Dear All,

The verse  (with some minor variations) usually cited in grammatical works is:

pade tu saṃhitā nityā nityā dhātūpasargayoḥ |
samāse ca tathā vākye sā vivakṣām apekṣate ||

The text given is the earliest attestation I have found, the Rūpāvatāra (p. 6, beginning of the Saṃhitāvatāra).

The optional sandhis are those at word boundaries, including between members of a compound, but not within a word or between a root and its upasarga. This maxim is at times invoked to explain irregular sandhis in Epic/purāṇic texts. Avivakṣā, however, is not a valid justification for failing to apply sandhi in verse according the authorities on poetry. Cf. Kāvyādarśa 3.159 (Dimitrov's edition).

A look at the various discussions of visandhi, a poetic blemish, will also reveal a number of interesting discussions about the application of sandhi in verse. For example, in commenting on KĀ 3.161, Ratnaśrījñāna explains that the lack of sandhi between the 2nd and 3rd pādas in KA 3.161 (api / āsu) should not be viewed as fault. He further bolsters his/Daṇḍin's view by citing a verse from the Abhijñāśākuntala 1.19 (tanoti  /  iyam). Similar statements are found in other works. There is also a rather lengthy back and forth between ālaṅkārikas over the ages concerning the correctness of using perfectly valid pragṛhya sandhis (duel endings in ī, ū, and e follwed by a vowel) ,but it is perhaps not of immediate interest.  

All the Best,
Victor D'Avella


On Thu, Jul 28, 2016 at 4:46 PM, Madhav Deshpande <mmdesh@umich.edu> wrote:
Yes, Jonathan, the external sandhi is optional according to grammarians, and yet the convention of metrical Sanskrit is that the sandhis are pretty much obligatory.  The break in Sandhi comes after the half-verse or ardha-ṛc in Vedic verses.  Only rarely is there a sandhi break within a metrical line.  On the other hand, there are hyper or double sandhis occasionally in older metrical texts, e.g. saḥ + eṣaḥ > sa eṣaḥ, but metrically occasionally one notices saiṣa, and Pāṇini allows such double sandhis for filling the metrical space (so'ci lope cet pādapūraṇam).  Breaking the sandhis in metrical lines is done as part of interpretation, but not as part of recitation, since with broken sandhis, the metrical lines are no longer complaint with rules of a specific meter.

Madhav Deshpande

On Thu, Jul 28, 2016 at 10:30 AM, Jonathan Silk <kauzeya@gmail.com> wrote:
I hardly dare to comment when my teacher, Madhav Deshpande, is on this list as well, and what little I know I know from him, but.... I recall very well learning that external sandhi is, according to the grammarians, always optional.
Because
One
Can
Always
Speak
Like 
This
If 
One 
Wants
To.
In other words, the use of sandhi is a convention, so the question might be slightly rephrased as: what are the conventions of the poets, and of the scribes. No?
(Perhaps, as is quite likely, of course, this was implied in the question, and I should have kept my ideas to myself.
In
Which
Case,
Sorry ;)

Jonathan

On Thu, Jul 28, 2016 at 4:26 PM, Valerie Roebuck <vjroebuck@btinternet.com> wrote:
Correction: I meant ‘at the end of one line’.

> I’ve just had a quick look at an edition of the Saundaryalaharī, in Śikhariṇī (17 syllables to a line), and sandhi is broken only between half verses and whole verses. For example, there's a ś  at the end of one half-line followed by a c at the beginning of the next.

Valerie J Roebuck
Manchester, UK

> On 28 Jul 2016, at 15:07, Harry Spier <hspier.muktabodha@gmail.com> wrote:
>
> Dear list members,
>
> In the longer sanskrit meters (Vasantatilaka for examplbe 14 syllables to a line) is Sandhi broken after each line or only after the half verse and end of verse.
>
> Thank you,
> Vasishtha
> _______________________________________________
> INDOLOGY mailing list
> INDOLOGY@list.indology.info
> indology-owner@list.indology.info (messages to the list's managing committee)
> http://listinfo.indology.info (where you can change your list options or unsubscribe)


_______________________________________________
INDOLOGY mailing list
INDOLOGY@list.indology.info
indology-owner@list.indology.info (messages to the list's managing committee)
http://listinfo.indology.info (where you can change your list options or unsubscribe)



--
J. Silk
Leiden University
Leiden University Institute for Area Studies, LIAS
Matthias de Vrieshof 3, Room 0.05b
2311 BZ Leiden
The Netherlands


_______________________________________________
INDOLOGY mailing list
INDOLOGY@list.indology.info
indology-owner@list.indology.info (messages to the list's managing committee)
http://listinfo.indology.info (where you can change your list options or unsubscribe)


_______________________________________________
INDOLOGY mailing list
INDOLOGY@list.indology.info
indology-owner@list.indology.info (messages to the list's managing committee)
http://listinfo.indology.info (where you can change your list options or unsubscribe)