s15.t3.6

² In a coin in the National Museum, New Delhi. The letter *ḻa* is also seen clearly in the Nagarjunakonda mould of the coin of Puḷumāvi (I.K. Sarma 1980: Fig. P-7, Pl. XVII).

c25.s15.t3.7

¹ C. S. Upasak 1960: Appendix No. 13, Figs. (9) & (10).

c25.s15.t3.8

***ṉ*: A closer examination of the legends on the reverse of these coins has revealed that the Tamil words *aracaṉ* 'king', and *makaṉ* 'son' are sometimes written with the alveolar *ṉ* of the Tamil-Brāhmī script earlier misread as *ṇ*. In the best-preserved specimens (see coins 2 to 4 in Fig. 5.9), the top horizontal line of the letter, unlike the bottom horizontal line, does not extend to the left of the vertical as in *ṇ*. The form is also unlikely to be Brāhmī *nā* which does not suit the context. The form is most probably a variant of the Tamil-Brāhmī alveolar *ṉ*, the difference being the absence of the downward stroke at the top right.² The occurrence of the alveolar nasal *ṉ* in the coin-legends is conclusive evidence of the language being Tamil and the script Tamil-Brāhmī, as this character and its sound do not occur in any other script. The word-final consonant *ṉ* in *aracaṉ* and *makaṉ* has to be read as basic in accordance with the orthographic convention in the Tamil-Brāhmī script and not as *ṉa*.

c25.s15.t3.9

² This form appears to be a further development of the tendency towards flattening of the curve at the top of the vertical in *ṉ*. Cf. Nos. *21* and *85* in the Corpus.

c25.s15.t3.10

***-ku*: The case-ending *-ku* is added to the nouns as in *aracaṉ-ku*, *makaṉ-ku*, etc. The earlier reading of this letter as *-sha* produced strange forms like *arahaṇa-sha*, *mākaṇa-sha*, etc., not attested elsewhere. The letter looks like *-ku* and adopting this value yields meaningful expressions appropriate to the context and corresponding to the genitive forms in the Prakrit original with the case-ending *-sa* on the obverse.

c25.s15.t3.11

***The puḷḷi*: It is well-known that the *puḷḷi* occurs with the basic consonant *ṭ* in the expression *vāciṭṭi* (< Pkt. *vāsiṭṭhī*) in the legends on the reverse of the coins of Vāsiṣṭhīputra Puḷumāvī, and those of his successors with the same *gōtra* name except Vijaya Sātakarṇi (see clear examples in coins 2, 3 & 4 in Fig. 5.9). The employment of the *puḷḷi* to mark the basic consonant is known only to the Tamil grammatical tradition.³ The earliest occurrence of the *puḷḷi* in the Tamil-Brāhmī cave inscriptions is at Anaimalai (No. 60, ca. 2nd century a.d.). The occurrence of the *puḷḷi* confirms that the legends on the reverse of the Andhra coins are written in the Tamil-Brāhmī script as no other variant of Brāhmī employed this device.

c25.s15.t3.12

³ *Tol. Eḻu.* 15-17. See also the discussion on the *puḷḷi* in section 5.14.

5.15.4 Chronology of Tamil-Brāhmī inscriptions in the light of the coin-legends

c25.s15.t4.1

A comparative study of the palaeographic features of the Andhra coin-legends with similar features found in the Tamil-Brāhmī inscriptions confirms the chronology of the latter. The late palaeographic forms of *c*, *m*, *ḷ*, *ṉ* and the *puḷḷi* noticed in the coin-legends (Fig. 5.10) occur in the cave inscriptions from about the end of the 1st century or the beginning of the 2nd century A.D. coinciding with the period during which the silver portrait coins with bilingual legends were known to have been issued by the Sātavāhanas.

c25.s15.t4.2

Figure 5.10. Tamil-Brāhmī characters in Sātavāhana coin-legends.

5.15.5 Language of the coin-legends on the reverse

c25.s15.t5.1

There can hardly be any doubt that the language of the legends on the reverse of the coins is not Prakrit. The only question for consideration is whether the language is Old Telugu or Tamil. *Prima facie*, the case for Old Telugu would seem stronger. The Sātavāhanas were Andhras; most of the silver portrait coins are found in the northern and northwestern regions of their dominions far away from the Tamil country. Telugu was the spoken language of Andhra from very ancient times, certainly much before the invention of the Brāhmī script. In spite of these considerations, there are formidable objections from the linguistic point of view to identify the language of the coin-legends as 'Old Telugu'. The terms *aracaṉ*, *makaṉ*¹ and *tiru* are not found in Old Telugu. The prothetic prefix *a-* in *aracaṉ* is a characteristic Tamil grammatical usage. The pronominal ending *-aṉ* as in *aracaṉ* and *makaṉ* is in Tamil; the corresponding Old Telugu ending *-(a)nṟu* is conspicuously absent from the coin-legends. The employment of the dative *-ku* as the genitive case suffix in certain contexts is sanctioned by Tamil grammatical tradition.² One has to conclude from the linguistic evidence that the language of the legends on the reverse of the coins is Tamil and not Old Telugu.

c25.s15.t5.2

¹ *magaṇṟu* with the meaning 'son' occurs rarely in Old Telugu inscriptions (K. Mahadeva Sastri 1969: Word Index), but not the form *makaṉ* seen in the Sātavāhana coin-legends.

c25.s15.t5.3

² *Tol. Col.* 94 cited in R. Panneerselvam 1969: pp. 285-286, n. 15.

5.15.6 Literal translation

c25.s15.t6.1

There is, however, a linguistic problem presented by the literal word-for-word translation of the Prakrit legends resulting in texts which are not in idiomatic or correct Tamil. Thus, while Prakrit requires every attribute to be in the genitive case to be in accord with the substantive noun in the genitive, there is no such usage in Tamil. The repetitive addition of *-ku* to the attributes as in *aracaṉ-ku* and *makaṉ-ku* is not in accordance with Tamil grammar. It is clear that the scribe was not too familiar with Tamil and mechanically translated each Prakrit expression literally into Tamil producing the irregular forms seen in the coin-legends.

5.15.7 Reason for the employment of Tamil in the coin-legends

c25.s15.t7.1

The final question to be considered is why the Andhra kings chose Tamil and not Telugu for the legends on the reverse of their silver coins. The question does not admit of an easy answer and one can only speculate on the probable reasons. My suggestion is that the Sātavāhanas employed Prakrit and Tamil in their coin-legends as these were the only two languages used for political and administrative purposes in South India at the turn of the Christian Era. The Sātavāhanas apparently wanted the legends on their silver coins to be understood by the people of the Tamil kingdoms with whom they must have had active commercial intercourse.