/jñ/ is has been a very unstable sequence in India for a very long time, and its varying treatment can be traced in part through Prakrits (see Pischel’s
Prakrit Grammar, §88, §119, §303). Already, in Aśoka’s edicts we see a simple (very rarely doubled) /n/ (Kalsa in N, Dhauli in E) or /ṇ/ (NW and far south) or /ñ/ (at Girnar) in forms like
ānapayati or āṇapayati, ‘orders’ (cf. Sanskrit ājñāpayati), with [n] or [ṇ]; Pali has /ñ/ (pronounced [ny]) initially and /ññ/ medially.
One can sum up by saying that the initial /j/ has a tendency to disappear, or to survive in a minimal way in western India as an unexploded /d/ — that is, the beginning of a /j/ = [dzh/ without the fricative which is dropped before the [ny], OR realizing
the /j/ with a velar stop instead.
The options by region, generally, seem to be [gy] with or without accompanying nasalization much of the Hindi belt, [dny] in Maharasthtra and thereabouts, and either [ny] or [gn]/[kin] in the Tamil sphere, where an epenthetical vowel is inserted to break
up the cluster. This also happens in northwestern India, where we get Ismaili works called
ginān (< jñāna) from Sindh, Gujarat, and Panjab.
What does NOT seem to happen much is what Westerners today most often produce: j + epenthetical vowel + n
The regional variation is discussed (too) briefly on p. 78 of:
Chatterji, Suniti Kumar. 1960. ‘The Pronunciation of Sanskrit’. Indian Linguistics 21, 61–82.
In my edition of the Nīlarudropaniṣad (p. 87), I use the Maharashtrian pronunciation as diagnostic to identify the origin of a particular variant reading in the text:
I discuss the Prakrit and Tamil reflexes of the jñ conjunct a bit here as well (on pp. 425–426):
Professor of Religion and Adjunct Professor of Law
Chair of the Department of Religion
Chair of the Middle East and South Asia Studies Program
204 Tucker Hall
Washington and Lee University
Lexington, Virginia 24450
Dear list members,
I've been asked about the pronounciation of the conjunct jñ .
"jña: the pronounciation of this varies wid is that it is a palatalisation so to speak en bloc of an original gn, In some places, for instance, it is like gya, in others dnya. . . . The point about jñ is that it is a palatal isation so to speak en bloc
of an original gn.. . . . Perhaps the most approprate of the modern pronunciations to adopt is therefore gnya which (by adding y to gn) thus crudely represent a palatalisation."
1) Is jña ever pronounced in India as jnya and if so where?
2) where in India are the different pronounciations Coulson lists (gya, dnya, gnya) located?