Module talk:sa-pronunc/sandbox

Classical Sanskrit pronunciation
, hey all, I'm trying to refactor this module to show more pronunciations (Pāṇini's, Śaunaka's, reconstructed, recitational, Classical). It's gonna be some work, but before I get to it, I've been trying to figure out how the stress accent in Classical Sanskrit works. There's much agreement that it's like an extended version of the Dreimorengesetz (it can get all the way to the preantepenult), but I don't know how this works in sandhi situations. Do I have to calculate the stresses before external sandhi? For instance, in phrases like, the classical rendering would be stressed /ˈtatˈtvaˈma.si/ (or /ˈtat ˈtvam ˈa.si/?), but would be /ˈso.ham/ (The phonetic values aren't correct; I'm just showing the stress placement). For, I assume that the two words' stress coalesce, but I can't find any sources for this. And what about something like +  =. Is that /ˈnvaˈgac.cham/, even though the stressed "u" in nu has been sandhied away? I'd love to know what y'all think or if you have any sources for this. —*i̯óh₁nC[5] 10:43, 19 December 2017 (UTC)
 * from what I've read, the stress falls on "heavy" syllables and nowhere is any mention of sandhi made. In तत्त्वमसि is the first syllable not heavy? The o in Sanskrit is always long or is technically a diphthong so /ˈso.ham/ makes sense. In न्वगच्छम् I can't find a heavy syllable (as the aspirate ch is considered to be one single sound rather than c + h). This would place the stress on the 1st syllable I'd guess. -- माधवपंडित (talk) 16:52, 19 December 2017 (UTC)
 * A few things, for, the second syllable would be stressed because gac.ch is a heavy syllable. More generally, my question was about multiple words sandhied together. Separately, contains three separate words, each of which would normally get a stress. I find it hard to believe that when put together, only one of the three words would receive stress. I'm certain that the heavy syllable rule only works within a single word before sandhi. Let me cast this net slightly wider: . —*i̯óh₁nC[5] 23:12, 19 December 2017 (UTC)
 * I always thought sandhi reflected how the word would be pronounced, without adjacent vowels across morphemes being merged, so I would assume there would be only one stress. Anyways I would do /'tat.va.ma.si/ and /nva'gac.cham/. Perhaps is a different case from  since  undergoes no changes with sandhi (but I'm just speculating). And there could also be the possibility of secondary stress, like in Hindi (which has a nearly identical system for syllable weight and stress as Classical Sanskrit). FWIW, The Cambridge Introduction to Sanskrit's chapter on syllable weight and stresses seems to de-sandhi the text when putting stresses. —AryamanA (मुझसे बात करें • योगदान) 23:38, 19 December 2017 (UTC)
 * So, at least in Vedic Sanskrit, a continuous block of characters could have multiple pitch accents (like tā́nī́dáhāni for tā́ni íd áhāni, RV VII.7.3a). There's no notion that the sandied words are not separate morpho-phonological units. As far as I can tell Classical Sanskrit should function similarly (i.e. each morphological word should receive an stress). —*i̯óh₁nC[5] 06:56, 20 December 2017 (UTC)
 * I got pinged, but I know nothing at all about stress in Classical Sanskrit, sorry. —Mahāgaja (formerly Angr) · talk 10:05, 20 December 2017 (UTC)

So would the module have to be given the individual words (like तत् त्वम् असि)? I feel like IPA for multiword entries is unnecessary. —AryamanA (मुझसे बात करें • योगदान) 15:03, 24 December 2017 (UTC)
 * I was thinking the usage could be something like tát would be the technique. Alternatively, tát tvám ási could work. Regardless, I was hoping to be able to handle multi-term inputs for phrases, etc. We'd have to implement the sandhi ourselves, given the nú + agáccham example above. —*i̯óh₁nC[5] 23:12, 24 December 2017 (UTC)