BnF INDIEN 528: Five Tamil Alphabets

The source of all the pictures of INDIEN 528 is BnF

BnF INDIEN 528 (ex- Tamoul 528; Cabaton, Catalogue sommaire, Paris, 1912, p. 74; Figs. 1‒2) is a bundle of 15 regular palm leaves of same dimensions (370 x 45 mm) protected by two cover wooden planks. It contains five Tamil alphabets or syllabaries (French “syllabaires”) from Sri Lanka.

Fig. 1 — BnF INDIEN 528, tied bundle — Courtesy of BnF

Fig. 2 — BnF INDIEN 528, untied bundle — Courtesy of BnF

The manuscript is in a very good state of preservation. It has one hole only, through which a thread passes in. The thread is sealed with a lead seal. The bundle, which is not foliated, is in disorder as it is ordered by the sealed thread. We have below referred to the folios according to our new foliation after having reordered the manuscript.

As for provenance, a label in ink on the top right corner of f15v provide a clue, as one reads “Tolfrey. 4.” (Fig. 3). The manuscript thus comes from the collection of a certain Tolfrey. Other Tolfrey’s manuscripts from Sri Lanka are kept at the BnF.

Fig. 3 — BnF INDIEN 528, f10v — Courtesy of BnF

Wikipedia tells us that there are several Tolfreys linked to Ceylon/Sri Lanka: Edward Tolfrey, William Tolfrey and Samuel Tolfrey. Ours seems to be Samuel as one of the manuscript of the BnF Tolfrey collection bears the label “Saml. Tolfrey Esqre” (see Linayaratne, Catalogue des manuscrits singhalais, Paris: Bibliothèque Nationale 1983, N° 30, p. 104).1

In the copy of Vinson unpublished catalogue, supplemented with further descriptions by Léon Feer, one reads on f120r ad N° 528: “Trois Syllabaires Singhalais”. Rather than a mistake, as the manuscript contains five different Tamil alphabets, this statement appears to concern the provenance only. And, Feer notes that there are five alphabets.

BnF INDIEN 528 is also mentioned by Fauriel (1891, “Notices sur des mss. cingalais, tamouls, palis, etc., ou relatifs à ces langues (fonds Tolfrey), par M. Fauriel” in Anciens catalogues de la Bibliothèque du Roi et de différentes collections qui sont venues l’accroître aux XVIIe, XVIIIe et XIXe siècles, f6r “Tolfrey N° 4”). The script is correctly identified as being Tamil.

Cabaton does not suggest any date for the manuscript. Considering the dates of death of the different Tolfreys, the end of the 18th c. or the beginning of 19th c appears to be reasonable guesses.

In terms of paratexts, the manuscript contains a series of five Tamil alphabets, all titled arivari in the left margin. The Madras Tamil Lexicon tells us s.v. “அரிவரி ari-vari, n.”: “< Hari +. Alphabet book, the first line of which begins with அரி, Viṣṇu.” This is not the case here as we have what seems a classical Tamil alphabet in lay-out and conceptualisation (e.g. order of letters), with, as we will see, a specific sign for the āytam. There are at least three different hands but, from the examination of physical object (palm-leaf type and dimensions), there is no question that we have only one codicological unit.

For the sake of clarity, the text is not inscribed in scriptio continua but the letters are clearly demarcated by space, dots, and lay-out.

The five Tamil alphabets are distributed as follows:

Alphabet 1 (Fig. 4‒5). 5 folios (i.e. folios 1‒5 in the present order of folios, which remains folios 1‒5 of the reordered manuscript), 10 written pages, 2 lines per page. All the letters (265) are numbered in smaller script (in Tamil and in Latin numerals). This is very carefully executed alphabet.

Fig. 4 — BnF INDIEN 528, Alphabet 1, f1r — Courtesy of BnF

Fig. 5 — BnF INDIEN 528, Alphabet 1, f1v — Courtesy of BnF

Alphabet 2 (Fig. 6). 2 ff. (i.e. ff. 6 and 8 in the present order of folios = ff. 6‒7 of the reordered manuscript), 3 written pages, 8 lines per page, except last written page (4 lines). This is definitely from another hand than Alphabet 1.

Fig. 6 — BnF INDIEN 528, Alphabet 2, f6r — Courtesy of BnF

Alphabet 3 (Fig. 7). 2 ff. (i.e. ff. 7 and 9 in the present order of folios = ff. 8‒9 of the reordered manuscript), 4 written pages, 6 lines per page, except last written page (2 lines).

Fig. 7 — BnF INDIEN 528, Alphabet 3, f8r — Courtesy of BnF

Alphabet 4 (Fig. 8). 3 ff. (i.e. ff. 15, 10 and 11 in the present order of folios = ff. 10‒12 of the reordered manuscript), 5 written pages, 4 lines per page.

Fig. 8 — BnF INDIEN 528, Alphabet 4, f10r — Courtesy of BnF

Alphabet 5 (Fig. 9). 3 ff. (i.e. ff. 12‒14 in the present order of folios = ff. 13‒15 of the reordered manuscript), 5 written pages, 4 lines per page.

Fig. 9 — BnF INDIEN 528, Alphabet 1, f13r — Courtesy of BnF

The order of presentation of the letters is the same in the five alphabets:

    • 1 line for the initial vowels series, complemented by the āytam. Note that Grantha alphabets of South India in the BnF collection give (visarga) and (anusvāra) at the end of the series of initial vowels.
    • 1 line for the series of 18 consonants, each with ­-a.
    • 18 lines, 1 per individual consonant, with the full range of vocalisations and again the consonant with -a. See, e.g., f10r3, ka kā ki kī ku kū ke (i.e. as the manuscript does not use puḷḷi) kai ko (i.e. ) kau ka (unless the final ka is meant for as in BnF INDIEN 432, which uses puḷḷi).

The 5 alphabets have the same number of lines (20) and letters (not thoroughly checked). There is no two-curled kompu ே, nor puḷḷi used in any of them.

As alphabets, these five examples share certain features, some that might be characteristic of Lanka.

First ― and this is not a big surprise ―, they use old forms of letters, e.g. for ணா ṇā, ணை ṇai, லை lai, or றா ṟā (Figs. 10‒11).

Fig. 10 — BnF INDIEN 528, Alphabet 1, letters ṇā, ṇai, ṇo, and ṇō — Courtesy of BnF

Fig. 11 — BnF INDIEN 528, Alphabet 1, letters lai, ḷai, ṟā, ṉā, and ṉai — Courtesy of BnF

The āytam (ஃ ) (Fig. 12) comes at the end of the initial vowel series. It has a peculiar form found also in BnF INDIEN 432. It looks like modern கூ kū, while the கூ kū given elsewhere in the alphabet is similar but slightly different, except in Alphabet 2, where and are not clearly distinct.

Fig. 12 — BnF INDIEN 528, letters and in alphabets 1 to 5 (from top to bottom) — Courtesy of BnF

Note also the forms of தி ti and தீ (Fig. 13) in the five alphabets. This form of தி ti is known from other manuscripts, but quite rare, if non extant, in inscriptions.

Fig. 13 — BnF INDIEN 528, letters ti and in alphabets 1 to 5 (from top to bottom) — Courtesy of BnF

The two-curled kompu ே, which differentiates medial ē and ō from e and o, is not used in any of our five alphabets, but other devices are used to distinguish, for some consonants, these vowels. For instance in ணோ ṇō (Fig. 14) is differentiated from -o in ணொ ṇo by the addition of a loop to the second element of vocalisation in Alphabet 1. In the Alphabets 2 and 5, the loop goes above the in ṇō. In Alphabet 3, ṇō and ṇo appear to have been interchanged by mistake. In Alphabet 4, ṇo and ṇō are not clearly differentiated.

Fig. 14 — BnF INDIEN 528, letters ṇo and ṇō in alphabets 1 to 5 (from top to bottom) — Courtesy of BnF

The letters ṟo றொ and ṟō றோ as well as ṉo னொ and ṉō னோ can be similarly differentiated (Fig. 15). While Alphabets 1 and 3 do not make any distinction, Alphabet 2 does only between ṉo and ṉō and Alphabets 4 and 5 use the supplementary loop (except for ṉō in Alphabet 4).

Fig. 15 — BnF INDIEN 528, letters ṟo, ṟō, ṉo and ṉō in alphabets 1 to 5 (from top to bottom)— Courtesy of BnF

Further examination of the BnF manuscripts will maybe enable us to know whether these script peculiarities are found only in Lanka.



Cite this blog post
Emmanuel Francis (2019, November 14). BnF INDIEN 528: Five Tamil Alphabets. Texts Surrounding Texts. Retrieved March 29, 2024, from https://doi.org/10.58079/uxyj

Emmanuel Francis

Chargé de recherche CNRS (CEIAS-UMR 8564, EHESS-CNRS, Paris)

More Posts

  1. The folio of BnF INDIEN which bears the label “Saml. Tolfrey Esqre” is in fact misplaced: after reordering of the manuscript, it is now my folio #10. The fact that the bundle bears, on a misplaced folio taken to be the first or last, the name of its owner/collector/donor seems to indicate that the manuscript has been disordered for long. []

Emmanuel Francis

Chargé de recherche CNRS (CEIAS-UMR 8564, EHESS-CNRS, Paris)

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Search OpenEdition Search

You will be redirected to OpenEdition Search