Pramāṇa · Search · About · API

Early South Indian Palaeography c12: Chapter IV: The Language of the Brahmi Inscriptions

Early South Indian Palaeography c12

Language
English

Chapter IV: The Language of the Brahmi Inscriptions

c12.2

The existence of lithic records incised in the Brāhmī script in North India and the Deccan has been known for a very long time. Even as early as the fourteenth century A.D. when Firoz Shah Tughlak shifted the Aśokan pillars from Topra and Meerut to Delhi, it is said, he invited a number of Sanskrit scholars to read the writings on the pillars, but they were not able to decipher their script. It is also said that Akbar, the Mughal Emperor, was inquisitive about the writings on the pillars and that his attempts to have them read met with no better result. And it was only a little over a hundred years ago that the mystery of the script of the Aśokan edicts was unravelled by western scholars like Ch. Lassen and James Princep in the fourth decade of the last century, and that forms the foundation of our knowledge of the ancient art of writing, language and history of India. But the provenance of early records written in the Brāhmī script in the Tamil country was not known till the beginning of the present century when the first Brāhmī inscription was found in 1903 in the cavern at Kīḷavaḷavu in the Madurai District and copied by the officers of the Epigrahpical Branch of the Archaeological Survey of India. But the importance of the record was not realised immediately. Three years later, in 1906, L. A. Cammiade, a Civil Servant in the District of Tirunelveli, reported the discovery of another cavern with Brāhmī inscriptions at Marugāḷtalai. This prompted the making of vigorous attempts to trace similar caverns with Brāhmī inscriptions in other parts of the Tamil country; and it bore ample fruit. The majority of the short Brāhmī inscriptions under study in this book were discovered in the southern districts of the Tamil country in the next five or six years by the officers of the Epigraphical Branch of the Archaeological Survey of India.

c12.3

The discovery of these records created for them a difficult problem, namely their decipherment and interpretation. The script in which these label inscriptions were incised had been known to them for long, and as such there was not much difficulty for them in transcribing a major portion of them in the modern script. There were, however, a few symbols peculiar to these records neither seen in the Brāhmī inscriptions found in other parts of the country, nor included in the standard script with its value definitely settled. Copies of the records were also sent to scholars outside for study and interpretation. H. Krishna Sastri even thought it worth drawing public attention to these records by publishing photographic copies of the ink-impressions of the inscriptions in the Annual Reports on South Indian Epigraphy for the years 1912, 1915 and 1918. In order to stimulate fresh inquiry and study in this direction he induced Rama Prasad Chandra, then Honorary Secretary of the Varendra Research Society "to take up these Brāhmī documents of South India for study and make an attempt, however slight it may be, to interpret them."

c12.4

These attempts did not produce any result worth mentioning. "Careful readings directly from stone and from estampages were prepared on the spot of almost all such records as had been collected by the department." Finally Krishna Sastri himself gave his own readings of these short epigraphs in a paper which he contributed to the first All-India Oriental Conference held at Poona in 1919, which he concluded with the request that these inscriptions might be taken up for earnest study and interpretation. In his readings of these inscriptions he assigned to the new symbols found in them, the values of the signs very nearly resembling them in the standard Brāhmī or their variations employed in inscriptions in other parts of the country, drawing at the same time pointed attention to them either in the body of his paper or in foot-notes. He also induced K. V. Subrahmanya Ayyar, another senior member of his Department to take up the study of these inscriptions.

c12.5

The long experience of K. V. Subrahmanya Ayyar in the Department, particularly in handling the Tamil records as also their preliminary study by H. Krishna Sastri, were of immense help to him in his examination of these records. According to him these inscriptions followed the Brāhmī script employed in the records recovered from Bhaṭṭiprolu. With commendable ingenuity he was able, apparently after great labour, to fix the values of the new symbols which had baffled scholars previously as those of the letters peculiar to the Tamil language and he was the first scholar to recognise that the inscriptions were written in the Tamil language. He read and interpreted the records as written in the Tamil language and presented to the Third All-India Oriental Conference held at Madras in 1924 a paper embodying the results of his study.

c12.6

But his conclusions were not fully endorsed by scholars. Some were not agreeable to or rather convinced of the possibility of these records being written in the Tamil language. Their main objection appeared to be that while all the inscriptions of the period written in the Brāhmī script and found in other parts of the country were composed in the Prākṛt language, how could it be possible for these records alone to be composed in Tamil? Further the Tamil language employed for writing these records as read, explained and interpreted, did not have much resemblance to standard diction and could be understood only with some emendations. The idea that these inscriptions also must have been written only in the Prākṛt language was strong in the minds of some scholars. C. Narayana Rao took their language definitely as Prākṛt, restored them in Sanskrit and endeavoured to explain them in a paper presented to the Tenth All-India Oriental Conference held at Tirupati in 1940 and wrote a paper on the same in the first number of the now defunct *New Indian Antiquary*.¹

c12.7

¹ pp. 362-76.

c12.8

However, it must be noted here that the reading, restoration in Sanskrit and interpretation of the inscriptions by Narayana Rao are not at all convincing, seem to be far fetched and in very many cases are not applicable to the area, period and the people who were supposed to be the authors of these inscriptions, or for whom they were intended. Further, he had taken only the tentative readings of H. Krishna Sastri as the basis, and treated them as final for his interpretation, ignoring the hesitancy and warnings of the original reader of the inscriptions. Further, the Prākṛt language envisaged by him as used in these epigraphs is quite different from the inscriptional Prākṛt of the period and not found anywhere else.

c12.9

In spite of the distinct contribution made by K. V. Subrahmanya Ayyar to a study of these enigmatic inscriptions it should be said that his interpretation of them is defective for two reasons:

c12.10

(i) Even though he has traced in these records the influence of the scheme of writing adopted in the inscriptions of Bhaṭṭiprōlu he has not used that system in fixing the values of individual letters. It is true that this system does not help in reading all the records. Some of them could be satisfactorily read only if the standard values of the Brāhmī symbols are applied to them and are unintelligible if the values of the Bhaṭṭiprōlu or Drāviḍī type are assigned, while the others will become clear only when they are read in the Drāviḍi style. This will necessarily show that these records are of two different types and probably belong to two different periods chronologically. However, when he studied these records it was generally held, mainly on palaeographical evidence, comparing the forms occurring in these with those in other Brāhmī inscriptions, particularly from North India, that they belonged to the third century B.C., the period of Aśoka. The inscribed potsherds recovered from Arikamēḍu near Pondicherry in the course of the excavations at the place in 1945 and from Uraiyūr in Tiruchirapalli in the course of the excavations at the place in 1965 both assigned to the first century A.D. or thereabouts from reliable evidence were obtained only longafter. Similarly the inscriptions in the Brāhmī script, found at Pugaḻiyūr in the Tiruchirapalli District and Māmaṇḍūr in the North Arcot District and closely resembling those of the inscribed potsherds from Arikamēḍu were also found and copied only later. These discoveries would clearly show that all the Brāhmī inscriptions recovered in the Tamil country so far do not belong to the same period and that there were differences among them with regard to their chronology. Further it is possible to envisage an evolution in the method of writing in the Brāhmī script in the Tamil country, some of the records adopting the Drāviḍi style and some others not. If such a hypothesis is accepted, the reading and interpretation of the records would have become much easier and more acceptable.

c12.11

(ii) K. V. Subrahmanya Ayyar has not also pressed into service the knowledge derived from a study of the classical Tamil literature. Any study of the epigraphical material should utilise and not ignore the early literature in the language in which such epigraphical materials are found. Some of the expressions found in the Brāhmī records in the Tamil country are better understood and explained by reference to classical Tamil literature.

c12.12

At this stage it is necessary to know in detail the special features of the script used in the inscriptions found at Bhaṭṭiprolu or the Drāviḍī style of the southern Brāhmī as it is otherwise called and their bearings on the Brāhmī inscriptions in the Tamil country.

c12.13

In this connection two important questions may be considered viz., the language of these records and the object and nature of their contents.

c12.14

(i) All these records are found in natural caverns high up on the summit or the slopes of hills made suitable for human occupation. They are incised either on the beds in the caverns or on the brows of the overhanging boulders or walls, and are mostly of single or double lines consisting of a few letters. The situation and length of these records would indicate that they are either dedicatory or donative. We have the records of Daśaratha², the grandson and successor of the Mauryan Emperor Aśoka, dedicating the caverns excavated in the Nāgārjunī hill in Bihar to the monks of the Ājīvika sect for use during rainy season. Similar records are found also in other parts of India. On the northern border of the Tamil country there is a record at Mālakoṇḍa in the Nellore District in Andhra Pradesh.³ In Ceylon are found a large number of similar records,⁴ some of which have been edited in the volumes of the *Epigraphia Zeylanica*. All these Ceylonese records mention the dedication or donation of the cave to the monks of religious sects for their use during the rainy season and for doing meditation or penance. The Brāhmī inscriptions of the Tamil country also should likewise be taken as mentioning the dedication of these caverns and beds for the use of religious monks and their penance.

c12.15

² *Ind. Ant.*, Vol. XX, pp. 364; *Ep. Ind.*, Vol. X, Luders' List. Nos. 954-956.

c12.16

³ This is given as an appendix at the end.

c12.17

⁴ See *Ceylon Historical Journal*, Vol. I, pp. 78-96.

c12.18

In addition to the followers of the Vedic religion, followers of other religions such as Buddhism and Jainism, as also the Ājīvikas had come to the south even before the beginning of the Christian era and had settled in South India obviously with substantial following among the masses. One of the *Tēvaram* hymns derides for instance the heretics of those days as *purṟ-ēri-y-uṇaṅkuvar*,⁵ those who (suffer pain and) wither away, climbing the natural crevices (high up on the hill). This would imply that those who lived in these caverns were not the followers of the Vedic religion but those who professed the heretical faiths, Buddhism, Jainism or the Ājīvika one.

c12.19

⁵ Sambandar, *Tēvāram, Tirumuṟai* 3, *Padikam* 322, v. 10.

c12.20

It is not clearly known if these labels are dedicatory or donative in nature, whether they mention the donor or the donee. It is usual that in such inscriptions or labels, the names of the persons who donated or dedicated the object are given. The names of the donees, whether those of individuals or religious sects or even of some gods, are at times given, but not always. In such cases the names of the donees alone would not have been given, without those of the donors. It is the common tendency among donors to take some pride and satisfaction in making such offerings and inscribing their names to show to the world that they were responsible for them. In most of these Brāhmī labels there are found some personal names. These names may be those of either the donors or the donees, very likely those of the former, as many of them contain only single names. Therefore it may not be wrong if it is presumed that these labels mostly contain the names of the donors or persons responsible for the donation of the caves, sometimes accompanied by the names of the persons to whom they were made. But it is to be understood that the donors of these caverns to the teachers of one religious faith need not invariably be the followers of that faith. It is possible that the followers of one faith could have made donation to the followers of another faith for some reason or other. That would show the popularity of the teachers and the esteem in which they were held by the people. In Ceylon, for instance, there are instances of Brahmans making such donations to the Buddhist organisation.⁶

c12.21

⁶ *Ep. Zey.*, Vol. I, p. 18.

c12.22

(ii) The language of these records has been a puzzle to scholars from the very beginning. The official view of the Archaeological Survey of India has been summed up in a way in the following words :

c12.23

PALAEOGRAPHICAL CHART OF SELECT LETTERS FROM EARLY SOUTH INDIAN INSCRIPTIONS

c12.24

Fig. 12. Palaeographical chart of selected letters from early South Indian inscriptions.

c12.25

"The exact nature of the language of these inscriptions is still open to question, but they appear to be in early Tamil (as distinguished from the Tamil found in early Tamil literature, as well as modern Tamil) with a sprinkling of Prākṛit."⁷

c12.26

⁷ *Ancient India*, No. 2. July 1946, p. 109. Even now it is thought that the Brāhmī records have not been satisfactorily interpreted and there is nothing definite about their language. For instance, P. B. Desai writes as follows:

c12.27

It is true that all the early inscriptions recovered in the country and written in the Brāhmī script are in the Prākṛt language. So far nearly a thousand such inscriptions, not taking into account the edicts of Aśoka, have been ably and critically edited by competent scholars in various publications. But the South Indian Brāhmī inscriptions have baffled the attempts of scholars in Prākṛt with regard to their interpretation. To determine the language of these enigmatic inscriptions one has to seek a clue from the records themselves. An analysis of the Brāhmī script used in writing these records would reveal the following peculiarities:

c12.28

(i) Of the hard consonants *ka*, *ca*, *ṭa*, *ta* and *pa*, only the first or primary letters are found in the Tamil alphabet, i.e., the *varga prathamas* of the Sanskrit alphabet are found in them.

c12.29

(ii) The soft varieties or the third letters i.e. the *varga tritīyas* of the above consonants in the Sanskrit are absent.

c12.30

(iii) Of the aspirated varieties, i.e., the second and the fourth letters of these consonants, the letter *tha* is found used sparingly in two or three places while *dha* may also be taken as having been used in one place (Ariṭṭāpaṭṭi A). The others are not met with.

c12.31

(iv) *Śa* and *ṣa* are also not found. Only *sa* is occasionally met with.

c12.32

"These records have still remained an unsolved riddle, though they were discovered over four decades ago. The contributory factors for this state of affairs are these: peculiar forms of the Brāhmi alphabet, crude and archaic nature of the language which may be an undeveloped variety of Prakṛit (?), difficulty in the grouping of syllables and words, prepossession in regard to their Buddhist origin to the exclusion of their Jaina relationship." (*Jainism in South India and Some Jaina Epigraphs*, p. 29, *n.* 3.)

c12.33

(v) No ligatures or conjunct consonants are met with.

c12.34

(vi) Among the vowels only *au* is not found;⁸ similarly the vowels *ṛ* and *ḷ* of the Sanskrit alphabet as also the *anusvāra aṁ* and the *visarga aḥ* are absent.

c12.35

⁸ The previous scholars who have studied these inscriptions have not recognised the primary vowel *ai* in them, and so included the vowel *ai* as one of the letters absent in them. But this is found in the inscription at Ānaimalai. (See the word *aiḷa*).

c12.36

So far as the vowels *e* and *o* are concerned only one symbol without any distinction for the short and the long, is found, used both among the vowels and combined consonants. K. V. Subrahmanya Ayyar says: "In the case of the combined consonants the occurrence of short *e* and short *o* deserves special attention, the two being the special characteristics of the Dravidian alphabets".⁹ But it has not been possible to make out from the available photographs this difference between the short and long in the vowels *e* and *o*, both as primary vowels and as combined with consonants. It is also to be noted in this connection that this peculiarity is found in the Tamil inscriptions in all periods.

c12.37

⁹ *Proceedings and Transactions of the Third Oriental Conference*, Madras, 1924, p. 283.

c12.38

While explaining the Drāviḍi style of writing found in the records at Bhaṭṭiprolu reference has been made to the *sūtra* in the *Tolkāppiyam*, which says that in writing, the pure consonants will have a dot added to it. The next *sūtra* mentions that "*e* and *o* are also of the same nature".¹⁰ An additional dot to denote the short in the vowels *e* and *o*, both primary as well as secondary, is found in Tamil epigraphy on very rare occasions¹¹ and is more conspicuous by its absence.

c12.39

¹⁰ *Tolkāppiam, Eḷuttadikāram, Sūtra* 16; see *ante* p. 117.

c12.40

¹¹ See the Veḷvikkuḍi plates of Parāntaka Neduñjaḍaiyan (*Ep. Ind.*, XVII, pp. 291-92.)

c12.41

The above analysis would show that the language employed in these records had only the *varga prathamas* with no sonants or other aspirated forms. This peculiarity together with the absence of *śa*, *ṣa*, *ha* etc. would indicate that the alphabets of the language closely followed the pattern of the classical Dravidian form or Tamil.

c12.42

It has been pointed out earlier that some new symbols are also found used in these records. The very fact that these symbols are not found in the Aśokan inscriptions would clearly indicate that they represented alien sounds not found in the standard Brāhmi script and thus would belong to a language quite different from that of the Aśokan records. The vowels assigned to these symbols would certainly be an important factor in deciding the language of these inscriptions as also in their interpretation. Such new symbols found in these rocords are :

c12.43

(a) •|• (b) 𑀮 (c) 𑀵 (d) 𑀯𑀴 (e) 𑀴 (f) 𑀷

c12.44

(a) The first symbol •|• which occurs ten times in the Brāhmī inscriptions of the Tamil country under consideration is not seen in the Brāhmī inscriptions of North India prior to the commencement of the Christian era. However, it is found in the Kṣatrapa, Sātavāhana and Gupta inscriptions and given the value of long *ī*. This symbol is found in the Tamil inscriptions of the eighth century and given the same value. It also occurs in the Brāhmī inscriptions of Ceylon and has been read as short *i*. In the Brāhmī inscriptions of the Tamil country also the symbol is read as short *i* as in the Brāhmī inscriptions from Ceylon.

c12.45

In the inscriptions of Aśoka the vowel *i* short is indicated by three dots arranged in the form of a triangle ( ∴ ) and long *ī* by four dots arranged in the form of a square ( :: ). The latter form found in the Brāhmī records of the Tamil country are taken as short *i* by K. V. Subrahmanya Ayyar. This difference in the evaluation of the vowels *i* short and long proposed by him is interesting. But this position did not last long, for the original position was restored later. It is however difficult to explain why this difference was short lived. The formation of the symbol of a vertical line with two dots, one on either side, indicating the vowel *ī* long may be explained as follows. From the *sūtra* in the *Tolkappiyam* it become clear that in the Tamil country, the addition of a dot to any letter or symbol indicated the dimunition of the sound, reducing the *mātra* or the sound length to one half of its original length.¹² According to the above rule, the letter written in the shape of four dots instead of denoting the long *ī* would indicate the form of the short vowel *i*. Very likely it was considered necessary to make some change in the form of long *ī*. This was effected by joining the two diagonally opposite dots, thus making them a line, leaving the other two dots, one on either side.

c12.46

¹² See M. Raghava Ayyangar, article "*Tolkappiyanarum Puḷḷi-y-eḻuttuk-kaḷum*" in his collection of Essays, *Ārāyccit-tokuṭi*.

c12.47

(b) The second symbol Ր which looks like the ‘inverted J’ occurs frequently, more than sixty times, in our inscriptions. It occurs sparingly in the inscriptions at Bhaṭṭiprolu and has been assigned by Bühler the value of *na* long. Though its sound value may not be quite different from that of the same symbol in the Bhaṭṭiprolu inscriptions, it has to be given a different value here. This is evident from the fact that in a particular inscription, viz. the Śittaṇṇavāśal record the Ր symbol is used along with the ordinary Aśokan *na* form. As such this symbol has to be taken to be different from the ordinary *na*, but denoting the same sound value. The peculiar alphabet *na* of the Tamil language answers to the description. Hence it has to be taken to represent the letter *ṉ* which occurs generally at the end of some words. In some cases the symbol has to be treated as a basic consonant found at the end of some words and indicating third person masculine singular, and in others as *ṉa* as in the case of other letters.

c12.48

(c) The third symbol Ⴈ, however, occurs rarely in the Brāhmī inscriptions in the Tamil country eg. Tirupparaṅkuṉṟam and Ariṭṭāpaṭṭi. Krishna Sastri writes about it as follows: “The formation of this letter is very peculiar. Its similarity with Khalsi Aśokan *Ja* given in Bühler's *Tafel* II—15, 2 is very slight. One can venture to say that it is somewhat like the modern Tamil *ḻa*”.¹³ K. V. Subrahmanya Ayyar also is of the same view.¹⁴ Their suggestion is acceptable. In fact no other value can be assigned to the letter.

c12.49

¹³ *Proceedings and Transactions of the First Oriental Conference*, pp. 366, *n*. 7.

c12.50

¹⁴ *Proceedings and Transactions of the Third Oriental Conference*, p. 285.

c12.51

(d) The fourth symbol which also occurs many times in these records resembles very much the Brāhmī letter *la* with a slight difference, namely the addition of a hook line stroke to the right arm. It is suggested that it may be taken to represent the lingual *ḷa*, a distinct feature of the Dravidian languages. Though the symbol *ḷa* occurs in the Bhaṭṭiprōlu script also, its form is slightly different there.¹⁵

c12.52

¹⁵ See *Ep. Ind.*, Vol. II, p. 324.

c12.53

(e) The fifth symbol is a very curious one. It looks very much like a badly shaped *ṭu*. It occurs once in each of the Āṉaimalai and Śittannavāśal records, twice in one of the Ariṭṭāpaṭṭi records and four times in another (record I) from the same place. Krishna Sastri says about them: "These are letters the formation of which is not found elsewhere. I have suggested the readings *ḍu* and *ḍai* on the strength of the remarks on the letter *ḍu* made by Dr. Bühler, on the palaeography of the Bhaṭṭiprōlu inscriptions.¹⁶ Still the letters *ḍu* and *ḍai* are not certain. They may be conjunct consonants as well".¹⁷ K. V. Subrahmanya Ayyar takes it to be *ṟa* and says: "the only letter that is allied in sound to *ṭa* which has no equivalent in Sanskrit and for which a new symbol had to be devised in the Aśokan code is the Dravidian *ṟa*. The difficulty of pronunciation of this letter and its kinship with *ṭa*, *ḍa* and sometimes with *ta* by which letters it is often substituted not only in speech forms but in allied languages will be evident to any observer. (*Cf.* Tamil *kuṉṟu*, Telugu *koṇḍa*; Tamil *naṉṟu*, Telugu *naṇḍu*......This likeness in sound has to account for the devised form given to it in our records ... ... ... Even in the Vaṭṭeḷuttu script, its formation is such as to be mistaken for *ta* (with but slight difference). What proves conclusively the correctness of the value suggested here is the occurrence of it after *ṉ* in the Āṉaimalai record in the word *kuṉṟa*".¹⁸ Thus the symbol may be taken to denote *ṟa*.

c12.54

¹⁶ *Ibid.*, pp. 323 ff.

c12.55

¹⁷ *Proceedings of the First All-India Oriental Conference*, p. 324, *n*. 3.

c12.56

¹⁸ *Proceedings of the Third All-India Oriental Conference*, p. 285.

c12.57

(f) This symbol is found in the Brāhmī inscriptions at Pugaḷiyūr in the Tiruchirapalli District. Comparatively it is later in date than that found in the Madurai, Ramanathapuram and Tirunelveli Districts, and very much resembles the letter found on the potsherds unearthed in the excavations at Arikamēḍu near Pondicherry. The symbol found here looks very much like the letter *ṇa* of the Vaṭṭeḷuttu script. Along side of this symbol we find also in the inscriptions at the place the symbol 𑀡 indicating the value of cerebral *ṇa* as also a few others. But the form is slightly different. The upper part of the symbol instead of being a vertical line with a curve to the right at the top, lies low very much like a semicircle, with the right end of the curve reaching the level of the base. There are also one or two other forms which may be considered as intermediaries between them. Only the value of cerebral *ṇa* suits the places where they occur. Evidently these forms represent the different stages in the evolution of the form of the letter. All of them have been assigned the value of *ṇa*.¹⁹

c12.58

¹⁹ See T. N. Subrahmanyan, *Temple Inscriptions*, Vol. III, pt. i pp. 1509—11; also his *Paṇḍai Tamiḻ Eḻuttukkaḷ* (in Tamil) pp. 21—22.

c12.59

It is to be noted in this connection that the shape of the hard *ṟa* employed in the Brāhmī inscriptions at the place, though retaining the same form has undergone some change. The upper part of it in the records of the Pāṇḍya country appears like the Brāhmī *ṭa*, while the lower part consists of two vertical lines like the symbol of the medial *u* added to it, so that it may look different from *ṭu*. But here the upper part has been reduced to a mere hook, the lower part appearing like a mere *ta*. But still a comparison of these forms will show that the latter form was evolved out of the former.

c12.60

The Brāhmī of these inscriptions, even though following very closely the Brāhmī of the Aśokan records, yet contains a few peculiarities and deviations from the latter. These may be stated briefly as follows :—

c12.61

*Vowels* :

c12.62

In the Aśokan inscriptions, the letter *a* appeared like an 'inverted k' with the two arms joined. Even that form is found in these records also in a majority of cases. The letter has both the arms as two seperate angular lines diverging in opposite directions, above and below Ɥ. The symbol for the vowel *ai* has not been recognised in these inscriptions by the scholars who have studied them till now. The letter *ai* is written in the northern inscriptions as a modified form of *e* *i. e.*, a horizontal line to the left at the top of a triangle with its apex above and base at the bottom, 𑀏, or in other words, with the addition of the symbol for medial vowel *e*, to the symbol of the primary vowel *e*. This is in conformity with the concept of the Sanskrit grammarians deriving the *vṛddhi* vowel *ai* from the *guṇa* vowel *e*. But the form of *ai* found in these records (in the one at Aḻagarmalai, in the combination of *Aiḷa*) is in the shape of a trident, similar to that of the Vaṭṭeḻuttu script of later days; and this same form is found in the Allūru inscription²⁰ though standing a little bit to the left. The word *Aiḷa* is found also in the Hāthigumphā inscription of the Kaliṅga king Khāravela.²¹ Here it may be taken to signify a Pāṇḍya, as the later members of the family claim descent from *Candra* (Moon) through his son *Budha* (Mercury) and the latter’s son *Iḷa*. This trident form in the south appears to be connected with the form of the semi-vowel *ya*.

c12.63

²⁰ *A. R. S. I. E.*, 331 of 1924; *Rep.*, p. 97 and plate facing it.

c12.64

²¹ *Ep. Ind.*, Vol. XX, pp. 80 and 86.

c12.65

The necessity for designing and introducing a new form for a letter which already possessed another form in the Brāhmī script would clearly show that the existing form was not suitable for indicating the intended letter while writing in the southern part of the country; evidently the existing form indicated in the south another letter or it was very similar to the form of another letter so as to cause confusion. Such a letter could have been the vowel *e* either long or short, depending on the value assigned to the letter having the triangular shape. Considering that in the case of the vowels *a* and *u*, the standard shape indicated the short sound, the long one being indicated by the addition of a symbol, we may consider that the case of *e* also was similar. Very likely the triangular shape might have originally denoted at least in the south the short *e*, the long vowel *ē* being written by the addition of another symbol. Perhaps this explains the need for designing a new form for the vowel *ai*.

c12.66

*Consonants:*

c12.67

In the Aśokan inscriptions the letter *ma* is written as a circle with two perpendicular arms above. But in these records it is written in the form of a broad cylindrical tube opening up with a cross-bar in the middle. However, the shape of the letter found in the Māmaṇḍūr inscription is somewhat different. It appears more angular there and the two perpendicular lines at the top appear instead as two slanting lines converging towards a point below and joining together. This form looks very much nearer the shape of the letter in the Grantha script as prevalent in the Tamil country in the seventh and eighth centuries A.D.

c12.68

The letter *ya*, though written with three arms, is found in the northern Brāhmī with the middle arm too long like an anchor. Though similar forms are found in these records also, they are very rare. The most common shape is to have all the three arms of equal length.

c12.69

*Medial vowels:*

c12.70

It has been noticed earlier that in northern Brāhmī the long *ā* is indicated by the addition of a horizontal stroke at the right to the top of the character, and that in Drāviḍī the letter denoted the consonant with the inherent *a* short, instead of *a* long. Sometimes we find here also two such strokes to indicate *ā* long.

c12.71

Medial *ī* is generally denoted by the addition of two straight lines, one horizontal and the other vertical, opening upwards at the top of the right; sometimes the angle is found smoothened, thus forming a sort of curve.

c12.72

Medial *o* is formed by a line at the top of the character projecting both to the left and to the right; sometimes it is found as two separate strokes, one on the right and the other on the left.

c12.73

One other peculiarity in writing vowelled consonants is observed in these inscriptions. Instead of adding the symbol for the medial vowel to the consonant, sometimes the consonant and the vowel are written side by side. For instance, in the inscription at Kongarpuḷiyankuḷam, for denoting the letter *ṉu* in the word Pākaṉūr, the name of a village, the symbols for the consonant *ṉa* and the vowel *u* (both of which make the word *ṉu*) are written side by side. Similarly in the record at Śittannavāśal the letter *ṭa* in the word *atiṭanam* is represented by two symbols, the consonant *ṭa* and the vowel *a* written side by side. Likewise in the inscription at Karuṅgālakkuḍi the letter *yū* in the word Eṭhuyūr is denoted by two symbols, the consonant *ya* and the vowel *ū*, written side by side. These examples will show clearly that in these Brāhmī records, the consonant letter without the addition of any vowel sign is to be treated as a vowelless or pure basic consonant (without the inherent *a*).

c12.74

It is very likely that this method of writing was adopted not out of ignorance of the rules of the formation of the vowelled consonants, but in consonance with some definite principles followed in writing in those days. The expression *ūr* which comes at the end of words like Pākaṉūr and Eṭhuyūr clearly indicates them as the names of villages. They were probably named after persons like Pākaṉ and Eṭhūy. Hence the above form was adopted to indicate that such villages were named after certain persons.

c12.75

In other respects, the Brāhmī of these records follows the standard pattern of the Aśokan inscriptions.

c12.76

The new symbols invented both for new sounds as also for a few letters for which symbols already existed in north Indian Brāhmī, but which were found unsuitable for use in the records, would clearly indicate that the language employed in them was not Prākṛt, but Tamil, the language of the people of the region where these inscriptions are found.

c12.77

*Language:*

c12.78

This view is further strengthened by a study of the words employed in these records.

c12.79

The real import of a few of the words that frequently occur in these records but have not been properly realised and interpreted so far by the scholars who have studied them may be considered first :

c12.80

Among them are *ataṉa*, *aṉatai* and *aritaṉ* :

c12.81

*Ataṉa* or *atāṉa*, *aṉatai* and *anitaṉa*, *aritaṉa* or *aritiṉa* are some of the words which are found largely used in these inscriptions. Judging from the context in which they are found, it would appear that they are the proper names of persons. Then the letter *ṉa* at the end of the words *ataṉa* and *aritaṉa* have to be treated as the basic consonant *ṉ* indicating the masculine singular suffix.

c12.82

Of the first word the most common form is *ataṇa* and occasionally *atana*, without the medial *ā* long for the letter *ta*. Adopting the system followed in the case of records at Bhaṭṭiprōlu described earlier, the consonant letters without the addition of any symbol represent the basic or pure (vowelless) consonants, those with the symbol of *ā* long represent the consonant with the inherent *a* short; and to denote *ā* long two such symbols are added. According to this principle the word has to be read as *ataṉ*. This looks very similar to the word *Ātaṉ* found in early classical Tamil literature as part of personal names, *Cf.* the Cēra ruler *Cēral Ātaṉ*. The initial *ā* is shortened instead of being lengthened.

c12.83

The second word written as *aṇatai* has to be read as *antai*; and this in turn looks very similar to the word *āntai* (with the initial long vowel shortened) occurring likewise in classical Tamil literature. *Cf.* the poet *Picir Āntaiyār*, whose compositions are included in the collection *Ahananūru*.

c12.84

The third word has to be read as *aritan* or *aritin*. This may be compared with the word *Āritaṉ* of the name of the Tamil grammarian *Aiyaṉ-Āritaṉ-ār*, author of the treatise *Puṟapporuḷveṇpamālai*.

c12.85

It may be taken that these three names are really *Ātaṉ*, *Āntai* and *Āritaṉ*, with the initial vowel shortened in all of them.

c12.86

*Āritaṉ* and *Āritiṉ* are the Tamil forms of the names Hāritā and Hāriti respectively. The name Hārītiputra is mentioned in many of the Prākṛt inscriptions of the early centuries of the Christian era found in Karnāṭaka and western India. The kings of the Kadamba dynasty who ruled from Vanavāsi (Banavase) claim to be *Hārītiputras* of the *Mānavyagotra*. The word *Aritaṉ* or *Āritaṉ* of these inscriptions should be taken as indicating that the bearer of that name was a member of the Hārita family.

c12.87

There is a specific *Sūtra* in the Tolkāppiyam, which states that the Tamil word *antai* is formed by the combination of two words *ataṉ* and *tantai* and that another word *pūntai* is formed similarly by the combination of the words *Pūtan* (*Bhūtan*) and *tantai*. This is the middle of three *sūtras* explaining the rules of *sandhi* when the word *tantai* having the meaning of father is added to the proper names of persons. This is a general rule. The commentators point out the following as examples: *Cāttan* + *tantai* = *Cāttantai*; *Kīran* + *tantai* = *Kīrantai*; *Koran* + *tantai* = *Korantai*. The suffix *an* of the standing personal name and the initial consonant *t* (retaining the medial vowel *a*) of the joining word *tantai* are dropped. The *sūtra* after that under discussion gives the special rule to be observed when the standing personal name happens to be *Ātan* and *Pūtan*. In addition to the changes mentioned in the preceding *sūtra*, the final consonant and the initial vowel of the standing word and the coming word respectively (remaining after effecting the above change) are also dropped. Thus in the word *ātan* the suffix *an* is dropped as per the preceding *sūtra*, and of the remaining *āt*, the final consonant *t* also is now dropped, thus leaving only the initial vowel *ā*. Similarly in the word *tantai*, the initial consonant *t* is dropped according to the preceding *sūtra* and of the remaining *antai* the initial vowel is now dropped, thus leaving only the final *ntai*. Thus combining with *ā*, the residue of the standing word, it becomes *āntai*. The third *sūtra* gives the exemptions, *i. e.*, the cases wherein the above rules will not apply. Accordingly if such proper names are preceded by adjectives, there is no change (*i. e.*, no dropping of letters). If the standing word is *Peruñcāttan*, *i. e.*, *Cāttan* with the adjective *Perum*, signifying great, there will be no dropping of letters when *tantai* is added to it. Thus *Peruñcāttan* + *tantai* will be *Peruñcāttan-tantai*.²⁶

c12.88

²⁶ *Iyarpeyar munnart tantai muṟaivarin* *mutarkaṇ meykeṭa vakara nilaiyu* *meyyoṭit tankeṭu mavviyar peyarē.* (348) *Ātanum Pūtanuṅ kūṟiya viyalpoṭu* *peyaroṟ ṟakaran tuvarak keṭumē.* (349) *Ciṟappoṭu varuvali yiyaṟkai yākum.* (350) (Tolkāppiyam, Eḷuttadikāram, Ed. by P. S. Subrahmanya Sastri.) There are also some more words in Tamil which have been formed likewise in combination with *tantai* eg. *entai* made up of *en* and *tantai* meaning 'my father', *nuntai* made up of *nun* and *tantai* meaning 'your father,' etc. These two words relate to the first and second persons of the grammatical formula. Similar combinations with the words *tāy* (mother) and *tampi* (younger brother) also occur for the first and second persons; *empi* made up of *eṉ* and *tampi* or *ampi* (my brother); *numpi* made up of *nuṉ* and *tampi* or *ampi* (your younger brother); also *yāy* (my mother) made up of *ed* and *āy*; *ñāy* (your mother) made of *nuṉ* and *āy*.

c12.89

The above *sūtras* would thus show that in those early days, many names in which the word *tantai* forms an integral part were in common use among the people of the Tamil country. That such was the position will become evident from the following names borne by some of the poets and others as gleaned from the Śaṅgam classics; Cāttantai, Kīrantai, Āntai, Pūntai, Koṟṟantai, Piṭantai, Eiṉantai, Vaṭukantai etc. On the analogy of *entai* (made up of two words *eṉ* and *tantai*) meaning ‘my father’ and *nuntai* (made up of two words *nuṉ* and *tantai*) meaning ‘your father’ the above words are taken as indicating the fathers respectively of Cāttaṉ, Kīraṉ, Ātaṉ, Pūtaṉ, Koṟṟaṉ, Piṭaṉ, Eyiṉaṉ and Vaḍukaṉ. Thus the word Āntai has been taken as indicating the father of Ātaṉ.

c12.90

But it would appear rather strange that persons bearing these names, themselves great poets, instead of being known by their own personal names, should be known as the father of their sons about whom nothing is known. It is usual to indicate the name of a person as “so and so the son of so and so”. This practice also appears to have been in vogue in those days as will be evident from the names *Maduraikkarikkavaṉarmagaṉār Nakkīraṉar*, *Ollaiyūr Kiḻārmagaṉ Peruñcattaṉ*, *Kaṭṭūr Kiḻārmagaṉār Kaṇṇaṉār* etc., occurring in the Śaṅgam classics. It might be possible to conceive of stray cases where the father of a famous son is referred to as “the father of a well-known son”. But the above cases do not appear to be stray ones. On the other hand these appear to be common as a class. Among the persons having the name Āntai or Āntaiyār we have Picir Āntaiyār, (author of *Naṟṟiṇai* 91) Ōtal Āntaiyār (author of the hundred stanzas, 301-400 of *Palai* in

c12.91

*yāyum jñāyum yār ākiyaro?*
*entaiyum nuntaiyum emmuṣaik kēḷir?*
*yānum nīyum ev=vaḻi aṟitum?*
*Kuruntogai*, 40.

c12.92

Such combinations do not, however, occur for the third person. The words *tāy* and *tampi* are generally taken as signifying such relations with reference to the third person, thus obviating the necessity of using any special word for that. (*taṉ* *āy* and *taṉ* *tampi*).

c12.93

*Ainkuṟunūṟu*) Añcil Āntaiyār (author of *Kuṟuntokai* 294 and *Narṟiṇai* 233) Cirakkuṭi Āntaiyār, (author of *Kuṟuntokai* 56, 57, 62, 133, 168, 222, 273, 300 and *Narṟiṇai* 16 etc. It is inconceivable that these poets were known as fathers of sons bearing the name of *Ātaṉ*, who are otherwise unknown.

c12.94

This difficulty appears to have been felt by some scholars even earlier.²⁷ They have therefore tried to explain the construction *e.g.* of the word *Āntai* made up of the two words *Ātaṉ* and *tantai* as referring to "a person who has *Ātaṉ* as his father", that is referring to the son of *Ātaṉ*. But rules of grammer do not admit of such a construction. It has therefore to be given up.²⁸ Further it would have been much easier and more normal to refer to a person as 'the son of so and so' instead of in a round about way as 'the person having so and so as his father'.

c12.95

²⁷ *Śentamil*, Vol. xix, pp. 281 ff.

c12.96

²⁸ *Tamil Poḻil* Vol. x, pp. 33, 410.

c12.97

The problem is further complicated by the mention of the names of the wives and sons of such persons. The incident relating to the lady referred to as *Kīrantai maṉaivi* (the wife of Kīrantai) is significant.²⁹ The term *Kīrantai maṉaivi* would mean 'the wife of *Kīrantai*' *i.e.*, the wife of the father of *Kīraṉ*. This is a circuitous way of referring to a person. She could be easily referred to as the mother of *Kīraṉ*.

c12.98

²⁹ See Ch. 29. It is said that a person known as Kīrantai went on a pilgrimage leaving his wife at Madurai, with no one else to look after her, thinking that the protection given by the Pāṇḍya king was enough.

c12.99

The *Tolkāppiyam* gives in continuation of the three *sūtras* mentioned above (codifying the changes that take place in the *sandhi* when the word *tantai*, meaning father, is added to certain proper or personal names) another *sūtra* detailing the changes when the word denoting son is introduced.

c12.100

*Appeyar meyyoliṭ taṉkeṭu vaḷiyu*
*Nirṟalu murittē yammeṉ cāriyai*
*Makkaṇ muṟaitokūu maruṅki ṉāṉa.*

c12.101

(*Tolkāppiyam*, *Sūtra*, 351.)

c12.102

According to it in the expression *Cattaṉ makaṉ korṟaṉ*, *i.e.*, *Korṟaṉ* son of *Cattaṉ*, the word *makaṉ* itself is not used (being only understood) while the suffix *aṉ* of the standing word is dropped and its place taken by the increment *am*. Thus it becomes *Cattaṅ Korṟaṉ*. This is followed by another rule which gives the exceptions in certain cases whether the word in question denotes the father or the son.

c12.103

*tānum pēṉuṅ kōṉu menṉu*
*mā muṟai yiyarpeyar tiripiṭanilavē* (352)

c12.104

According to this *sūtra* there is no dropping off if the words *tāṉ*, *ēṉ* and *kōṉ* are either followed by the word *tantai* or have the word denoting son understood after them *e.g.*, *tāṉ* + *tantai* = *tāṉ-ṟantai*; *pēṉ* + *makaṉ* + *Korṟaṉ* = *pēṉ-koṟṟaṉ*.

c12.105

These would show that just as in the case of *tantai* or father, the word denoting a son was also largely used in the names of persons in those days.

c12.106

The author of *Tiṇai-moḻi-y-aimpatu*, one of the eighteen minor classical poems, known by the collective name of *Padineṇ-kīḻ-kaṇakku* is known as *Cāttantaiyār-makaṉār kīraṉ-koṟṟaṉār* *i.e.*, *Kīraṉ Koṟṟaṉār* son of *Cāttantaiyār*. The name of the poet *Kīraṉ Koṟṟaṉār*, is to be interpreted as *Koṟṟaṉ* son of *Kīraṉ* and who at the same time was the son of *Cāttantai*, the son of *Cāttaṉ*, *i.e.*, the person who was the father of *Cāttaṉ*.

c12.107

*Cāttantaiyār makaṉār Kīraṉ Koṟṟaṉār*
*Kīraṅkoṟṟaṉār* = *Kīraṉ makaṉ Koṟṟaṉār*.

c12.108

The full name will be:

c12.109

*Cāttaṉuḍaiya tantaiyiṉ makaṉārāṉa Kīraṉ makaṉ Koṟṟaṉ*.

c12.110

Such an interpretation does not, however appear to be correct.

c12.111

This very same combination with the substitution of the name *Piṭantai* instead of *Cāttantai* occurs in one of the Brāhmī lables now under discussion and found at Pugaḻiyūr. The actual reading in the inscription is

c12.112

*Piṭantai makaṉ Kīraṉ Koṟṟaṉ*

c12.113

Under the circumstances we have to find some other explanation for similar constructions as also the word *makaṉ*.

c12.114

There are some words in Tamil with the suffix *makan* added to them like *aṇmakan*, *paṇimakan*, *perumakan* etc. In all of them the suffix *makan* does not mean a son. In fact *aṇmakan* means *aṇmai uḍaiyavan* (one who is courageous or brave), and *paṇimakan* means *paṇiśeivan* (one who does service or work). Similar expressions are found in Sanskrit, such as *Vaṇik putra* meaning only a merchant (not a merchant's son).³⁰ In all these cases the word *makan* and *putra* obviously indicated only a spiritual son and not a natural son. It is only by such an interpretation that we can explain the formation of the name *Kīran-Koran* which indicates *Kīran makan Koran i.e.*, *Koran* son or follower of *Kīran* son of *Piṭantai*. It is also to be noted in this connection that the names *Śattan piṭan* (Bhaṭāra) *Kīran* (Kubera, etc.) are all holy ones.

c12.115

³⁰ It may be mentioned in this connection that the purohit Brahmins of Gayā (*Gayāvalis*) are called *Gaṅgāputras* and the Vira Balingas of the Telugu country (*Viravalañjiyar* of the Tamil country) are called *Parvatīputras*.

c12.116

Considering the above it would appear that the word *tantai* should have had some special significance or meaning in addition to its literal meaning, father. The term 'father' denotes five persons, the bodily or physical father, the religious preceptor, the teacher who gives education, king and one who gives food.³¹ There are different tests by which these five persons are distinguished.³² But the religious preceptor in addition to the bodily or physical father is always included in the list and occupies an honoured place. The Sanskrit word *guru* meaning a preceptor or teacher also refers to the bodily father. In the inscriptions of the Western Cāḷukya king Vikramāditya I this word is used in that sense (*svagurośśriyam*: the royalty of his father).³³

c12.117

³¹ *Eg.* The Gadval plates, *E.I.*, Vol. x, p, 103, ll. 13-14.

c12.118

³² Vide the Tokai (தொகை) section of the Tamil *nighaṇṭus*.

c12.119

³³ See *Tamil Lexicon*, S. V. Vol. I Ainkuravar.

c12.120

Evidently the word *tantai* used in these compounds gives out some such meaning as a 'spiritual father'; possibly it denoted a religious teacher in the same way as the word 'Father' is now applied to Roman Catholic Priests or an *acarya* among the Hindus. Therefore the word *antai* (*ātan* + *tanai*) may be possibly taken as meaning 'the Holy Father *Ātan*,' instead of merely as 'the father of *Ātan*.'

c12.121

Then the word *Ātaṉ* which is frequently met with in these inscriptions should be taken as denoting a person of high religious merit respected and venerated by people. The word seems to be derived from the Sanskrit word *āpta* and the Tamil *nighaṇṭus* give it as one of the synonyms of the *arhat*.

c12.122

This word is also found largely used in the *Tēvāram*, when referring to the persons of the heterodox religious orders *i.e.*, persons belonging to religions other than Saivism like the Buddhist, Jaina, Ājīvika orders. This word is explained by modern commentators as a low people (*kīḻmakkaḷ*)33a and as ignorant persons, fools (*aṟivillatavar*)33b

c12.123

33a *Tirujñānasambandar Tēvāram*, First Tirumurai: Ed. by Dharmapuram Adhinam, p. 97.

c12.124

33b Second Tirumurai, II, 107–10, p. 504.

c12.125

If really the word *ataṉ* meant only low people or ignorant fools, then would it have been borne as personal names, even by royal personages: *cf.* Cēral Ātaṉ i.e. Ātaṉ, the Cera king? Further from the context where this word occurs in the *Tēvāram* hymns, it would appear that it should denote one of the many heterodox religious groups then prevalent in the country.

c12.126

It has been pointed out above that the word is derived from *āpta* which is a synonym for *arhat*. The word *arhat* is found in use not only among the Jainas, but also among the Buddhists and the Ājīvikas. Further, from the Tamil poem *Nīlakēśi*,³⁴ it is learnt that the word *āpta* was used by the Ājīvikas, to denote exclusively the founder of that religion (Makkhali Gosāla), *Markali*.³⁵ Even in a late Kannaḍa inscription of the thirteenth century A.D. from Śravaṇa Beḷgoḷa the word *āpta* is used, so as to suggest the Ājīvikas.³⁶

c12.127

³⁴ See *vv*. 670, 671, 683 *etc.* (A. Chakravarti's edn.). See A. L Basham, *History and Doctrines of the Ājīvikas*, pp. 79, 125, 244 and 276.

c12.128

³⁵ Basham thinks *āptan* to be 'a rather unusual title which may have had a specifically Ājīvika connotation'. (*op. cit.*, p 79).

c12.129

³⁶ *Ep. Car.*, Vol. II, (Rev. Edn.) SB No. 234.

c12.130

It is therefore possible that the words *Ātaṉ* and Āntai (Holy Father Ātaṉ) are terms denoting the people of the Ājīvika faith.³⁷

c12.131

³⁷ It is very likely that similar combinations Pūntai, Kirantai, and S'āttantai also mean respectively Holy Father Pūtan (Bhūta), Holy Father Kiraṉ, and Holy Father S'āttaṉ, and that these names related to the various religious faiths then prevalent in the Tamil country. The *S'ilappadikāram* mentions *Paṣaṇḍa S'āttaṉ* (Canto. IX, 11. 12, 15). (cf. also *Jayanti Ramayya Pantulu Commemoration Vlume*, pp. 156.8).

c12.132

*Paḷi* is another word found thrice in these inscriptions, at Kīḻavaḷavu, Karuṅgāḷakkuḍi and Ariṭṭāpaṭṭi. This word has an additional symbol *ī* at its end in the record at Kīḻavaḷavu and thus the reading of the word is *Paḷi ī*, while the record at Ariṭṭāpaṭṭi has the letter *ya* at the end of the word, which therefore reads *pāḷiya*. Subrahmanya Ayyar takes the word in the record at the latter place as *Paḷiy*, with the expletive use of *y* at the end as in accord with the spoken dialect. The suffix *y* is commonly found in several Tamil inscriptions of the early period (up to the 10th century A. D.). But he thinks that the *ī* at the end of the word in the Kīḻavaḷavu inscription is the pronoun denoting 'this'. But this letter *ī* as indicating the pronoun is not found used either in Tamil literature or inscriptions. It is also very doubtful if it was ever used in that way in the other Dravidian languages in early times. Again its occurrence at the very end of the record also shows that the letter could not have been used in that sense, as it is quite against the Tamil or Dravidian syntax. Subrahmanya Ayyar himself is aware of it and so he adds, "if its use at the end of a sentence is objectionable and unlikely it may form part of *pāḷi*". He thinks that the word *paḷi y* is the Tamil word meaning an excavation in stone set apart for the residence of monks and that it is not unlikely that it is connected with the Prākṛt word *paliya* as suggested by Krishna Sastri. He also quotes two extracts from the Tamil *Nighaṇṭu* in support of the meaning he has suggested.³⁸

First 131 of 197 passages. Read on

Linked texts (25)