Orthographic conventions (Re: New discovery in Tamil Nadu

Jean-Luc Chevillard jean-luc.chevillard at UNIV-PARIS-DIDEROT.FR
Sun Jun 28 11:15:38 UTC 2009

Dear Stefan Baums,
Dear George Hart,
dear list members,

it is possible that the spelling is simply dictated by the orthographic 
conventions of that time.

In the system which is described/codified  for ancient tamil in the 
Tolkāppiyam Eḻuttatikāram (TE)
(i.e. the first book [having 483 sūtras] inside of the Tolkāppiyam [a 
treatise consisting of 3 books])
there are 12 vowels (uyir) and 18 consonants (mey),
to which must be added 3 special items (over-short u, over-short i and 
which were probably not written with special signs at the time
but which are important in the precise description
of the conditions of application of a number of morpho-phonological rules
and also for the correct scansion of verse.

The twelve vowels are: a ā i ī u ū e ē ai o ō au
and the eighteen consonants are: k ṅ c ñ ṭ ṇ t n p m y r l v ḻ ḷ ṟ ṉ

All the vowels (except au, to my current knowledge)
are attested inside the corpus of early inscriptions
but there are nevertheless a number of things to be noted.

"ai" is nominally a long (neṭil) vowel but is in fact "reduced" in 
duration most of the time
See what sūtra TE57i says [i.e. sūtra 57 according to the numbering by 
the commentator Iḷampūraṇar]:

ōraḷa pāku miṭaṉumā ruṇṭē // tēruṅ kālai moḻivayi ṉāṉa (TE57i).

"There are also environments/places wherein (they) become one aḷapu 
(unit measure of time).
When one examines, (it will be seen that it happens so when they are 
employed) in words"
 [TE57i translation by V.S. Rajam, 1981]

Later grammarians explain that, except when naming itself (in a 
metalinguistic context),
the vowel "ai" is reduced [they call it "aikārak kuṟukkam] and has a 
duration of 1 measure,
which seems to mean that what is written "ai" should in fact be understood
as referring to a short vowel [which we could prefer to write "ä"], with 
a duration of 1 unit of time.
["aikārak kuṟukkam" was precisely the topic of a presentation I made at 
the last ICHoLS conference, at Potsdam, in september 2008,
and which should be published at some point in the future]

However, one must add that several grammarians insist on the fact that 
"ai" has sometimes a duration of 1,5 measure (especially in the first 
syllable of words).

That seems to mean that:

-- in the initial syllable of a word, what is written "ai" stands for 
"ay" or "äy" (the opposition a/ä being neutralized before "y"), which 
has a duration of 1.5 measures

-- in other syllables, what is written "ai" stands for the short vowel 
"ä" ["aikārak kuṟukkam" = reduced "ai"), which has a duration of 1 measure.

However, a number of other facts have to be mentionned:

TE54i and TE56i describe what is often explained [by later writers] as 
alternate spellings for "ai" [they mention those spellings as "pōli 
eḻuttu"], but it is possible that it was in some cases the only way to 

The spelling convention described in TE56i makes use of "y"
(Iḷampūraṇar says that "aivaṇam" can be written "ayvaṇam"

As explained by I.Mahadevan[2003:183], the earliest occurence of "ai" as 
standalone "initial vowel" is found in a 6th century A.D. vaṭṭeḻuttu 
inscription at Thirunatharkunru (see Mahadevan[2003:473] for an 
idealized facsimilé). However, combined with a consonant, the "medial 
-ai" sign is found from the 2nd century B.C. onwards (see 

I.Mahadevan [2003:269] notes that "-ai" is often followed by "paragogic" 
suffix "y".

All this may suggest that the spelling "vayra" (as given by the article 
in The Hindu) might be the spelling [va-ya-ra] used in those days
to write what would be written nowadays as vai-ra (or vai-ra-m).

It may have been pronounced "väyra" (or vära ?).

Of course, it is slightly (very? :-)
unreasonable to speculate solely on the basis of an article in The Hindu,
but hopefully, we shall have in the near future more scholarly sources 
of information,
thanks to scholars such as Y.Subbarayalu, K. Rajan and  I.Mahadevan.
But of course, journalists are always impatient.

Best wishes

-- Jean-Luc Chevillard (Paris)

Stefan Baums a écrit :
> The Pali (and other MIA) word is vajira with epenthesis, so I
> imagine that that is what was borrowed into Tamil. The further
> sound change of intervocalic j to y could have happened either in
> MIA (certainly in the Northwest at the same time) or, apparently,
> in Tamil. – Why it would be written vayra (not vayira) in the new
> inscription remains puzzling, especially if it is still vayiram in
> modern Tamil, but maybe the newsreport just omitted the i. Well, I
> assume photos will be published in due course, and then we will
> see.
> Best regards,
> Stefan Baums

More information about the INDOLOGY mailing list