Wikidata talk:Lexicographical data/Archive/2017/07

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

L-items

I am sorry if I lose some conversation, but I can't find about it. Is it planned to create "L-items" for lexemes? Why not to simply create page "apple" for lexeme "apple"? All lexemes are different and it is convenient to refer them like https://www.wikidata.org/wiki/apple ... --Infovarius (talk) 14:29, 15 June 2017 (UTC)

For "apple" it could work, and I don't want to compare to it, but it doesn't work for "orange": here there would be at least six different lexemes in as many languages. Note that lexemes are per language, unlike Wiktionary pages, who put all lexemes from all languages that are homographs on the same page. --Denny (talk) 16:28, 15 June 2017 (UTC)
So orange@en will be at different page from orang@fr? Ah so... Can we at least have "disambig" page https://www.wikidata.org/wiki/orange for all these "oranges"? In which we have links to all Wiktionary oranges? --Infovarius (talk) 11:15, 16 June 2017 (UTC)
https://www.wikidata.org/wiki/orange would conflict with existing URLs, e.g. someone loading https://www.wikidata.org/wiki/Q1 could mean Universe (Q1) or lexemes such as wikt:fr:Q1. I would expect something like that to be a special page, maybe something like "/wiki/Special:LexemesByLemma/orange". - Nikki (talk) 15:57, 16 June 2017 (UTC)
I believed that there are no such lexemes... wikt:fr:Q1 is very unexpected... --Infovarius (talk) 22:24, 5 July 2017 (UTC)
On the English Wiktionary, we've started mapping the senses of certain words to corresponding Wikidata items, e.g. wikt:Paris (view the source and look for the "senseid" templates). I don't see a particular need to have separate treatment for lexemes, though. Lexemes fit just fine into the existing structure of items (Q....), since they are "things" that can have properties. As for the naming orange@en, that won't work either, because there's many lexemes spelled "orange" in English alone. You can think of a lexeme as a single part-of-speech section on the English Wiktionary. So there's orange an English noun, an English adjective, and an English verb. Sometimes there are also multiple noun lexemes with the same spelling in the same language, e.g. wikt:ban has three English nouns and one English verb. CodeCat (talk) 21:08, 16 June 2017 (UTC)
Correct. That's why all of them have L-pages with numbers, and nothing based on the Lexeme. In theory, it could be "lexeme-pos-language", such as orange(n)@en, but then you also have the case where even those can be multiple, just as the German See, where you have the same word, same POS, and same language, and yet two different Lexemes... Language is awesome :) That's why L-Items seem simpler. --Denny (talk) 20:16, 27 June 2017 (UTC)
I don't think we should hide or get rid of individual lexeme pages, but I do think there will be more demand for an easy way to list all lexemes with a given spelling because that is what people are used to. It's what Wiktionary does (OmegaWiki too) and it's very similar to how paper dictionaries work. Without it, the only option would be to wade through search results, which isn't particularly user friendly and will make it hard to navigate and (since I can't think of a good English word) unübersichtlich.
For example: When you are looking up the meaning of an unknown word, it's useful to see all the lexemes in order to get a proper understanding of what it means and how it's used. When someone inevitably adds information to the wrong lexeme, you want to quickly navigate to the right lexeme in order to move it there. Likewise, when someone inevitably creates a duplicate, you want to be able to easily spot it and easily find the right lexeme to merge it into.
If we have a special page like the one I suggested above, I could imagine having something like breadcrumbs at the top of the lexeme page, e.g. "orange > orange (en) > orange (en) (L123)" where the first one would link to a special page showing all lexemes for "orange", the second would link to the same page filtered by language and the third would be the current lexeme page. It could maybe include parts of speech too (e.g. "cat > cat (en) > cat (en) (noun) > cat (en) (noun) (L124)").
The Cognate extension could perhaps be extended to include automatic links from Wiktionary to the Wikidata special page and maybe even from the Wikidata special page back to the Wiktionary pages. That would make navigating between Wikidata and Wiktionary projects easier.
- Nikki (talk) 10:53, 3 July 2017 (UTC)
That sounds like an interesting proposal, and I am certainly not opposed to it. In fact, it is very well compatible with the current work plan, and once the first few structures are there, this can probably be easily added. --Denny (talk) 21:59, 3 July 2017 (UTC)

Meaning of pesonal names

Hello, Do you know how can I have statments for personal names which stands for meaning of name?

I find the question of how to represent meaning indeed a very interesting question, and I, for one, have no idea how to do it. A few ideas, but really just trying them out. Nothing that we can really capture with Wikidata yet, nor given the Wiktionary for Wikidata plan. I think that is pretty much an open question, and it would be great to tackle it. --Denny (talk) 22:01, 3 July 2017 (UTC)