Pramāṇa · Search · About · API

Elements of South-Indian Palæography, from the Fourth to the Seventeenth Century A.D. c5.s2: § 2. The Grantha-Tamiḷ Alphabets

Elements of South-Indian Palæography, from the Fourth to the Seventeenth Century A.D. c5.s2

Language
English

Contents

§ 2. The Grantha-Tamiḷ Alphabets

A. Cēra

c5.s2.t1.1

The Grantha, Modern Tamiḷ, Malayāḷam and Tuḷu alphabets all have their origin in the Cera character, a variety of the 'Cave character' which was used in the Cera kingdom during the early centuries A. D. From the third to the seventh century appears to have been the most flourishing period in the modern history of this kingdom; it then extended over the present Mysore, Coimbatore, Salem, Toṇḍaināḍu, South Malabar and Cochin. It was, however, one of the three great old Dravidian kingdoms and existed already in the third century B. C. What civilization it had before the period referred to, there is no information; nor is there the least trace as yet of any inscription before the early centuries A. D.¹ The existing inscriptions show that about the fourth or fifth century A. D. the rulers of this kingdom received the Jains with great zeal, and made most liberal endowments to them in the territory that constitutes the modern province of Mysore.

c5.s2.t1.2

¹ The history of the Cera kingdom is excessively obscure, and will, probably, always remain so. Like in most Indian kingdoms that have preserved an existence for several centuries, there were, in all probability, many revolts of feudatories and changes of dynasty; it is thus very little use to accept the "Kŏṅyudeśarājakkaḷ" as an authority, for it bears evident signs of being a very recent compilation from grants and local traditions most clumsily put together. It is translated in the *Madras Journal*, vol. xiv. pp. 1-16. The most important investigation (as yet) respecting the Cera kingdom is by Profr. Dowson (in *Journal of the R. A. S. of London*, vol. viii. and also printed separately).

c5.s2.t1.3

The Cera alphabet changed but little during a considerable time; the earliest and latest authentic inscriptions which are in existence, and which belong to a period of about four centuries, show very few innovations. Two varieties of this character must be distinguished; the first, which was in use in that part of the Cera country which constitutes the modern Mysore and Coorg up to the final end of the kingdom which was conquered by the Coḷas about 877 A. D., and which then fell into disuse being soon supplanted by the Western Cālukya and transitional characters; and the second, which was used in Toṇḍaināḍu (the neighbourhood of Madras) which was a feudatory of the Cera kingdom till about the end of the seventh century when it fell under the Coḷas. This last alphabet then became under the new dynasty the medium of introducing brahmanical culture to the Tamiḷ country².

c5.s2.t1.4

² In the third century B. C., the Açoka Edicts show that Keraḷaputra (*i. e.* the Cera sovereign) was one of the three great powers of the South. Ptolemy (2nd century A. D.) and the Periplus of the Red Sea (3rd century A. D.) prove (§ 34) that this was still the case. According to the former (vii., 1, 86) Karūr was then the capital. Hiouen-Thsang (about 640 A. D.) does not mention this kingdom, but under the name of the kingdom of Kŏṅkaṇapura (the present Kŏṅkaṇa-haḷḷi) he describes a part of it ("Pèlerins Bouddhistes" iii. pp. 146-9). The dynasty which the inscriptions mention extends from

c5.s2.t1.5

The earliest unquestionable inscription as yet known is that of which the alphabet is given in Pl. ii. and which has been published in facsimile in the *Indian Antiquary*¹; the date is about 467 A. D. A later inscription of the same dynasty is also given in the same Journal². Its date is, though not clearly put, as there is an obvious error of the engraver in omitting a letter in the date, beyond doubt. This runs (v. *l.* 8): "ashṭanavatyuttareshu ṭchateshu çakavarsheshv atīteshu". The ṭ in ṭchateshu is clear, and though 'sha(ṭ)' is entirely wanting, yet as 'shaṭ' is the only possible numeral it must be read 698 (=777-8 A. D.) The difference in character between the alphabets of the two inscriptions is so slight that I have not thought it worth while to give both.

c5.s2.t1.6

¹ Vol. I. The transcript needs some corrections. Jinālakke is clearly "for the Jinālaya" (Jain temple) and not "for the destruction of the Jains", as the whole inscription is Jain in style (cfr. the mention of the Vasus). I have been able to examine the original plates of this very valuable document, through the kindness of the Rev. G. Richter of Mercara.

c5.s2.t1.7

² Vol. II. 155 ffg. See especially Dr. Eggeling's remarks (iii. pp. 154).

c5.s2.t1.8

In Pl. xi. I have given the alphabet of a Cera inscription which, if genuine, would be (being dated about 247 A. D.) one of the oldest Indian grants known; it is, however, a forgery³. As nevertheless even forged grants have their value as evidence, if not of the early centuries A. D. down to the ninth, but it was probably in these later times a feudatory revolted against the older dynasty to which Açoka and the classical authors refer. The Mercara grant (Mr. Richter's) gives the kings as follows:

c5.s2.t1.9

³ The reasons are: 1. Č and bh open at the top as here do not occur before the tenth century.
2. U, kh, gh, ṅ and j are also modern forms of the letters, and of about the same date.
3. Subscript u is written in two ways, a practice comparatively recent.
4. The stroke in ph to distinguish ph from p is also late (about 10th-11th century).
5. The historical data contradict more or less those of other inscriptions.
6. The Çaka era was not used in S. India so early as the third century. In the fifth century it is very unusual.
7. Lastly (to judge from an impression) the plates are far too well preserved; the letters are all sharp and clear; this would not be the case if the grant was engraved in the third century A. D. There are other grounds, but these are, I think, sufficient for rejecting this grant.

c5.s2.t1.10

Kŏṅgaṇi Koṅgiṇi or Koṅguṇi (i.) The eighth king of the so-called chronicle! (about 350 A. D.)
Mādhava (i.)
Ari-(*i. e.* Hari)varmā (The grant *d.* 247 A. D. is attributed to him!) The true form of his name appears from the grants of çaka 435 (ii. *a*, 6.) "çrīmaddharivarmmamahādhirājasya," and less plainly in that of ç. 276 (i. *b.*, 2).
Vishṇugopa
Mādhava (ii.) in 454 A. D.
Koṅgaṇi (ii.) in 467 A. D.

c5.s2.t1.11

The Nāgamaṅgala grant continues:

c5.s2.t1.12

Durvinīta (? From 478 A. D. was reigning in 513 A. D.)
Mushkara
Çrīvikrama
Bhūvikrama
Prithivī Koṅgaṇi (? A. D. 727-777).
Rājmalladeva (?)
Satyavākya (987 A. D.) Mr. Kittel has edited 3 important Canarese inscriptions of this king of 780 (?), and 809 ç. s.; the third having no date (I. A. vi., pp. 101-3); I have photographs of these through his kindness.

c5.s2.t1.13

The great Inscription at Tanjore (11th century) mentions a Šēramān, but also a king of Karuvai (or Karūr) and a Govindacandra (king of Kannāḍa). There are, however, many difficulties about the genealogy and succession which remain to be cleared up. It would be well to term this 'the later Cera dynasty'.

c5.s2.t1.14

facts, yet palæographically, I allow this one a place. It shows the condition of the N. Cera character about the tenth century, which was then fast becoming assimilated to the Cālukya and transitional alphabets of the North. This was, no doubt, owing to the conquest of the Cera kingdom by the Coḷas in the ninth century, and the separation which followed between the two divisions of the Cera kingdom, that above, and that below the Coimbatore Ghauts. The first became assimilated to the northern kingdoms, the later had a new development under the Coḷas. Thus the old Cera alphabet of the North became superseded by the Telugu-Canarese, and that of the last developed into the Grantha-Tamiḷ. This tendency appears to have existed in the eighth century; the fall of the Ceras rendered it much more rapid. The chief distinction between the Cera and Cālukya characters is the tendency of the former to preserve the old subscript forms of **u** etc. when attached to a consonant, whereas in the latter these are gradually moved up to the right side of the preceding consonant.

Eastern Cēra

c5.s2.t2.1

What I have termed the Eastern Cera is of interest as being the source of the Coḷa Grantha, and hence of the modern S. Indian Sanskrit alphabet. I have used the term "Eastern Cera" rather to indicate the source from which it was derived, than with reference to the reign of the Ceras over the sea-coast of the North Tamiḷ country, a fact hardly doubtful, though, as yet, not fully supported by the evidence of inscriptions.

c5.s2.t2.2

This alphabet was confined to the old Tŏṇḍaināḍu or Pallava kingdom of Conjeveram¹, and is an offshoot of the early Cera before the full development of the horizontal line at the top of the letters of that alphabet; it is, therefore in origin, very near the Cave character; and the introduction of this alphabet into Tŏṇḍaināḍu is, probably, to be placed about the fourth century. In the second century A. D. (as we know from Ptolemy) this country was inhabited by nomads. In the seventh century Hiouen-Thsang found a small kingdom of which Kańśi (or Conjeveram) was the capital. He calls it Ta-lo-pi-tcha or Draviḍa[^1). The name of the family of kings of which inscriptions occur at Seven Pagodas (Māmalaippuram, the old port of Conjeveram) was Pallava, and they appear to have been formidable enough to have been attacked by the Western Cālukyas about the middle of the seventh century. Still later (about the eighth or ninth century) the country was conquered (according to Ellis) by the Cōḷas who had revived again after a long eclipse.

c5.s2.t2.3

¹ The account of the divisions of this kingdom by F. W. Ellis [in his Paper on Mirasi Right (pp. 51-9) edited by C. P. Brown, Madras, 1852] is still unquestionably the most valuable contribution to S. Indian Ancient Geography that has been written. It is much to be desired that Mr. Ellis's papers be collected and published in an accessible form, so as to be a lasting memorial of a truly great scholar. About the time that Bopp laid the foundations of the comparative philology of the Aryan languages, Ellis did the same for the Dravidian family [preface to Campbell's Telugu Grammar (1816) and "Dissertations"]; he was the first to decipher and explain the grants to the Israelites of Cochin, and he did this in a way that is still a model. (See *Madras Journal*, vol. xiii. part ii., pp. 1-11.) His labours to promote the study of Hindu Law and of Tamiḷ (annotated edition and translation of the Kuṟaḷ left unfinished, etc.) are still of the highest value. He was also the first to collect the S. Indian inscriptions. He died (accidentally poisoned through the carelessness of a native servant) while on an archæological tour in the Madura Province. His monument (at Ramnad) has an inscription in English and Tamiḷ, the former of which runs: "Sacred to the memory of Francis Whyte Ellis Esq. of the Madras Civil Service whose valuable life was suddenly terminated by a fatal accident at this place on the 9th March 1819 in the 41st year of his age. Uniting activity of mind with versatility of genius he displayed the same ardour and happy sufficiency on whatever his varied talents were employed. Conversant with the Hindoo Languages and Literature of the Peninsula, he was loved and esteemed by the Natives of India with whom he associated intimately[^1.1]), and his kind and playful disposition endeared

c5.s2.t2.4

1.1 S. Gordon, "Researches in South-India 1823-8" (London 1834), says (p. 54): "The natives are grateful for this favor (permission for a procession) to Mr. Ellis, then Collector of Madras, who was poisoned at Ramnad, on his way to Ramiseram; he assumed the native dress, and adopted their modes." (?!) him to his own countrymen among whom he was distinguished no less by his capacity as a public servant than by a mind fraught with intelligence and alive to every object of interest or utility. The College of Fort St. George which owes its existence to him is a lasting memorial of his reputation[^1] as an Oriental Scholar, and this stone has been erected as a tribute of the affectionate regard of his European and Native friends.” So little interest in science is there in S. India, that this eminent man is chiefly recollected among the Native Roman Catholics by some quasi-devotional poems in Tamiḷ which are attributed to him. [^1): 1) Al-Bīrūnī (11th century) mentions Draviḍa but as distinct from Kańśi, though a Cōḷa province (Reinaud, “Fragments”, p. 104). I see there is a paper by E. Burnouf on Draviḍa in the *J. As.* for Oct. 1828, pp. 241, ffg., but it is not accessible to me. [^2): 2) The genealogy of two inscriptions was first given by Profr. Eggeling (*I. A.* ii., p. 272; iii., p. 152) and has been finally extended and settled by Mr. Fleet (do: v., p. 154). The dates are not known, but it is safe to attribute these sovereigns to the 5th and 6th centuries.

c5.s2.t2.5

¹ It lasted but a short while; the Madras University is a different institution and does not even continue old traditions of scholarship.

c5.s2.t2.6

Of these Pallavas but little is known. The general use of the title ‘varman’ is common to them and the Veṅgi kings; and this and some other facts make it most probable that they belonged to one family. They had numerous contests and alliances by marriage with the Cera, Cōḷa and Cālukya families, but in the 11th century they were mere feudatories of the Cōḷas, and on this account were, apparently, repeatedly attacked by the Western Cālukyas. They disappear about the 14th century A. D.[^2)

c5.s2.t2.8

As regards the name ‘Pallava’, Mr. Kittel (“*Nāgavarma*”, p. xxi. *n.*) connects it with pŏllava and Tel. pallĕ=a rustic. The Paḷḷis (a cognate Tamiḷ caste) are now very degraded, but they may not have been so always. The Cōḷas and Pāṇḍiyas were merely Kaḷḷar or ‘Robbers’, a low caste at the present time.
The ‘Palakkaḍa’ of the first of these inscriptions must be the modern Pulicat. In the 11th century the Pallava

c5.s2.t2.9

² If this Skandavarmā or his grandson be the prince of that name who is mentioned in a Cera grant of 513 A. D. (?) as is likely, it may yet be possible to establish a synchronism. The references in the earlier Cālukya grants are all vague. capital was Conjeveram (Tamiḻ inscriptions and Bilhaṇa's Vikramān̄kadevacarita). The later Cālukyas never then attacked Tanjore, the real capital of the Coḷas, but only Conjeveram and (according to Bilhaṇa) Gān̄gakuṇḍa which is clearly Gañyaikkōṇḍānšōlapuram to the north of the Coleroon, and which was a great Coḷa town. 'Daçanapura' (=Tooth-town) as Mr. Fleet pointed out, must be a translated name, and it appears to me that the original must be Palakkaḍa. For pallu=tooth; kaḍa=place. Is this, then, one of the places where there was a tooth-relic of Buddha? There was such a place in Kaliṅga according to the Pali "*Dāṭhāvamso.*" [^1): Mr. Fergusson long ago stated this. ("*History of Indian and Eastern Architecture,*" pp. 175, 327 ffg.) [^2): Mr. Kittel has kindly informed me that according to the Č. Basava Purāṇa (ch. iv. 3-6) the first liṅga was found in Kerikāla Coḷa's time or c. 950 A. D. None of the great Çiva temples of S. India can be traced back beyond the 11th century A. D. (See Caldwell, "**Dravidian Grammar**", p. 86). This statement is fully supported by the inscriptions, and it is now certain that the liṅga worship is an importation from the North into S. India in, comparatively, recent times. See also Kittel's "**Liṅgakultus**", p. 16. ffg. [^3): An often engraved temple of this description is the one at Seven Pagodas on the sea-shore which is washed by the waves at high tide; another is on the northern bank of the Pālāṛu also near the sea and a few miles south of Seven Pagodas. These are the best examples that I know, but there are many others (often more or less altered) in the same neighbourhood. In the first, the Vishṇu cell is behind that in which the liṅga is found; in the others I know, the two cells are side by side. There is a correct plan of the first temple in No. 6 of the large map in Major Carr's book, and an incorrect one in pl. xxiii. [^4): Çan̄karācārya must be put at about 650-700 A. D. See my "**Sāmavidhānabrāhmaṇa**" vol. I. pref. p. ii. *n.* He preached at Conjeveram, it is said, but the tradition has little to recommend it.

c5.s2.t2.10

The character used in the two inscriptions of this dynasty that are in existence is a slightly developed form of the Veṅgi character, in the direction of the florid form found in the Seven Pagodas inscription (Pl. xii.). The secondary form of **ī** is the same as in the Veṅgi, but is clearer in the Pallava inscriptions. The letters **m**, **l**, **v**, **ç** and **h** are nearer to the Grantha forms. **Ç** and the secondary form of **u** are identical in the Pallava and later character found in the Seven Pagodas inscription.

c5.s2.t2.11

There can be no question that the caves and monoliths at Seven Pagodas, and in the neighbourhood, are of Buddhist-Jain origin[^1); the sculptures on the so-called *rathas* (monoliths) show (if anything at all) a slight admixture of Çȧiva notions, such as appear in the later Buddhism. Over several of the figures are, however, Vaishṇava names (*e. g.* *çrīNarasimhaḥ*) which ill-agree with the representations. In some of the caves are pure Vaishṇava and Çaiva mythological scenes. Taking into consideration the fact that this place is not mentioned by Hiouen-Thsang together with the nature of the sculptures, the original work is to be attributed to Jains of about the fifth century, and the alphabet of the inscriptions corresponds with this date. But as the caves now exist, they have been subsequently extended and adapted to the worship of Çiva[^2), or to the combined worship of Vishṇu and Çiva in the same temple, which is so remarkable a feature in the older and unaltered temples in the neighbourhood of Madras[^3), and which can only be attributed to the influence of the Vedānta doctrine as preached by Çan̄karācārya[^4). It is to the period of the adaptation that the dedicatory inscription, from which the alphabet in Pl. xii. is taken, belongs. The king under whom it was done is termed 'lord of the Pallavas' (Pallaveçvara) with the epithets "victorious in battle"

c5.s2.t2.12

(raṇajayaḥ), or "very fierce in battle" (atiraṇacaṇḍaḥ), and had, therefore, come under the northern brahmanical influence¹).

c5.s2.t2.13

¹ The Çambhu of these inscriptions is shown by the sculptures to be Mahādeva-Çiva; one inscription mentions Pārvatī.

c5.s2.t2.14

The inscriptions in question are not dated; the earlier ones (which consist of merely a few words in explanation of the figures on the so-called *rathas*) are in a character very near to the Veṅgi and early Cera, but distinguished from them by a few important variations. The first of these is the use to a considerable extent of secondary forms of ā, e and o separated from the consonant to which they belong and follow in pronunciation²); thus **rā**, **vā**, **çā** and **hā** occur in these words with the ā separated only, and in **kā**, **bhā** and **rā** both united to the consonant and also separate. **To** and **no** occur with the **o** separate. Besides these variations some of the letters, and especially **ç**, show an approach to the Grantha form. These and the two Pallava inscriptions represent the earlier stage of development.

c5.s2.t2.15

² See plates 16, 17 and 18 in R. A. S. Transactions ii. and in Major Carr's Collection of Papers relating to the Seven Pagodas (Madras 1869, 8vo.). I put at the editor's disposal my copies of the inscriptions at Seven Pagodas and also at Šāluvaṅkuppam, as well as the results of excavations which I had made in 1867, and some of these are printed by Major Carr (pp. 221-225).

c5.s2.t2.16

A still further development in the direction of the Grantha forms is to be found in the inscription on a monolith at Seven Pagodas, now used as a Gaṇeça temple; and also again in a still more developed form at Šāluvaṅkuppam. There can be no doubt that these inscriptions must be put at about 700 A. D. The first four lines of the Gaṇeça temple inscription describe Çiva in a way that was only possible after Çaṅkara's development of the Vedānta; and as the rest states that a Pallava king built "this abode of Çambhu", the inscription cannot be later than the eighth century; for the Coḷas about then conquered Tŏṇḍaināḍu³), and rendered such an inscription in praise of a king of the old dynasty, impossible⁴). Again, decidedly archaic forms of letters occur; *e. g.* the secondary form of ā which is occasionally turned up instead of down, and which early disappeared in the Cālukya and Cera characters. That again this inscription is later than those on the so-called *rathas*, follows from the words “atyantakāmapallaveçvara çrīhā (!?) raṇajayaḥ” being written in this character over a nondescript figure on one of them. Were all these explanatory labels over the figures of one date or of about the same date, such a difference in the writing would not have occurred. There is another circumstance which corroborates the date I have assigned to this inscription—the existence of a Nāgarī transcript of some verses selected from it with additions at Šāluvaṅkuppam. The Nāgarī is precisely that of the eighth or ninth century, and it is accompanied by a transcript in old Grantha very near to that of the eleventh century as given in Pl. xiii.

c5.s2.t2.17

³ According to Ellis. (Madras Lit. Trans. I.) I cannot find the authority.

c5.s2.t2.18

⁴ Major Carr has given my transliteration of this inscription (in Nāgarī) on pp. 221-2; as, however, it is not quite correct, I give it again here. (I mark the half-çloka by ;, ).
1. Sambhavasthitisamhārakāraṇaṁ vītakāraṇaḥ; bhūyād atyantakāmāya jagatā(m) kāmamardanah |||
2. Amāyaç citramāyo 'sāv aguṇo guṇabhājanaḥ; save (?) nirantare jiyād ********
3. Yasyā 'ngushṭhabharākrāntaḥ kailāsaḥ sadaçānanaḥ; pātālam agaman ma * çrīnidhis ta * * * * * |||
4. Bhaktiprahveṇa manasā bhavam bhūshaṇalīlayā; doshṇā ča yo bhūm(au) * * jiyāt sa çrībharaç čiram |||
5. Atyantakāmo nṛipatir nirjitārātimanḍalaḥ; khyāto raṇajayaḥ çambhos tene 'dam veçma kāritam |||
6. * * * prāṇanishkaḷaḥ . . . . . vijayatā çaṅkarakāma (rda) naḥ |||
7. Rājarājo navaraçmaç čakravarttijanārddanaḥ; tārakādhipatiḥ svastho jayatāt taraṇāṅkuraḥ |||
8. Çrīmato 'tyantakāmasya dvishaddarpāpahāriṇaḥ; çrīnidheḥ kāmarājasya harārādhanasaṅginaḥ |||
9. Abhishekaǰalāpūrṇe čitraraktāmbujākare; āste viçāle sumukhaḥ çirassarasi çaṅkaraḥ |||
10. Tene 'dam kāritam çambho-(r bhavanaṁ bhūtaye bhuvaḥ kailāsa-)mandiračubham prajānām ishṭasiddhyartham 11 . . . . . shashṭi . . . . . çivam . . . . . yeshā(m) na vasati hṛidaye kupathagativimo—
12. xako rudraḥ ||| atyantakāmapallaveçvara çrī— See pl. 14 in Major Carr's Collection of Papers, and in Dr. Babington's article (Trans. R. A. S. ii.). The translation given by the last (pp. 266-7) and reprinted by Major Carr, is not satisfactory. For Major Carr's "known as Raṇajaya" (p. 224), "famed, victorious in battle" should be substituted. Line 10 is completed from the Šāluvaṅkuppam inscription. In Dr. Babington's transcript the last lines are mixed up.

c5.s2.t2.19

It will be seen that I have put together the Cera and Pallava characters on the ground of their original similarity, and after-development in the same way.

B. Coḷa-Grantha and Middle Grantha

c5.s2.t3.1

The development of the early stages of the Grantha character is very difficult to trace, for the reason that the N. Indian civilization, when it got as far down in the peninsula as the Tamiḷ country, found there a people already in possession of the art of writing, and apparently a cultivated language[^1). Thus Sanskrit did not regulate the Tamiḷ phonetic system, nor did it become more to the people than a foreign learned language; it thus remained almost exclusively in the knowledge of the Brahmans, and the Grantha alphabet is nothing more than the character the Tamiḷ Brahmans used and still use, for writing their sacred books in a dead language. As there are no old MSS. written in this character, the records we possess of its early stages are most imperfect, and consist chiefly of Sanskrit words which casually occur in Tamiḷ inscriptions. I am aware of the existence of only a few Sanskrit inscriptions in the Grantha character more than three centuries old, and these are not dated, except one of 1383.

c5.s2.t3.2

The only interest this character possesses is the proof it affords of the derivation of the modern Grantha alphabet from the Cera, and thus from the Southern Açoka character. The first traces, I have found of it, are a few words in the grant B. to the Persian Christians, and which are, therefore, to be referred to the early part of the ninth century A. D. The letters are somewhat carelessly formed, but are almost identical with the Cera of the same period (cfr. Pl. xxvi.). To the fall of the Cera kingdom in

c5.s2.t3.3

[^1): This is proved by the entire absence of *old* inscriptions in the Tamiḷ country in the Grantha or Grantha-Tamiḷ characters, all such are in the Vaṭṭeḷuttu. See § 3 (below).

c5.s2.t3.4

the ninth century must be attributed the sudden appearance of brahmanical culture in the Tamil country and Malabar¹).

c5.s2.t3.5

¹ The succession of the Coḷa kings is as follows:

c5.s2.t3.6

The letters in the upper part of Pl. xiii. are taken from two sources: 1. the inscription round the shrine of the great temple at Tanjore (which belongs to the end of the eleventh century A. D.²); and 2. an undated inscription near Muruᵞamaṅyalam (in the Chingleput district) which is evidently of about the same date.

c5.s2.t3.7

The only point to which it is necessary to call attention is the advance made in about two centuries in the separation of the secondary forms of **ā**, **ē**, **ai** and **ō**, as shown by these inscriptions. In 825 A. D. only the ෙ (**e**) was clearly separated from its consonant; about 1100 A. D. **ā** is also generally separated (cfr. Pl. xii., **kā**, **tā**, **nā**, **pā**, **mā**, **yā**, **rā**, **hā**); the form in which it is attached being rare (cfr. **cā**, **çā**). The modern Grantha alphabet dates from about 1300.

c5.s2.t3.8

Owing to the long occupation of Ceylon by the Coḷa kings, the Grantha character probably affected the modern Singhalese alphabet³).

c5.s2.t3.9

Kerikāla-Coḷa (? about 950 A. D.)
|
Rājarāja-Coḷa *alias* Narendra¹⁾ (40 or 41 years) 1023 to 1064.
|
Vīra-Coḷa (D.) *alias* Kulottuṅga-Coḷa (i.) *alias* Rājarājendra (Rājarāja) Kŏppākesarivarmā²⁾ (49 years) 1064 to 1113. His abhisheka took place in 1079.
|
Vikrama-Coḷa (15 years) 1113 to 1128.
|
Kulottuṅga-Coḷa II. 1128 to ? ³⁾ Ruled over the whole Tamil country (Caldwell² p. 135) for, at least, 30 years.
.
.
Vikramadeva, reigning in 1235⁴⁾.

c5.s2.t3.10

² Letters taken from this are marked * . This immense inscription was photographed by Capt. Tripe in 1859 and published by the Madras Government. There is little Sanskrit in it except an introductory verse (Svasti çrīḥ | etadviçvarūpaçreṇimaulimālopalābhītam | çāsanam Rājarājasya Rājakesarivarmaṇaḥ ||) which belongs to a part of the inscription dated in the 26th year of the king's reign (=1090 A. D.), and a few words in the Tamil text.

c5.s2.t3.11

³ The Singhalese put the invasion of Ceylon by the Coḷas in 1023, and state that it was in 1071 the revolt began. (Tennent's "Ceylon" i., pp. 402-3.) This can hardly be correct, though it is according to Turnour.
1) F. (see above, p. *n.*) calls him Rājarājanarendra; E. Rājarāja-Coḷa; D. Narendra. This king must have restored Tanjore which according to Al-Bīrūnī was in ruins at the beginning of the 11th century (Reinaud, "Fragments", pp. 92, 121; "Mémoire", p. 284). This fact confirms the earlier Cālukya boasts of conquest, and was certainly owing to them.
2) That Kŏppākesari v. is the same as Rājarājendra is proved by the inscriptions at Tanjore and at the Varāhasvāmī temple at Seven Pagodas. (See *Madras Journal*, xiii., pt. 2; p. 36.) He seems to have been a great patron of Brahmans and of Çaivism, but he must also have been liberal to Buddhists, for Buddhamitra (the author of a Tamil grammar) called his work Vīraśoḷīyam after him.
3) He was reigning in 1134 A. D. the date of E. In his time there must have been a great many Buddhists in Tanjore, as Parākrama Bāhu (king of Ceylon 1155-1186) fetched his priest from there according to the Mahavamso:
4) Inscription near Seven Pagodas. (*Madras Journal*, xiii., pt. 1; pp. 50-1.) Kaliṅga was lost in 1228 A. D.

c5.s2.t3.12

In the following century (1310) the Coḷa and Pāṇḍya kingdoms were conquered by Muhammedans (Elliot, iii. pp. 51, 90, 203, etc.) and then by Vijayanagara; the inscriptions of the latter dynasty claim this, and grants in the S. Arcot district prove that it was the case so early as 1380, but even before the Muhammedan invasion the Coḷa kingdom had much declined, and the Madura kingdom was the chief one in the South, though it, with part of Coḷa, had been (as Mr. Rhys Davids proved) conquered by Ceylon (about 1173 A. D. Turnour, Mahavamso p. lxvi. *Bengal As. Soc. J.* No. 2 of 1872).

c5.s2.t3.13

Athā ’pi Coladeçīyam nānābhāsāvisāradam |
Takkāgamadharam ekam mahātheram susaññatam ||
Rājā rājaguruṭṭhāne ṭhapitvā tassa santike | etc. (Ceylon J. 1867, p. 26).

C. Modern Grantha and Tuḷu-Malayāḷam

c5.s2.t4.1

The materials for the history of this section of the S. Indian alphabets are also excessively defective. These alphabets were up to quite recent times in very limited use, and except in Malabar, are still applied merely to write Sanskrit. The name ‘Grantha’ by which the E. coast variety has been known for some centuries[^2), indicates that it was merely used for ‘books’ or literary purposes. This being the case, it is hopeless to look for old specimens, as palm leaf MSS. perish rapidly in the Tamiḻ country where they are mostly written on leaves of the ‘Borassus flabelliformis,’ far inferior to the Talipat leaves in beauty and durability. The oldest MS. I have been able to discover is Tanjore 9,594 which must be of about 1600 A. D. Autographs of mediæval authors who must have used this character (*e. g.* Appayya Dīxita in the sixteenth century A. D.) appear to be no longer in existence.

c5.s2.t4.2

There are at present two distinct Grantha hands. The brahmanical or *square* hand (cfr. Pl. xiv.), and the *round* or Jain hand which has preserved the original features of the early Grantha far better than the other. The first is used chiefly in the Tanjore province; the last by the Jains still remaining near Arcot and Madras.

c5.s2.t4.3

By far the largest number of Grantha MSS. now existing are brahmanical, and the lesser or greater approach of the writing to the angular Tamiḻ forms, is a certain test of the age of a MS. Such a hand as that shown in Pl. xxxiii. became quite obsolete by 1700 A. D. The only modern MS. that I have seen at all like it, came from Palghat (Pālakkāḍu); but occasional Malayāḷam forms of letters show its origin[^3).

c5.s2.t4.4

The Tuḷu-Malayāḷam alphabet is a variety of the Grantha, and like it, was originally applied only to the writing of Sanskrit; it is, therefore, the Grantha of the West, or the original Coḷa-Grantha as modified in course of time in a country secluded from all but very little communication with the east coast of the peninsula[^4). The importation of this alphabet into the S. W. coast must obviously have occurred after the Grantha had

c5.s2.t4.5

[^1): 1) The first complete representation of the E. Grantha alphabet is in “A Sanskrit Primer” by Harkness and Visvambra Sastri, (*sic*) (4°, Madras, College Press, 1827); the letters are, however, badly formed. The type now in use at Madras is very little better in this respect. [^2): 2) See the reference to the Basava-purāṇa (1369 A. D.) on p. 31. [^3): 3) MSS. in all these hands, and of different ages occur among those I presented to the India Office Library in 1870. [^4): 4) The history of the west coast is very obscure. There were, it appears, in the earliest times as down to the present, a number of small kingdoms given as twenty-five by the Portuguese. In the 11th century the west coast was more than once invaded by the Coḷa king, and it is termed Malanāḍu in the Tanjore inscription, though it is not clear to what extent of country this name should be applied.

c5.s2.t4.6

assumed its characteristic forms, or about the eighth and ninth centuries A. D.¹ But it is remarkable that the Tuḷu-Malayāḷam character preserves older forms which were modified at later times in the Grantha. (cfr. the Grantha ‘mu’ of the 11th century with the modern Grantha [Grantha character] and the Malayāḷam [Malayalam character] etc.)

c5.s2.t4.7

¹ See the words (from the grant to the Persian Christians), given in pl. xiii.

c5.s2.t4.8

Up to about 1600 A. D. the Tuḷu² and Malayāḷam alphabets (as shown by Sanskrit MSS.) are identical, and hardly differ from the modern Tuḷu hand given in Pl. xiv. MSS. from Malabar proper are generally written in a very irregular sprawling hand³, those from the Tuḷu country are neater. This character was termed in Malabar Ārya-eḻuttu, and was only applied to write Sanskrit works up to the latter part of the seventeenth century when it commenced to supplant the old Vaṭṭeḻuttu hitherto used for writing Malayāḷam. In the Tuḷu country it cannot be said ever to have been used for writing the vernacular language—a Dravidian dialect destitute of a written literature.

c5.s2.t4.9

² I have been told by a Brahman of the Mādhva sect that the founder (Ānandatīrtha, † 1198 A. D.) wrote his works in this character on palm leaves, and that some are still preserved in a brass box and worshipped at Uḍupi. It is probable, but I have not been able to get any corroboration of this story. The MSS. (if still existing) must be reduced by time to the condition of tinder; for the oldest MS. that I have seen in S. India which was of the 15th century, could not be handled without damage to it.

c5.s2.t4.10

³ The types used in printing the first edition of the Malayāḷam Gospels (at Bombay in 1806) exactly represent it.

c5.s2.t4.11

The application of the Ārya-eḻuttu to the vernacular Malayāḷam was the work of a low-caste man who goes under the name of Tuñjatta Eḻuttacchan, a native of Tṛkkaṇḍiyūr in the present district of Malabar. He lived in the seventeenth century, but his real name is forgotten; Tuñjatta being his ‘house’ or family-name, and Eḻuttacchan (=schoolmaster) indicating his caste. It is probable that there was a scanty vernacular literature before his time⁴, but it is entirely owing to him that the Malayāḷam literature is of the extent it is. He translated the Sanskrit Bhāgavata, and several similar mythologico-religious poems, leaving, however, a large infusion of Sanskrit, and writing his composition in the Ārya character. His translations are often erroneous, and beyond adopting the Vaṭṭeḻuttu signs for ṟ, ḻ and ḷ, (റ, ഴ and ള) he did nothing whatever to systematize the orthography which till lately was most defective⁵, or to supply signs for letters (*e. g.* ṵ) which are wanting in most of the other Dravidian languages. The Sanskrit literature was, after this, no longer a secret, and there was perhaps no part of S. India where it was more studied by people of many castes during the eighteenth century.

c5.s2.t4.12

⁴ Dr. Gundert considers the Malayāḷam Rāmāyaṇa to belong to a period of perhaps some centuries before the arrival of the Portuguese.

c5.s2.t4.13

⁵ The distinction between ĕ and ē, and ŏ and ō was first made within the last thirty years by Dr. Gundert. In a new fount of types used by the Carmelites at Kunamāvu (Cochin territory) an attempt is made to separate the secondary forms of u and ū.

c5.s2.t4.14

Tuńjatta Eḻuttacchan's paraphrases were copied, it is said, by his daughter. I have seen the ᴍs. of the Bhāgavata¹ which is written in a round hand sloping to the left (or backwards), and thus precisely agrees with the current hand used in Malabar proper, and which was imitated in the types cut to print Spring's Grammar in 1836. The modern types vary considerably. The Travancore hand is more angular².

c5.s2.t4.15

¹ This is preserved at Puḷakkale, a village in the Cittūr Tālūk of the Cochin territory, and not far to the south of Palghat (Pālakkāḍu). The MS. was much broken and injured by damp when I saw it in 1865. The author's stool, clogs and staff are preserved in the same place; it thus looks as if Tuńjatta Eḻuttacchan was a sanyāsi of some order.

c5.s2.t4.16

² There are some MSS. in this hand, among those I presented to the India Office Library in 1870; including one of the Mādhavīya Dhātuvṛitti. The types used to print books at Trevandrum follow this model. The first printed specimen of the Ārya-eḻuttu that I have seen is in the preface of vol. i. of Rheede's "Hortus Malabaricus". But Malayāḷam was printed already in 1577 when J. Gonçalves, a Spanish lay-brother of the Jesuits, cut type with which a Catechism was printed at Vaypicota near Cranganore. (Sousa, "Oriente Conquistado", ii., p. 110). The complete alphabet was printed by the Propaganda at Rome in 1772, 8º. "Alphabetum Grandonico-Malabaricum."

c5.s2.t4.17

The Sanskrit ᴍss. in this character (inscriptions there are none to my knowledge) present a peculiarity which deserves notice—the substitution of l and ḷ for a final t or ṭ, when these letters *unchanged* precede other consonants, or are final. Thus for tatkāla we find തൽക്കാല (talkāla), and for tasmāt തസ്മാൽ (tasmāl). This practice is totally wrong according to all authorities, and probably arises out of the tendency of the people of Malabar to slur over all surd consonants³.

c5.s2.t4.18

³ P. Paulinus a St. Bartholemeo followed this practice in his "Vyacaranam". (Sanskrit Grammar), and was in consequence ridiculed, but most unjustly, by Leyden and the Calcutta Sanskrit scholars of the last century.

c5.s2.t4.19

Apart from this singular practice, the Sanskrit ᴍss. from Malabar are among the best that can be had in India. Up to quite recent times the study of Sanskrit literature, and especially of the mathematical and astrological treatises, appears to have been followed in Malabar with more living interest than anywhere else in the South.

c5.s2.t4.20

It is hardly necessary to remark that the Ārya-eḻuttu or modern Malayāḷam alphabet is necessarily affected by the old Tamiḻ orthography as far as it is applied to the writing of Dravidian words. So in a Malayāḷam sentence ത, except if initial, should be pronounced δ in a Malayāḷam word, but t in one that is Sanskrit; ക should also be pronounced γ and g in the same circumstances. This, however, is but little observed, and Sanskrit words are commonly Dravidianized.

c5.s2.t4.21

The Tamiḻ and Canarese grammars give rules for Dravidianizing Sanskrit words⁴, but the subject deserves more attention than has yet been paid to it. These influences unquestionably affect the orthography of Sanskrit ᴍss. written in S. India.

c5.s2.t4.22

⁴ Nannūl, iii. sūtras 19-21. "Çabdamaṇidarpaṇa" pp. 46, ffg. cfr. also the introduction to the excellent Tamiḻ-French Dictionary, published at Pondichery (in 2 Vols. 8vo.) "par deux Missionnaires Apostoliques".

D. Grantha-Tamiḷ

c5.s2.t5.1

¹ The first specimen printed in an European book that I know of is of 1625 (Purchas’ “His Pilgrimes” vol. i., Bk. i., p. 185.) But this was not the first: In 1578 Father João de Faria cut Tamiḷ types, and printed on the “Pescaria” (Tinnevelly) coast in the same year: “O Flos Sanctorum., a doutrina Christãa, hū copioso confessionario, & outros livros” (Sousa, Oriente Conquistado. ii., p. 256.) I give the exact words as several different accounts have been given which are incorrect. F. de Sousa S. J. compiled his work from MSS. (not now in existence) at Goa in the 17th century. It was printed in 1710, and he died at Goa in 1713. Tamiḷ was printed also in the 17th century at a place in the Cochin territory called Ambalakkāḍu where the Jesuits had a house. F. Paulinus says (“*India Or. Christiana*”): “Anno 1679 in oppido Ambalacāta in lignum incisi alii characteres Tamulici per Ignatium *Aichamoni* indigenam Malabarensem, iisque in lucem prodiit opus inscriptum: *Vocabulario Tamulico com a dignificação Portugueza composto pello P. Antem de Proença da Comp. de Jesu, Miss. de Maduré.*” This writer was born in 1624, went to India (Madura) in 1647, and died at Toṭṭiyam (Madura district) in 1666. Barbosa Machado (“Bibliotheca Lusitana” i., p. 182.) from whom I have taken the dates of F. da Proença’s life, did not know of his Tamiḷ Dictionary, and since Profr. de Gubernatis has made a fruitless search in the library of the Propaganda, there is little hope that a copy will be discovered. If any exist, it must be in the ‘Mission House’ at Halle, or at Goa. The first (engraved) Tamiḷ alphabet that I know of, is in Baldæus (1672) “Beschrijving der Oost-Indische Kusten Malabar en Choromandel” p. 191 ffg. The first printing press in India was that of I. de Endem at Goa—1563.

c5.s2.t5.2

The earliest inscriptions in which this character occurs are of the tenth century, and belong to the earlier kings of the revived Coḷa kingdom; they are at Conjeveram and in the neighbourhood of Madras and in the Kāverī delta. South of Tanjore, there are few old inscriptions in this character²).

c5.s2.t5.3

² The old Grantha-Tamiḷ alphabet was given by Babington in Pl. xiii. of vol. ii. of the Transactions of the Royal As. Society of London; he apparently took it mostly from the inscription of Šaluvankuppam, which is probably of the year 1038 A. D.; but he added letters from other inscriptions of later times and from other places. I have examined this inscription which is very roughly cut, and therefore preferred that at Tanjore which is of various dates chiefly from 1073 A. D. to 1089. It includes a large number of grants with many clauses in each.

c5.s2.t5.4

The origin of this Tamiḷ alphabet is apparent at first sight; it is a brahmanical adaptation of the Grantha letters corresponding to the old Vaṭṭeḻuttu, from which, however, the last four signs (ḷ, ḻ, ṟ and ṉ) have been retained, the Grantha not possessing equivalents. The form of *m* is also rather Vaṭṭeḻuttu than Grantha. Çaṅkarācārya is said to have preached with much success in the Coḷa kingdom; that it was the seat of a great brahmanical mission in the tenth century is shown by the inscriptions. This alphabet, accordingly, represents the later brahmanical Tamiḷ culture as opposed to the older culture of the Jains of Tanjore and Madura, and the Buddhists of Tanjore; but the earliest stage of the history of this alphabet is very obscure before the 11th century.

c5.s2.t5.5

Inscriptions in this character abound in all the Northern Tamiḷ country, where there is scarcely a temple of any note which has not acres of wall covered in this way. I need only mention the great temples of Conjeveram and Tanjore. It is, however, very unusual to find any with dates that can fully be identified, most being only in the year (āṇḍu) of the king’s reign (or life?), and genealogical details being very rarely given in them. As the list of the Coḷa and Pāṇḍya kings is uncertain, it is thus impossible to procure a series of palæographical standards, and I, therefore, give only three specimens¹). These will show how very little alteration and development occurred between 1073 and 1600 A. D. A very near approach to the modern Tamiḻ character must already have been reached about 1350 A. D. (cfr. Plates xiv. and xix.). The most important point is the conversion of the curve at the top of **k**, **s**, **t**, **n**, and **r** into an angular stroke, thus [Figure:], and this feature appears in an inscription of about 1200. The last letter to finally assume the modern form was **k** about 1500. The greatest development has occurred in this present century owing to the increased use of writing, and to the arbitrary alterations of the type founders²).

c5.s2.t5.6

¹ Our information about the Coḷa kings is confined to the 11th and 12th centuries (see above, p. 39, note). As regards the old Pāṇḍya kings we know still less. (See "The Madura Country" by J. H. Nelson, 1868). It is possible that the mention of some of these kings by the Ceylon annalists, Marco Polo, and the Muhammedan Historians may eventually furnish a clue. What information there is, has been last collected and discussed by Dr. Caldwell, "Dravidian Grammar" pp. 139-146. 536. According to an inscription at Chillumbrum (Śidambaram), Vikrama Pāṇḍya was succeeded in the 11th century by his son Vira Pāṇḍya who was conquered by Koppākesari the Coḷa. He established his younger brother on the throne there. This prince's name was Gaṅyai-kkŏṇḍa Coḷa, but he took the name Sundara Pāṇḍya Coḷa (Inscr. at Karuvūr). In 1173 Madura was conquered by the Singhalese who put a Vira Pāṇḍya on the throne, and about 1310-1315 by the Muhammedans, under whom it continued till about 1370 when it fell under Vijayanagara. From about 1100 down to 1310 there is much confusion, and it is impossible to identify the kings; this is probably to be explained by contests of the Pāṇḍyas and Pāṇḍya-Coḷas for the throne. Dr. Caldwell gives the last Pāṇḍyas as follows: Parākrama Pāṇḍya (1516-1543) Vikrama Pāṇḍya (1543-1565) Vallabhadeva, Ativirarāma (1565-1610) He was alive in 1605; in literature, he is chiefly remarkable as a diligent translator or patron of translators from Sanskrit. I have a grant by a Sundara Pāṇḍya d. Rudhirodgāri thirteenth year of his reign; this can only be 1623. Vallabhadeva, therefore, died in 1610. But all these Pāṇḍyas had no real power.

c5.s2.t5.7

² The first edition of the N. T. in Tamiḻ (4°. Tranquebar, 1714) is printed with type that exactly reproduce the character of the Tamiḻ inscriptions of the fifteenth and sixteenth centuries.

c5.s2.t5.8

The Grantha-Tamiḻ differs from the Grantha-alphabet in precisely the same way as the Vaṭṭeḻuttu, as far as the reduplication of consonants and the expression of the absence of the inherent vowel (virāma) are concerned. The puḷḷi or dot above the consonant which serves the purpose of the virāma, does not occur in any of the inscriptions I have seen, and it is omitted in the earliest printed books³). The famous Jesuit C. J. Beschi (in India, 1704-1744?) is the author of a great improvement in modern Tamiḻ orthography—the distinction between the long and short **e** and **o**. This he effected by curving the top of the [Figure:]- used to express the short e, thus [Figure:], and the same sign serves (in the compound for **o**) to express the long **ō**⁴). Before then, he states, the short **a** and **o** were *occasionally* distinguished by a stroke (the Sanskrit prosodial mark) above them. In the alphabet as given by Baldæus (1672) the same character stands for **ĕ** and **ē**, **ŏ** and **o**; but he notices the fact that these letters are long and short.

c5.s2.t5.9

³ It appears to have been known to the Tamiḻ grammarians.

c5.s2.t5.10

⁴ "Grammatica Latino-Tamilica, in quā de vulgari Tamulica lingua" etc. (Tranquebar, 1739, 12°).—"longis (e et o) nullo notatis signo brevibus superscribendum docent illud signum(-). Attamen nullibi hæc signa præterquam paucis aliquot dictionibus ex inertia fortasse amanuensium superscribi vidi unquam.......addo excogitasse me alium et faciliorem modum distinguendi e et o longa a brevibus: scilicet, cum utrique inserviat littera *[Figure:]* combu dicta; si hæc simplici formā scribatur, erit e breve et o breve: si autem inflectetur in partem superiorem, ut infra dicam de i-longo, sic *[Figure:]*, e at o erunt longa."

c5.s2.t5.11

Beschi omits to mention the information given by the Tamiḷ grammarians which is of some interest. The Tŏlkāppiyam (i., 1, 16) states that a dot is to be put over **ĕ** and **ŏ**. This practice was, therefore, recognized by the grammarians about the 8th century. In the 11th century the Vīraśōliyam of Buddhamitra (i., 6: “ĕɣaravŏɣaramĕyyiṟ puḷḷi mēyvum”) repeats this rule, and so does the Nannūl in still more recent or, as Dr. Pope considers, in quite recent times. Of this also, I have not been able to find the least trace in the inscriptions.

c5.s2.t5.12

The angular form of this Tamiḷ character is owing to a wide-spread practice in the South of India, of writing with the style resting on the *end* of the left thumb nail; in Malabar and the Telugu country the roundness of the letters is to be attributed to the practice of resting the style *òn* the left *side* of the same thumb.

c5.s2.t5.13

The map shows a great extension of the Grantha-Tamiḷ alphabet to the North extending over the deltas of the Kṛishṇā and Gōdāvarī; this occurred under the Coḷa rule in the eleventh and twelfth centuries. Inscriptions in Tamiḷ and in the form of character given in Pl. xviii. still exist (or existed till lately) in some of the islands of the Gōdāvarī delta, and the village accountants were originally all Tamiḷ Brahmans¹). The ritual of many temples was also in this language. This however did not continue long, for in 1228 the Coḷas lost Kaliṅga which was conquered by the Ŏruṅgal king²), and by the beginning of the fourteenth century, Telugu inscriptions and grants only appear.

c5.s2.t5.14

¹ This remarkable extension of Tamiḷ to the north was first pointed out by F. W. Ellis (“On the Law-books of the Hindus” in Madras Lit. Trans. i.); I was able to verify it for myself in the Nellore province.

c5.s2.t5.15

² “About the year . . . A. D. 1228, the fourth prince of this [the Ŏruṅgal] line drove the Shózha Raja entirely out of Cālinga; this fact and date being proved by a remarkable inscription on a stone now standing on the westernmost point of the island of Dive” . . . Ellis, *u. s.*

c5.s2.t5.16

Tamiḷ is remarkable among the Southern languages for using a number of abbreviations for common words such as month, year, fanam etc. These appear in common use at the beginning of the 16th century³).

c5.s2.t5.17

³ They are to be found in most Tamiḷ Grammars.

c5.s2.t5.18

In the 16th and rarely in the 17th century, an abridged style of writing Tamiḷ is not unusual in which the long **ā** is connected with the preceding consonant; thus [Figure:]=**rā**⁴). Other letters *e.g.* [Figure:] (=**ṭṭ**) are also joined in a similar way; this manner of writing exists now in the cursive hand only.

c5.s2.t5.19

⁴ See inscription of ç. 1454 (=1532-3 A. D.) on the Gopura of the Piḷḷaiyār Temple in the Fort at Tanjore. Thus in the Digambara cosmogony called ‘Trilokasāra’ the gāthas 848-850 describe the persecution of some Jain ascetics by Kalki (a king said to have lived 394 years after the Çakarāja). These run:

Linked texts (11)