Template talk:inc-extension

@Bhagadatta I want to add more lines like

Ashokan Prakrit -𑀓- (-ka-) not later Maharastri Prakrit -𑀅- (-a-)

at Template:inc-extension/documentation. Should there be a discussion before putting more lines like that? Unless you can find extensions like Ashokan Prakrit -𑀓- (-ka-) in the inscriptions, wouldn't extensions like Ashokan Prakrit -𑀓- (-ka-) be reconstructions without the * symbol for convenience? Kutchkutch (talk) 12:01, 2 September 2020 (UTC)Reply

@Kutchkutch: I believe you mean they should be reconstructions with the asterisk if they're not attested through inscriptions? But I think, these particles are better off without the * symbol. But in spite of that, I wouldn't push for an asterisk to be affixed to that particle.

However, the wiktionary practice gives these non lemma forms an asterisk too. For instance, in the etymology of उग्र (ugrá), we write: PIE *h₂ug-ró-s, zero grade of *h₂ewg- + *-rós.

But Ashokan Pkt's case is not entirely same as PIE. PIE is entirely unattested but Ashokan is partially. A word like kataïllaka as a whole is unattested. But two of its components seem to be, viz. kata and -ka (the latter having been used in many other forms too). Sooo..? Maybe add the asterisk to the part that is not attested? I know this appears to be rather muddled but we are treading in a territory hitherto unexplored at least on the Wiktionary and on this scale. Our going into this much detail may seem like overdoing it for now but in the years to come we will set standards for Middle Indo-Aryan linguistics at wiktionary. -- Bhagadatta (talk) 15:14, 2 September 2020 (UTC)Reply

@Bhagadatta: The wording of the question about the * symbol was awkward but you understood what I meant. The comparison with PIE non-lemma forms is helpful, and the reason for asking these questions is exactly as you put it:

"We are treading in a territory hitherto unexplored at least on the Wiktionary and on this scale. Our going into this much detail may seem like overdoing it"

It's really easy to start making stuff up when "treading in a territory hitherto unexplored", especially with reconstructions using these extensions. Since you have experience with this when reconstructing PIA/PIIR/PIE, it's insightful to to hear what you think. The literature about Old Marathi discusses these extensions when explaining how MIA becomes NIA. Kutchkutch (talk) 07:37, 3 September 2020 (UTC)Reply

@Kutchkutch: Is the Old Indo-Aryan ancestor of MIA -illa- (in *kataïllaka) known? Or is this component (-illa-) used in any attested Ashokan Prakrit form? If it is used in any attested form then we need not use the asterisk for that component even in reconstructions. So I think -ka- does not need a * symbol as it is attested both in Old and Middle IA. A component like -illa- may need to be treated differently if its use was never recorded.. -- Bhagadatta (talk) 09:31, 3 September 2020 (UTC)Reply

@Bhagadatta, Kutchkutch: MIA -illa-, -ulla-, -alla have been connected by modern scholars (e.g. {{R:inc:Zoller:2016}}) to Proto-Indo-European *-ilo-, *-ulo-, *-elo, which produced verbal gerundives in Tocharian, various verbal forms in Anatolian (Hittite and Lydian) and Slavic. Interestingly, there's no known OIA form; some suffix like that may be present in Sanskrit बहुल (bahulá) but it's not as widespread as in MIA. —Aryaman^{A (मुझसे बात करें • योगदान)} 13:55, 3 September 2020 (UTC)Reply

Also, I think not having the reconstruction asterisk is fine, since these affixes were attested in MIA words and not totally made up by us. Really, it's what we do with every affix in every attested language; if it occurs commonly as a clearly distinct morpheme we treat it as attested. —Aryaman^{A (मुझसे बात करें • योगदान)} 13:56, 3 September 2020 (UTC)Reply

@AryamanA: Wow {{R:inc:Zoller:2016}} is some really good stuff! -- Bhagadatta (talk) 14:14, 3 September 2020 (UTC)Reply

@Bhagadatta: I am especially interested by his claims of a large number of NIA vocabulary (at the fringes of IA) that can be tied to PIE without any known OIA intermediaries. His Indus Kohistani dictionary also lists a lot of connections to OIA (and modern IA cognates) that update Turner's work. Probably, continuing work on Dardic will be very useful for us. He's publishing "Indo-Aryan and the linguistic history and prehistory of North India" soon which will probably be very cool! —Aryaman^{A (मुझसे बात करें • योगदान)} 14:43, 3 September 2020 (UTC)Reply

@AryamanA: Thanks for chiming in and sharing the paper. It is certainly interesting (and even mentions Old Marathi among other things). After your thoughts on the reconstruction asterisk, the consensus would be to continue omitting it.

@Bhagadatta: Despite the consensus to omit reconstruction asterisk, you make an intriguing point with your question. The only way to discover whether these affixes/extensions are attested in Ashokan Prakrit are to become more familiar it. Corpus linguistics on Ashokan Prakrit might be one way to do this. Perhaps 𑀜𑀸𑀢𑀺𑀓 (ñātika) proves the existence of -𑀓- (-ka-) in Ashokan Prakrit.

Unless there is an error in assigning Prakrit lects by Pischel, Sauraseni Prakrit 𑀧𑀸𑀟𑀮𑀺𑀧𑀼𑀢𑁆𑀢𑀅 (pāḍaliputtaa) is an exception to the Ashokan Prakrit -𑀓- (-ka-) → Sauraseni Prakrit -𑀕- (-ga-) generalisation (see 𑀧𑀸𑀝𑀮𑀺𑀧𑀼𑀢 (pāṭaliputa)). However, one exception is not enough to challenge the generalisation . Kutchkutch (talk) 04:58, 5 September 2020 (UTC)Reply

@Kutchkutch: I agree. This may be just one exception which may have happened because of influence from Maharashtri, or maybe the sound laws worked slightly differently in unusually polysyllabic words like the Ashokan Prakrit word pāṭaliputtaka with its six syllables. -- Bhagadatta (talk) 05:35, 5 September 2020 (UTC)Reply

@Bhagadatta, Kutchkutch: See these - आलू#Etymology 1, हीरा#Etymology, मुंडा#Etymology. I've edited these etymologies and added from {{inh|hi|sa|<term>क}}, from {{m|sa|<term>}} {{inc-ext|hi|ka}} because I think these are from Sanskrit आलुक, हीरक, and मुण्डक respectively, rather than आलु, हीर, and मुण्ड because the ā of Hindi comes from Sanskrit -क. In the discussion above, it's said that the "ka" extension is of Ashokan Prakrit, so can it be used beside Śaurasenī or Skt.? Also, I don't think Sanskrit word being extended with MIA extension makes a lot of sense. I think there should be an extra parameter which hides the "Middle" leaving only Indo-Aryan ka, or which replaces it with Old-Indo-Aryan (+in Devanagari script). Categorisation wouldn't have to change since it is only "Indo-Aryan" in the categories and not MIA (although we may think of Brahmi -> Devanagari ). What are your thoughts? 🔥शब्दशोधक🔥 16:00, 29 January 2021 (UTC)Reply

And yes, I changed those etymologies because of this edit of @AryamanA. 🔥शब्दशोधक🔥 16:03, 29 January 2021 (UTC)Reply

@शब्दशोधक: The -ka extension differs from the others because it is not a Middle Indo-Aryan innovation. It existed in the earliest Vedic Sanskrit and also in Proto IIR because we have cognate formations in Avestan and Old Persian too. So in cases where the term with the -ka extension is attested in Sanskrit, there should be no MIA extension mentioned in my opinion. I'll let you do all the edits to those pages you mentioned now. -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 16:28, 29 January 2021 (UTC)Reply

@Bhagadatta: Yeah, that's what - from Sanskrit हीरक, from हीर + Middle-Indo-Aryan -क seems a bit weird, doesn't it? An additional parameter to hide the word "Middle" should be fine then? And (may be obvious, but I don't know about that), why do we use Brahmi script for Prakrits? Devanagari would be much more convenient, I think. 🔥शब्दशोधक🔥 16:38, 29 January 2021 (UTC)Reply

@शब्दशोधक: With Hindi words ending that were extended with the suffix -ka, there are two possibilities: One, the -ka suffix existed in Sanskrit and is attested and two, the Prakrits took a Sanskrit word and attached a -ka suffix to it and the Hindi term came from there. In the first case, I don't see why it should be templated at all. If the term with the -ka suffix in Sanskrit is attested, then the further info should be in the Sanskrit entry where again a simple {{m|sa|term}} + {{m|sa|-क}} will suffice because you can rarely tell if the -ka suffix was added at the Sanskrit level or at an earlier level (Proto-Indo-Aryan, Proto-Indo-Iranian etc). In the second case (where the word with the -ka suffix is not attested in Sanskrit), you CAN have From Sanskrit + Middle Indo-Aryan -ka because it implies that a Sanskrit term was taken and a -ka suffix was added in the MIA level. So no edits are required to the template. That also means that in the above entries you mentioned, you should remove the MIA extension template because the -ka suffix was attested at the Sanskrit level (I'm too lazy to check now, I'll assume they all are attested, I'll let you check and do the edits). -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 16:58, 29 January 2021 (UTC)Reply

PS: About the script for Prakrits, Prakrit entries used to be in Devanagari until the Brahmi script was standardized for Prakrit in 2017 because they were originally used for Prakrit. I personally would not mind having alternative script form entries for Prakrit in Devanagari because it would help people find these entries on wiktionary better (people would be more likely to search in Devanagari). Let's see what we can do after the merger. -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 17:03, 29 January 2021 (UTC)Reply

@Bhagadatta: Ok, but what about the categorisation, then? Is it fine if we don't categorise them into Category:Hindi terms extended with Indo-Aryan -𑀓-? And where should all the descendants given here should be - at हीरक or हीर (I'm going to add the Sanskrit entry soon there)? 🔥शब्दशोधक🔥 04:30, 30 January 2021 (UTC)Reply

@शब्दशोधक: Category:Hindi terms extended with Indo-Aryan -𑀓- will contain those Hindi terms which come from a Sanskrit term where the -ka suffix is not attested. For example: Hindi गूढ़ा (gūṛhā) will be {{inh|hi|sa|गूढ}} + MIA -ka because *गूढक (gūḍhaka) does not exist in Sanskrit (the word is there in one dictionary where it means something else and might have been coined later). For Hindi terms that come from Sanskrit words with and attested -ka suffix, why is categorization required? The suffix was not added at an MIA level. So Hindi आलू (ālū) will be simply {{inh|hi|sa|आलुक}}.

The descendants of हीरक will be at हीरक and हीरक will be in the derived terms section of हीर. -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 05:01, 30 January 2021 (UTC)Reply

@Bhagadatta: And regarding doublets - should we consider आलु, आलू and आलुक (and also हीरा, हीर and हीरक) Hindi doublets? Also, I think Hindi hīr is a learned borrowing from Sanskrit हीर (hīra) as R:CDIAL doesn't mention Hindi hīr; it only mentions Hindi hīrā. R:hi:McGregor says , but is Pa. Prakrit here (generally Prakrit is Pk. in it)? 🔥शब्दशोधक🔥 05:49, 30 January 2021 (UTC)Reply

@शब्दशोधक: They are distinct doublets, yes. In the Oxford Hindi dictionary, the abbreviation Pa. is Pali as given in page 18 of the introduction section of that dictionary. No idea why Pali was mentioned here, maybe for the sake of comparison. It is borrowed from Sanskrit hīra and Pali has nothing to do with it. -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 07:04, 30 January 2021 (UTC)Reply

@Bhagadatta: Yeah, even I thought it may be related to "Turner's "Pa."" or Pali, so thanks for clarifying. So our problem is solved now and I know when to use MIA -क and when not. 🔥शब्दशोधक🔥 07:28, 30 January 2021 (UTC)Reply

@Bhagadatta, शब्दशोधक: Regarding -ka:

Now that this has been discussed, I agree with the consensus of only using {{inc-ext}} if

the -ka suffix is not attested in Sanskrit

and

the Sanskrit term was taken and a -ka suffix was added in the MIA level

Would the same reasoning apply to

CAT:Terms extended with Indo-Aryan -𑀇𑀓𑀸- vis-à-vis -इका (-ikā). Such as:

Marathi चांदी (cāndī) and Hindi चाँदी (cā̃dī) and

Marathi साडी (sāḍī) and Hindi साड़ी (sāṛī)?

CAT:Terms extended with Indo-Aryan -𑀉𑀓𑀸-. Such as:

Marathi वाळू (vāḷū) and Hindi बालू (bālū)?

Regarding Prakrit in Devanagari:

I created Devanagari → Brahmi redirects (such as गुज्जरत्ता (gujjarattā), पसरइ (pasaraï), अक्खइ (akkhaï)) because the documents containing the quotations are in Devanagari. I also wouldn't mind expanding the idea of having Devanagari → Brahmi redirects for any Prakrit entry

because it would help people find these entries on wiktionary better (people would be more likely to search in Devanagari)

Perhaps it might help to have a Prakrit equivalent of T:pi-alt and T:sa-alt, but we would have to determine if there should any scripts in addition to Brahmi and Devanagari. मागधी is written in 8 scripts at Magadhi Prakrit. Kutchkutch (talk) 09:17, 30 January 2021 (UTC)Reply

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Yeah, it would for Skt. -ikā also. 🔥शब्दशोधक🔥 09:29, 30 January 2021 (UTC)Reply

@Kutchkutch: Yes, I think no Sanskrit suffix should be called an MIA suffix unless it can be established that the Sanskrit term with suffix was coined at a much later age. -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 11:25, 30 January 2021 (UTC)Reply

@Bhagadatta: How should the etymology be here? Sanskrit अर्धक (ardhaka) is an attested word, but then will we have to reconstruct the Sauraseni term like अद्धग? 🔥शब्दशोधक🔥 16:51, 1 February 2021 (UTC)Reply

@शब्दशोधक: Yes you can but it's optional. You can also directly say it's from अर्धक. -- 𝓑𝓱𝓪𝓰𝓪𝓭𝓪𝓽𝓽𝓪^{(𝓽𝓪𝓵𝓴)} 04:48, 2 February 2021 (UTC)Reply