Template talk:pa-decl-noun

Disputed note

Hi @عُثمان - I hope you're well. I was hoping that I could remove the disputed template since it seems to be affecting every Punjabi lemma that includes the declensions, and also wanted to know how you think it should be resolved? نعم البدل (talk) 01:00, 10 December 2022 (UTC)

@نعم البدل Hi, I have removed it because when I added it I thought it would only appear on the template page.

I have a tenatative solution, if you are interested to help with this it would be greatly appreciated as it will require some work to get there.

I am thinking of using Wikidata Lexicographical Data to provide word-specific information about the forms. I have implemented an example noun declension table doing this here on pnbwiktionary. This same data can be used to produce a Gurmukhi or Shahmukhi table from the item at L729524 on Wikidata; in these case the forms being added manually means the conservative spellings could be accounted for.

I think the main issues are:

• How do we identify masculine nouns ending in -a which do not change in the plural direct form or singular oblique? There are only a few of these, but some are common, like بھرا and اللہ. I will check how these are treated after this; they should at least have their own tables somehow to keep them distinct.
• How do we handle singular-only and plural-only nouns? There are patterns with food words - for lack of a better name, I have made up the terms "saag gosht nouns" and "haldi chawal nouns." Saag gosht nouns never change form no matter what. There is no saagaan, this is impossible. Saag gosht nouns require measure words to be counted. For gosht, boti is the measure word. Haldi chawal nouns are either only singular if they are feminine or only plural if masculine. These require measure words as well. Some nouns like اجّ are singular only without being food related.
• Ablative, instrumental, vocative, and locative case forms should have some kind of attestation proving they have been used. We can only guess why some words have these and some do not, but it ultimately must be related to the underlying meaning.

I think these issues are where Wikidata can really come in handy. If you take a look at form number 5 at the bottom of نخرہ here, I have added an attestation from a poem that shows use of an ablative singular form. I did not find any for other additional case forms. A table drawing from this could include a footnote with a reference to this poetry anthology where one can find an example of the form being used. What would really be helpful is just adding more form-specific data like this, so that we have a more nuanced representation of the forms in use rather than applying the same assumptions everywhere. عُثمان (talk) 08:16, 10 December 2022 (UTC)

@نعم البدل Oh and I should show you this, you will find it interesting: https://www.wikidata.org/wiki/Lexeme:L741317

This is an old Punjabi declension pattern which has fallen out of use. Feminine nouns ending in ـُ -u changing to ـُو -ū in the plural form. This pattern has been preserved in a few Saraiki words in contemporary usage. It would be interesting to represent some of these older forms. عُثمان (talk) 08:21, 10 December 2022 (UTC)

@عُثمان: This is certainly a lot to take in. While I'm quite impressed with attaching declensions to the Wikidata, my issue with it is the fact that the declensions will have to be manually added to Wikidata and that for all the Punjabi lemmas - is it meant to be replacement of the templates (or is it for specific lemmas only?). Also you suggest attestation of other forms, which I'm not opposed to, but it's already difficult enough to attest any Shahmukhi lemmas, wouldn't that make a little difficult?

Also kudos for trying to edit or fix Punjabi Wiktionary! نعم البدل (talk) 13:19, 10 December 2022 (UTC)

@نعم البدل It is possible to automate the addition of forms to a certain extent, there are some tools that can be adapted for this. I am still trying out different approaches, since I think this will be necessary for verbs which get much more complex. There could even be a combination approach used:

• For the direct case and oblique case plural forms of a “regular” noun, the module could show the declensions based on what we would expect if there is no additional data available. What I mean by regular here is sort of nebulous, but you can kind of get a sense for whether we can expect these four forms to all be in use or not or if we have to check based on the meaning. For countable concrete objects, like wheels or buildings, we can expect at least all of these to be usable. For abstract nouns, it makes sense to double check. 99% of the time ਮਤਲਬ مطلب is an uncountable noun, but ਮਤਲਬਾਂ مطلباں in the plural oblique has actually been attested occasionally. I would not have guessed this without checking; it likely is limited to metaphorical or poetic uses but it exists nonetheless. For “important” words—kinship terms, religious terms, proper nouns, foods—more manual checking is warranted as well.
• If we do have additional data—alternate spellings, attestations, pronounciations, dialect indications, and so on—we can incorporate that from the Wikidata forms.
• I consider an attestation in either script to be an acceptable reference for a given form, each of which has spellings in both attached. Consider that the additional “synthetic case forms” as some writers describe them are in more widespread use in colloquial/spoken Punjabi than in writing, and ultimately we are representing the spoken language as well as these two written registers of it. Some forms, particularly vocatives for nouns, only appear in quotations and dialogue because they do not make much sense in prose. To that end, even citing use in songs or videos would work. For finding more Shahmukhi attestations, some of the classic poetry like Waris Shah can be helpful. (Admittedly, I find it harder to search for these without getting Urdu results mixed in, but much of the older poetry written in Shahmukhi has been transcribed to Gurmukhi, so you can even find the original by searching in Gurmukhi first.)

عُثمان (talk) 16:47, 10 December 2022 (UTC)

@نعم البدل To break this up into smaller parts we can start doing something about, I created this page for brainstorming: User:عُثمان/Punjabi declension paradigms . I will add more details when I have some time but the idea is just that each noun listed follows a different pattern, in forms available and/or in morphology. For example, we can consider گھر as representative of a group of nouns which decline like گھر and so on. Feel free to add any words you think may be a good example of a particular pattern. To start, we could just review what the tables for existing entries of these words look like and then go from there.

It looks like enwiktionary does not have Shahmukhi declension tables yet, so this is something we could implement to. عُثمان (talk) 21:03, 10 December 2022 (UTC)

@عُثمان: Btw Shahmukhi Punjabi does have declension tables. I'm not sure why they aren't appearing in Category:Punjabi noun inflection-table templates, but if you search "Templates:pnb-noun" then all the declension templates should come up! نعم البدل (talk) 10:57, 13 December 2022 (UTC)

@نعم البدل Ah, I see, good. Although the only one which appears for me is ساوا ; the others show red links. I will investigate to see what that is happening. I would also suggest one change with that table – for masculine plural oblique, leave out the zer as ساویاں and for the feminine plurals include just the zer without jazam as ساوِیاں as this is the convention employed by Salahuddin in Waddi Punjabi Lughat and it is a bit simpler. I have seen jazam applied more liberally in Urdu/Hindustani dictionaries but I have not really seen it used like this in Punjabi. عُثمان (talk) 16:53, 14 December 2022 (UTC)

@نعم البدل Ok, I have found some of the others now. I have a few changes in mind; I can share these when I have them ready and see what you think. عُثمان (talk) 16:58, 14 December 2022 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ @عُثمان: I'm all ears, feel free to make any changes. I only wish for it to be consistent. I've been longing to get feedback from another Punjabi user. Honestly these declensions have confused me quite a lot. I based them on 'A Descriptive Grammar of Hindko, Panjabi, and Saraiki' - which is the only recent and detailed study of Pakistani Punjabi (as well as Hindko/Saraiki) that I've seen.

I've also been struggling on how to highlight the difference between 'ਿਆ' and 'ੀਆ' in Shahmukhi lemmas. نعم البدل (talk) 17:14, 14 December 2022 (UTC)