Wikidata talk:Lexicographical data/Archive/2022/08

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

spelling variants should be forms

I think woman fashion (L203929) should be a form of (L203930). Change my mind 😬 –Shisma (talk) 09:04, 9 July 2022 (UTC)

@Shisma: I wont change your mind, I agree with you. If there is no objection (@SixTwoEight:), I'll merge them next week. Cheers, VIGNERON (talk) 17:55, 13 July 2022 (UTC)

@VIGNERON: I likewise agree that they should be merged. My bot should've checked for alternate forms on Wiktionary before creating Lexemes. SixTwoEight (talk) 17:59, 13 July 2022 (UTC)

No comments in a month, I merged them. @Shisma, SixTwoEight: Cheers, VIGNERON (talk) 13:23, 19 August 2022 (UTC)

Is numeral really a category at the exclusion of others in so many languages?

In Punjabi, every numeral also has a "regular" part of speech, and numerals can be adjectives, adverbs, or nouns. These numerals take on the properties of their respective grammatical category; for example, a nominal numeral must have a grammatical gender, and whether an adjectival numeral has to inflect for gender depends on its inflection class ("red" or "black" adjectives, named after how the adjectives for those colors inflect). Adjectival cardinal 97 has gendered inflections, while adjectival ordinal 97th does not. Meanwhile, adjectival cardinal 94 lacks gendered inflections, but adjectival ordinal 94th is inflecting. I have been using items like "numeral adjective" and "numeral noun" which are subclassed to both "numeral" and the part of speech, so that these lexemes are not completely disjoint should someone want to do a multilingual query of numerals. (After all, adjectival ordinal 97 in Punjabi is still equivalent to numeral 97 in English or any other language with a comparable numeric system.)

I am curious if this may apply to other languages, because I am somewhat surprised how many languages so far have these simply categorized as "numeral." Maybe there are others which could be using "numeral noun" or "numeral adjective"? I speak English and admittedly I can't tell what part of speech numbers are supposed to be in English. Middle river exports (talk) 18:14, 25 August 2022 (UTC)

Yeah, I don't think treating them as special parts of speech makes a lot of sense. In English "ten" is there as a noun and a numeral, but "tenth" is only there as a noun and adjective. If we were unifying them and treating the ordinal as a special form of the cardinal then maybe it would make sense, but otherwise I don't think it's helpful. Better to link ordinal with cardinal via properties I think, if needed. ArthurPSmith (talk) 21:09, 25 August 2022 (UTC)

I had not realized that for English the "numeral" category had not been given to nominal cardinal numbers or to ordinal numbers. In reading more background this, it seems like various English grammarians have considered the numeral category as it pertains to English to be determinative, in the same sense "a," "and," and "the" are. When we say two horses or three dollars we are using them as determiners, and determiners are sort of an overlapping concept with adjectives that we have more specific properties for.

It turns out that determiner (Q576271) is a superclass of numeral (Q63116), which would indicate that numeral as a category for English would inherit all the properties of a determiner, but I think this also means many subclasses of numeral (Q63116), including the ones I added, and the items for cardinal, ordinal, and distributive numerals, should be linked to something else instead.

For example, cardinal number 1 ਇੱਕ/اِکّ (L672623) is just an adjective in Punjabi, and despite having similar spelling, ਇਕ/اِک (L686328) is a determiner which is not numeric as makes an "individual vs. many" distinction rather than a quantitative distinction. I have just using instance of (P31) to indicate type of number, but there should be a way to say something is a cardinal number without saying it is a determiner.

Perhaps we should relabel numeral (Q63116) to be something more specific, like numeric determiner or determinative numeral, to make it clear that this is really just indicating a grammatical option for numerals that some languages have, and create a new item that is a meta-class for numerals generally regardless of category. This would require thorough investigation on where to link numbers in various languages though. The advantage of creating the new meta-class would be that we can create this and use it right away without breaking anything else. Middle river exports (talk) 14:28, 26 August 2022 (UTC)

I don't consider numeral a proper lexical category because it doesn't describe how the word behaves in a sentence, so I try to use the actual part of speech (noun (Q1084), adjective (Q34698), etc) as the lexical category, with instance of (P31) cardinal numeral (Q1329258), ordinal numeral (Q923933), etc to describe that it's a numeral. I do still use numeral as a lexical category some cases though, either because I don't know enough about a language's grammar to be confident in picking another category, or when a language has numerals that don't clearly fit into any particular category. I especially like using instance of (P31) to say that things are particular types of numerals because then it doesn't really matter what happens with the lexical categories, I don't have to rely on everyone agreeing on which ones to use. - Nikki (talk) 06:14, 27 August 2022 (UTC)

I have come around to doing it this way, since even the combined category like "numeral adjective" could be taken to mean something different depending on definition of numeral. I ended up just moving determiner from "subclass" on the numeral item to "partially coincident with," since from what I could find reading about it numerals are not necessarily determinative in all the languages linked to it. (In Polish apparently, some of the numbers 1-10 may be considered determiners, but others aren't.) The "nominal" item numeral is linked to seems sufficient anyway as this term describes nouns, adjectives, and numeric determiners all at once.

I want to try reworking the query on the Ordia numerals tab to use instance of (P31) to see what this looks like across languages at the moment. I kind of want to update the English "numeral" lexemes to determiner since that distinguishes them more clearly from the noun forms, but don't want to surprise anybody since a lot of people edit lexemes in that language. Middle river exports (talk) 16:27, 27 August 2022 (UTC)