Talk:हड्डी

Words without compensatory gemination

@AryamanA There are a lot of such words in Hindi. Earlier, we were using {{bor+|hi|inc-nwe}} to indicate these words. It seems you've changed a lot of these.

I disagree with using Template:inherited on these pages. The template documentation says that inheritance is reserved for an "unbroken chain of inheritance from the source term in question" and should "should not be used for terms that were reformed morphologically during their history". In my view then, there should be a 1-to-1 between a Hindi inherited term and the Sanskrit etymon. In this example, the word हाड़ (hāṛ) exists ([1]) and is the "true" inherited descendant in Hindi according to expected sound changes. I do agree though that using {{bor+|hi|inc-nwe}} was a poor choice, because it oversimplified the fact that Hindustani came from a dialectal mixture and contains a large number of words with some features more typical of the Northwestern-zone than Central. I think a good solution would be to use {{der+|hi|inc-nwe}} (or at least {{der+|hi|pra}}, {{der+|hi|sa}}, ...).

This also touches on another point, which is that Hindi has a lot of words which were extended by + Middle Indo-Aryan -𑀓- (-ka-), etc, and in these cases based on the same logic above I'm inclined to say these are "derived" from Prakrit and Sanskrit as opposed to "inherited". We used "inherited" rather liberally on a lot of such pages (e.g. अंगूठा (aṅgūṭhā), which is from अङ्गुष्ठ (aṅguṣṭha) + Middle Indo-Aryan -𑀓- (-ka-)). There are lots of minimal pairs where one word was historically extended, e.g. दाँता (dā̃tā), and another was not, e.g. दाँत (dā̃t), and in my understanding only the latter has the right to be considered directly "inherited". The historically-extended words should generally be considered "derived" from Sanskrit since they underwent suffixation.

Pinging other Hindi contributors for thoughts @Benwing2 @Kutchkutch @Svartava Dragonoid76 (talk) 23:52, 25 October 2024 (UTC)

@Dragonoid76:

• I had asked AryamanA about this point on WT:Discord much earlier and he gave an excellent answer to this, but basically since Hindustani came from a dialectal mixture, there are often irregularities in inheritances (rather than indicating a borrowing from NWe) that shouldn't make us consider them not unbroken. Inherited: Denotes words which etymologically are not borrowings but derive through regular or sporadic sound change.
• As for words with extensions, I/we generally prefer just writing inherited for simplicity when the extensions like -ka-, -ikA- are used which are clearly found in Sanskrit or inherited from Sanskrit. My reasoning is that writing:

Inherited from Sanskrit अङ्गुष्ठ (aṅguṣṭha) + Middle Indo-Aryan -𑀓- (-ka-).

is the simpler way of writing

Inherited from Sanskrit *अङ्गुष्ठक (aṅguṣṭhaka), from अङ्गुष्ठ (aṅguṣṭha) + Middle Indo-Aryan -𑀓- (-ka-).

which legitimizes the inheritance since *अङ्गुष्ठक (aṅguṣṭhaka) is a valid Sanskrit reconstruction because of -ka- being clearly found in Sanskrit.

The first one can be written as a more crisp etymology for the same thing except just that the joint *अङ्गुष्ठक (aṅguṣṭhaka) is not written. – Svārtava (tɕ) 03:49, 26 October 2024 (UTC)

@Svartava See the point below, but basically in my view this should be more like

Derived from Prakrit 𑀅𑀁𑀕𑀼𑀝𑁆𑀞 (aṃguṭṭha​) + Middle Indo-Aryan -𑀓- (-ka-), from Sanskrit अंगुष्ठ (aṃguṣṭha). Here, we're using "derived" which seems like the correct idea here based on the definition of "inherited". And this etymology is quite succinct. Dragonoid76 (talk) 17:51, 26 October 2024 (UTC)

@Dragonoid76 inherited should not be used for terms that were reformed morphologically … Hindi has a lot of words … extended by Middle Indo-Aryan -𑀓- (-ka-), etc, and in these cases … I'm inclined to say these are "derived" … We used "inherited" rather liberally on a lot of such pages

Although extensions such as -𑀓- (-ka-) are technically affixes, their ubiquity and the lack of concrete definitions for them has led to the impression that they are a part of the inheritance process.

There are lots of minimal pairs where one word was historically extended, e.g. दाँता (dā̃tā), and another was not, e.g. दाँत (dā̃t), and in my understanding only the latter has the right to be considered directly "inherited"

Since these derivationally interesting, perhaps there should be some way to distinguish them.

I think a good solution would be to use {{der+|hi|inc-nwe}} (or at least {{der+|hi|pra}}, {{der+|hi|sa}}}}

Borrowing from another New Indo-Aryan a language should only be done in the most obvious cases where inheritance can be completely ruled out. So, {{inh+|hi|pra}}, {{inh+|hi|sa}} is preferable to {{bor+|hi|inc-nwe}} even for areal features as a result of a sprachbund.

Kutchkutch (talk) 03:53, 26 October 2024 (UTC)

@Kutchkutch @Svartava What is wrong with using {{der+|hi|pra}} and {{der+|hi|sa}} on these pages. The user sees "Derived from ...", which is perfect, and we don't have to make any claims about "inheritance" or "borrowing". Dragonoid76 (talk) 17:49, 26 October 2024 (UTC)

@Dragonoid76: Derived would basically make half the terms ending in -ā as not inherited, while these are in general studied as inherited/descended and thus are expected to be in Category:Hindi terms inherited from Sanskrit as tadbhavas. The alternative way would be to type out the extended term with the reconstruction asterisk, but that is clearly not quite succinct. – Svārtava (tɕ) 18:02, 26 October 2024 (UTC)

@Svartava I understand that view. But basically we have a situation where हाड़ (hāṛ), *हाड़ा (hāṛā), *हाड़ी (hāṛī), हड्डा (haḍḍā), हड्डी (haḍḍī) are all in Category:Hindi terms inherited from Sanskrit. Despite all having undergone very real diversions from the "expected" path, and it seems worthwhile to call this out.

I'd like to propose an alternative then. What if we have a category Called Category:Hindi tadbhavas (along with Category:Hindi tatsamas, ardhatatsamas, etc) which can accommodate all of these words regardless of suffixation or presence of Northwestern-zone sound change.

But Category:Hindi terms inherited from Sanskrit should be reserved for words that underwent expected Central zone sound changes. There's no need to say अंगूठा (aṅgūṭhā) and अंगूठी (aṅgūṭhī) are "inherited" from *अङ्गुष्ठक (aṅguṣṭhaka) and *अङ्गुष्ठिका (aṅguṣṭhikā). We need not make an assumption about a form that may or may not have existed in Old Indo-Aryan. We can say that the Hindi term is "derived" from अङ्गुष्ठ (aṅguṣṭh), but categorize it as a Tadbhava. Dragonoid76 (talk) 18:12, 26 October 2024 (UTC)

@Dragonoid76: Having other categories like Hindi tadbhavas is kind of impractical, and it is likely that it will not be well-maintained, leading to a messed situation. According to the definiton given at inherited, an irregular sound change shouldn't rule out an unbroken inheritance chain.

Turner writes out other extensions explicitly but never writes out -ka- or -ikA- given they are very common and productive in OIA as well (and to an extent, a part of how NIA inherits words - for example, without these, the feminine form of an OIA word would inherit into the same word as it's masculine form because of final vowel deletion), so writing अंगूठा (aṅgūṭhā) as "derived" rather than "inherited" from अङ्गुष्ठ (aṅguṣṭha) is unnecessarily pedantic. – Svārtava (tɕ) 18:52, 26 October 2024 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ @Dragonoid76

I'd like to propose an alternative then. What if we have a category Called Category:Hindi tadbhavas

Having other categories like Hindi tadbhavas is kind of impractical, and it is likely that it will not be well-maintained, leading to a messed situation.

• That has been tried before, and it did not work. See Category talk:Hindi Tadbhava
• In addition, such a concept would also have to encompass other Indo-Aryan languages by analogy, which may lead to a messy situation on a much larger scale.

The alternative way would be to type out the extended term with the reconstruction asterisk

• Having reconstructed terms either only in etymology sections or even full-fledged reconstructed entries is preferable to Category:Hindi tadbhavas
• However, the drawback of this approach is that reconstructions are not just one-offs is not favoured in Indo-Aryan historical linguistics except in cases in which there is no alternative approach. This can be deduced by the low number of reconstructed Proto-Indo-Aryan terms.

Kutchkutch (talk) 03:49, 27 October 2024 (UTC)

@Dragonoid76: Re: words lacking compensatory lengthening, these words are inherited. Let me explain why I believe this to be true. First, the lack of compensatory lengthening is actually almost entirely regular: monosyllabic words like हाड़ (hāṛ) undergo it, but words with more syllables generally do not, include हड्डी (haḍḍī), सच्चा (saccā), पत्थर (patthar), मक्खी (makkhī), अच्छा (acchā), etc. Additionally, many verbs fall into this class of words which we never called borrowed, e.g. रखना (rakhnā), सकना (saknā). The regularity of this change is strong enough that it is unfounded to say they are all loans. Additionally, this is all explainable by a regular law relating Apabhramsha to modern Hindi.

I agree that there are historical exceptions in Middle/Old Hindi like माखन (mākhan) (superseded by मक्खन (makkhan)); the argument I make in this regard is that modern Hindi, being founded on the dialect of Delhi, is a koiné that has levelled a great deal of variation in Hindi belt languages. Earlier in its history, the prestige lects of the Hindi belt were more to the east (e.g. Braj, Awadhi, even Maithili to an extent) and those lects have regular compensatory lengthening without exception. The more you go west, the less the tendency is to compensatorily lengthen (e.g. Punjabi doesn't do it at all). Delhi is sort of at the boundary, and as modern Hindi developed there (at a big urban centre where speakers of many lects co-existed) it mixed features of all these lects and that's why we end up with some exceptions to the rule and the fact the modern Hindi is less lengthening than older varieties of Hindi.

Finally I do want to cite page 232 of Kogan (2020), which agrees with me on this.

Re: on labelling Indo-Aryan extensions as non-inherited, I think we all agree that there doesn't seem to be an obvious rule governing when pleonastic suffixes were added to words, beyond that monosyllabic words are more liable to get extended than longer words. My argument for calling these terms inherited is that these suffixes are basically meaningless; they only change morphological features of a word (for nouns, declension pattern and/or gender) and not the meaning or part of speech. Due to that, I don't see them as normal derivational suffixes. Compare Telugu -ము (-mu) which gets added to Sanskrit neuter-gender borrowings: does added that suffix mean we have to use {{der|te|sa}} instead of {{bor}}? That doesn't make sense to me since it's merely adapting the word to Telugu morphology.

Additionally, we don't actually know at what stage of the inheritance chain the suffixes were added. The -ka and -ikā suffixes show up in words at least as far back as Proto-Indo-Iranian (maybe even PIE?) and almost all of the rest exist by Early MIA (Pali/Ashokan Prakrit). Modern languages are also free to add masculine -ā/-o and feminine -ī to words. So for any given words, the pleonastics may have been added at any time from PII to the present day. So etymologies could end being a helluva mess if we stop following the standard we have right now on pleonastics. —Aryaman^{A (मुझसे बात करें • योगदान)} 05:45, 27 October 2024 (UTC)

@AryamanA I suppose I understand the reasoning now. Thanks for the explanation. The journal article by Kogan that you linked is quite interesting. Dragonoid76 (talk) 06:37, 27 October 2024 (UTC)

If you're not convinced definitely say more! This is just what we arrived on based on earlier discussions. There are a bunch of such editorial issues that are a bit tough (both sides have some good arguments) but we have to choose something in order to maintain standardisation of our entries. —Aryaman^{A (मुझसे बात करें • योगदान)} 08:00, 27 October 2024 (UTC)

@AryamanA I'm convinced at the very least that it would be hard and messy to maintain a Category:Hindi tadbhavas distinct from Category:Hindi terms inherited from Sanskrit, or to change the etymologies of a large number of existing words. I agree that Middle Indo-Aryan -𑀓- (-ka-) and Middle Indo-Aryan -𑀇𑀓𑀸- (-ikā-) alone are special suffixes that are not necessarily derivational suffixes due to their ubiquity and having no real meaning (perhaps a diminutive meaning in some cases). I'm okay with {{inh+|hi|sa|अङ्गुष्ठ}} being a visual/maintenance preference over {{inh+|hi|sa|*अङ्गुष्ठक}} on अंगूठा (aṅgūṭhā). Perhaps changes in gender should be considered grounds for a breaking of inheritance? For example अंगूठी (aṅgūṭhī) has a real change in meaning from अङ्गुष्ठ (aṅguṣṭha) due to the Middle Indo-Aryan -𑀇𑀓𑀸- (-ikā-) diminutive suffix. -ई (-ī) is regularly used derivationally to form feminine equivalents of masculines, and so it is not a meaningless pleonastic suffix in many cases. Dragonoid76 (talk) 18:39, 27 October 2024 (UTC)