9. Wrapping up the vowel terminations (pragŗhyam) that block sandhi changes

We are looking at the series of sutras starting from 1.1.11, which give the different circumstances under which word-ending long vowels īT ūT eT (long ī, long ū, and long e)  are left unmodified by the normal sandhi rules (given elsewhere in Panini), in which case they are termed pragŗhyam (which I fancifully translate as “constrained”). We saw the following conditions so far: sutra 1.1.11 says that a dual form (could be a verb or noun) ending  in the long vowel will not take part in sandhi, i.e. will be pragŗhyam; next, in 1.1.12, we were told that forms of adas “that” having endings –m+long vowel would be pragŗhyam; and in 1.1.13, certain Vedic pronoun forms ending in e like asme, yuşme, are also pragŗhyam, i.e. they are not combined with following vowels under usual sandhi rules (which introduce a y sound for an i, or  v sound in place of u, but that is a different sutra, 6.1.77 iko yaņ aci which we will keep for later).

Now the next three sutras give some more contexts in which a terminal long vowel (īdūded or rather, īdūdet if we quote the string without sandhi modification of the final sound) is pragŗhyam, constrained. These are given below and discussed briefly:

1.1.14 nipāta ekāc anāŋ (the symbol ŋ stands in for the ng sound)

This is interesting both for the principle it establishes, and for the brevity of the terminology. The import is straightforward: a particle (nipāta) of one (eka) vowel (aC, which we know from our Siva-sutras, see Page above) except (an-) the long ā (āŋ or āŊ to show the second letter is a nonsense marker, an iT). We supply the predicate to complete the idea, which is pragŗhyam, constrained, i.e. does not get modified by sandhi.

Terms like ekāc and anāŋ are delectable, as they express a whole phrase (except for…, which is not…) by such a concise formulation. Examples given with i, u, e denote the single vowel (ekāc) as exclamations, like ‘O adversity’ which would not be pronounced as ‘Ovadversity’ with sandhi: a apehi “hey, stay away”. But there is a complication about the particle ā; in some cases it is still a pragŗhyam, in apparent disregard of the sutra, and will remain unmodified. In such cases it is not technically āŋ, and therefore behaves like any other single vowel ekāc. So what are these mysterious anāŋ (not- āŋ) that buck the rule? They are used like a preposition meaning “a little” or a pre-verb meaning  “up to” rather than being just an exclamation or a sigh:   ā uşņam = oşņam, a little hot. Not a pragŗhyam!

The next sutra

1.1.15 ot

assigns pragŗhyam character to a particle (nipāta, supplied from the preceding sutra) ending in the long vowel o: āho iti “emphatic interrogative particle”, āho atra “what now!”. The particle need not be of one syllable (ekāc). 

The next three sutras refer to forms of vocative case (hailing, calling someone) ending in o (carried from the previous sutra), or ending in u, or in a nasalized ū (we don’t have a simple symbol for this), before a “non-Vedic” iti. These are special cases which need not detain us here, but they do say something about how closely the sage has observed the spoken tongue.

The next sutra, and the final one in this span of sutras on the contexts of   pragŗhyam (constrained endings) is a little more interesting:

1.1.  19  īdūtau ca saptamyarthe

Here is our old friend īT, now attached to the companion ūT, and the compound word put in nominative case, dual number (1/2), hence meaning “the two vowels long ī and (ca) long ū”. What of them? When they are word endings in the seventh (saptami) case meaning (arthe), i.e. in the locative case, they will be  pragŗhyam (constrained endings), i.e. they will not be modified as per sandhi rules when another vowel follows (as has happened to the short i in saptami+arthe= saptamyarthe in the sutra!).

The span of sutras 1.1.11 to 1.1.19 displays many characteristics of Panini’s styles and conventions: the coining of compact words to denote conditions (e.g. ekāc, anāŋ) taking advantage of the ground already laid by his intelligent (crafty!) grouping of letters and giving them range names by interposing iT markers (ekāC, anāŊ); the carrying over of words from previous sutras by implication, and which have to be supplied mentally by the reader; the inter-relation between sutras in different parts of the treatise; and so on. Above all, is the overwhelming austerity of thought and expression: almost as if the man wished to be ever ready to go to his Maker, so sparse was his verbal baggage. But this very austerity of expression meant that various aids were required to keep Panini’s thought alive: we shall touch on these in the next post.

(A question to explore: in the term anāŋ, we understand āŋ to stand for the sound ā alone; what determines the choice of the terminal iT, is it random, or does the choice of  ŋ (Ŋ) rather than say t (T) signify something technical?)