Talk:बद्ध

Ashokan Prakrit

@AryamanA, Kutchkutch Hi, I just wanted to know if you have any objections if the Prakrit descendants (all of them 𑀩𑀤𑁆𑀥 (baddha)) are derived not from Ashokan 𑀩𑀥 (badha) but from *𑀩𑀤𑁆𑀥 (*baddha) which clearly is the older Indo-Aryan form. We could also apply the same policy to some other terms where the Prakrit descendants are closer to the Sanskrit form than the (attested) Ashokan term is. There are a bunch of such entries; Ashokan 𑀧𑀼𑀮𑀺𑀲 (pulisa) is one I can think of now. So there too, we could derive Maharashtri 𑀧𑀼𑀭𑀺𑀲 (purisa) from Ashokan *𑀧𑀼𑀭𑀺𑀲 (*purisa) instead of 𑀧𑀼𑀮𑀺𑀲 (pulisa). -- Bhagadatta (talk) 02:02, 2 September 2020 (UTC)

How well does the Ashokan spelling reflect geminates? I recall claims that geminates were generally not written as such in the earliest Brahmi, which to me would imply that Ashokan 𑀩𑀥 (badha) is phonetically ambiguous between /badʰa/ and /baddʰa/. This is rather like the earliest Latin not recording double consonants. If I have this right, the correct approach would be to grit our teeth and start recording broad 'standard' pronunciations. --RichardW57 (talk) 05:54, 2 September 2020 (UTC)

If we follow the Wiktionary links for 'Ashokan Prakrit', the only mention of it we find is 'the language of Ashoka's inscriptions'. It seems to be a chronological stage rather than a real language, with differences being explained as adjustments to local dialects (Ashoka had ensured they didn't have armies so as to make them languages). For the purisa/pulisa difference, postulating multiple forms is probably the only way out of the problem of the non-descent of the alleged daughter languages. --RichardW57 (talk) 05:54, 2 September 2020 (UTC)

@Bhagadatta: For 𑀩𑀥 (badha) vs *𑀩𑀤𑁆𑀥 (*baddha): The issue stems from the Brahmi script.

Despite the impression of uniformity given by the Unicode encoding and the available fonts (Noto Sans Brahmi and Segoe UI Historic), there are actually various stages of the Brahmi script as outlined at Brahmi script#Evolution_of_the_Brahmi_script: Early, Middle and Late. Ashokan Prakrit uses the earliest stage of the Brahmi script: Early Brahmi. Early Brahmi did not have a have a mechanism for expressing gemination even if gemination was present. The gemination is reconstructed based on the etymon बद्ध (baddha) and the Dramatic Prakrit descendants (Maharastri, Sauraseni, Magadhi): 𑀩𑀤𑁆𑀥 (baddha).

Initially, en.wikt used the Devanagari or Latin scripts for the Dramatic Prakrits since these are the scripts that are more commonly used today, but User:DerekWinters persuaded User:Aryamanarora to use Brahmi (see this discussion) since Devanagari and Latin are just modern conveniences for the original script(s): Middle Brahmi or Late Brahmi (after Early Brahmi).

The mechanism for expressing geminates using the virama/halant and conjunct characters was introduced after Early Brahmi. The later Brahmic scripts are somewhat unusual since they indicate details such as gemination, which may not be as common in other scripts just like User:RichardW57 alludes to with Latin. This is somewhat comparable to how Old Persian cuneiform doesn't accurately record all details, and the {{m|inc-ash|ts=}} parameter is used to show the reconstructed transcription. So to resolve the 𑀩𑀥 (badha) vs *𑀩𑀤𑁆𑀥 (*baddha) issue, perhaps the {{m|inc-ash|ts=}} parameter on 𑀩𑀥 (badha /⁠baddha⁠/) should be used to show the gemination instead of reconstructing *𑀩𑀤𑁆𑀥 (*baddha).

However, when reconstruction is necessary such as *𑀓𑀼𑀢𑁆𑀢 (*kutta), the gemination should be indicated with the virama/halant and conjunct characters since the Brahmi script is being used to represent the spoken 'Proto-Prakrit' rather than written Ashokan Prakrit. *𑀣𑀧𑀢𑀺 (*thapati) is an exception to this. Since User:AryamanA has been the primary creator of these 'Proto-Prakrit' entries thus far, his opinion is needed on whether to standardise the presence or absence of Brahmi gemination in 'Proto-Prakrit'.

For 𑀧𑀼𑀮𑀺𑀲 (pulisa) vs *𑀧𑀼𑀭𑀺𑀲 (*purisa): The issue stems from the implications of the political expansion of Magadha kingdom.

Among the Mahajanapadas, the Magadha became so large and powerful that people who spoke other dialects in the surrounding kingdoms hypercorrected their speech to match the Magadhan dialect. The merger of /l/ and /r/ → /l/ was a feature of the Magadhan dialect. Thus, people who spoken other dialects in the surrounding kingdoms, also acquired the merger of /l/ and /r/. The Prakrits spoken further west and south, such as the Ashokan ancestor Maharashtri Prakrit, were so far from Magadha that they were less affected by these 'Magadhisms'. The Ashokan ancestor of Sauraseni Prakrit probably had this merger but lost it after the decline of the Maurya Empire.

User:AryamanA's opinion would be helpful here as well to resolve the 𑀧𑀼𑀮𑀺𑀲 (pulisa) vs *𑀧𑀼𑀭𑀺𑀲 (*purisa) issue. User:RichardW57 has used the biological term atavism to denote this undoing of Magadhan influence outside Magadha. Linguistic purism is another term. Perhaps one solution would be to use an existing or create a new CAT:Foreign derivation templates for this atavism. Kutchkutch (talk) 08:26, 2 September 2020 (UTC)

@Kutchkutch, RichardW57: Cool, thanks both. Kutchkutch has made this quite clear. I do recall someone, maybe DerekWinters mentioning before that germinates were not written as such and often used a single character which was also used for non germinates. It just so happened that I momentarily forgot it and was looking at the Brāhmī through a Devanāgarī tinted lens: I attribted to the Brāhmī the rigour and consistency of the Devanāgarī and concluded that the word being spelled as "badha" meant that the word itself was badha. I like the suggestion of entering "baddha" into the ts parameter and will do so right away. For pulisa/purisa, I think we can say that merger of l and r never completely occurred in Ashokan. If it were true, we will then be saying that the original form of every Maharashtri/Sauraseni etc Prakrit had an /l/ which later got Sanskritized and adopted the /r/ to match Skt /r/ and only those forms with a corresponding Skt /l/ got to keep the /l/.

Why not something like this?

{{desc|inc-ash|𑀧𑀼𑀮𑀺𑀲}}, {{l|inc-ash|*𑀧𑀼𑀭𑀺𑀲}} ({{q|< *पुरिष (puriṣa)}})

So basically, keep the Ashokan pulisa but put a *purisa next to it. Mention that they come from Skt. *puriṣa. This way we can account for every last descendant. -- Bhagadatta (talk) 10:01, 2 September 2020 (UTC)

@Bhagadatta: For purisa, I agree with this solution but it seems to contradict the notice at WT:About Sanskrit.

A related issue is how to treat 'inscriptional Prakrits' that are not Ashokan. 𑀤𑀔𑀺𑀦𑀸𑀧𑀣𑀧𑀢𑀺 (dakhināpathapati) is perhaps the only example so far. Like dakkhiṇāpatha, the 𑀔 (kha) actually represents the geminate 𑀓𑁆𑀔 (kkha). This can be addressed with the {{m|inc-ash|ts=}} parameter that you have just endorsed. Also, 𑀧𑀇 (paï) is the standardised Maharastri descendant of पति (páti) rather than 𑀧𑀢𑀺 (pati). So 𑀧𑀢𑀺 (pati) is either atavism for पति (páti) or the 𑀢 (ta) → ∅ change had not happened yet in this variety. Kutchkutch (talk) 10:46, 2 September 2020 (UTC)

@Kutchkutch: For 𑀧𑀢𑀺 (pati), maybe it was borrowed from Sanskrit or inherited from a variety where the change hadn't yet occurred as you said.

How does showing the form *पुरिष (puriṣa) contradict the notice? It says that Sanskrit is being taken to be equal to the Old Indo-Aryan dialect continuum just like Ashokan is considered to be equal to the MIA dialect continuum. That does not mean we cannot show unattested varieties of attested words. झापयति is another example, it has a reconstruction entry here. It means that this variety existed but was not recorded anywhere and thus did not make it to the standardized Epic/Classical Sanskrit. -- Bhagadatta (talk) 10:56, 2 September 2020 (UTC)

"{{l|inc-ash|*𑀧𑀼𑀭𑀺𑀲}} ({{q|< *पुरिष (puriṣa)}})" would be untrue. The /r/ is a conservative pronunciation (in these cases), not an innovation from Sanskrit. Do we have an identification of 'Ashokan {Prakrit)' anywhere? The Wikipedia link is one of many unhelpful redirects littering Wikipedia. Is Ashokan a purely inscribed language? --RichardW57 (talk) 11:03, 2 September 2020 (UTC)

Well, I meant *पुरिष as opposed to the attested पुरुष (puruṣa); I didn't mean *पुरिष as opposed to the also hypothetical *पुलिष (puliṣa). I know pulisa was an innovation and purisa is conservative. -- Bhagadatta (talk) 11:34, 2 September 2020 (UTC)

@Bhagadatta: There's no contradiction if what you mean by *Reconstruction is a redirect to Attested form. Thanks for the clarification. Kutchkutch (talk) 11:50, 2 September 2020 (UTC)

Ah, I think we've hit the reason we're not supposed to use notations like '<' in Wiktionary. What you mean, I'm glad to now see, is "{{l|inc-ash|*𑀧𑀼𑀭𑀺𑀲}} ({{q|inherited from *पुरिष (puriṣa)}})" as opposed to "{{l|inc-ash|*𑀧𑀼𑀭𑀺𑀲}} ({{q|borrowed from *पुरिष (puriṣa)}})". I used to be familiar with a longer symbol for inheritance than for borrowing. However, what is the qualifier supposed to be qualifying? 𑀧𑀼𑀮𑀺𑀲 (pulisa) is also inherited from *पुरिष (puriṣa)}!

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ @Bhagadatta, Kutchkutch, RichardW57: First, about the Brahmi: Ashokan Prakrit does not indicate geminates in the way modern Brahmic scripts do. We should treat it as a normal part of the inheritance chain either way. I support using |ts= in these cases, but it should be noted we don't really know anything about the phonology of MIA. For example, the Dramatic Prakrits consistently use retroflex ṇ, but that probably wasn't spoken as a retroflex. So we should be careful making assumptions (but this one about geminates is probably safe).

Second, I will expound on my thoughts that have been somewhat already expressed in regards to "inherited from Sanskrit". The status of the MIA languages is (1) we only have textual attestions, and (2) those texts only represent a few, presumably standardized and somewhat artificial even, MIA dialects. Now, just like with Sanskrit, just because a word is not attested in the exact form that is reflected in a descendant, doesn't mean that form did not exist at that time. Very little (compared to modern languages and even to Sanskrit) is written down from the MIA stage, so we should be liberal with reconstructing at that stage especially when the OIA form is attested. So I espouse a format like:

From Ashokan Prakrit *𑀧𑀼𑀭𑀺𑀲 (*purisa), 𑀧𑀼𑀮𑀺𑀲 (pulisa), from Sanskrit *पुरिष (puriṣa), पुरुष (púruṣa).

(This probably deserves some more etymological notes, since it's an interesting IA innovation.) The advantages of this format are that it's not too cluttered, every form is mentioned, and all the information on whether a form is attested or not is visible at a glance. I already do this for Hindi in cases where the best form in intermediary Sauraseni is not attested but a different form is. While the mention of an atavism is a good idea, I think that should follow in a separate paragraph in the etymology rather than in the chain of inheritances itself, to reduce clutter.

In general, we should treat Ashokan Prakrit as a "Proto-Early-MIA" parallel to what (Vedic) Sanskrit is being treated as. I think this stage is extremely important, since when we have widespread reflexes in NIA, too often we don't have the necessary comparative data (that is best supplied by Dardic usually) to reconstruct all the way back to Sanskrit. Too many cluster assimilations, consonant droppings, and other irrecoverable changes have occurred. As Ashokan Prakrit is being used as a general Proto-Early-MIA label, we should stick with representing geminates etc. as they would have been phonologically. It is not useful to reconstruct with the idiosyncracies of Ashokan Brahmi being applied, in comparative linguistics we care about the phonology not orthography.

I'll be starting a WT:About Ashokan Prakrit to ensure we have our standards written down. It would be good to have a w:Ashokan Prakrit too; I might start one once I have more time. —Aryaman^{A (मुझसे बात करें • योगदान)} 21:16, 2 September 2020 (UTC)

@RichardW57: That is exactly what I meant. I used that symbol because I've seen it being used in some linguistic papers to show derivation, not necessarily a borrowing. It can be misleading because wiktionary uses the arrow, albeit a very distinguishable one in an opposite direction, to show a borrowing. The qualifier template renders the text in italics and I've been using it ever since I saw it being used at *Háwǰas, where an Avestan term was listed as a descendant of *Háwǰas but it was made clear using a qualifier that the actual ancestor term of that Avestan word was the unpalatalized *Hawgás. So I propose to write "{{desc|inc-ash|𑀧𑀼𑀮𑀺𑀲}}, {{l|inc-ash|*𑀧𑀼𑀭𑀺𑀲}} ({{q|inherited from *पुरिष (puriṣa)}})" in the descendants section of the page for पुरुष (puruṣa). This will state that both pulisa and purisa are from *पुरिष (puriṣa).

@AryamanA: Very cool, thanks for making this clear. Happy to be on the same page! -- Bhagadatta (talk) 01:20, 3 September 2020 (UTC)