Wiktionary talk:Language families

qfa
The bit on "qfa" languages is confusing, since none of the examples given have "qfa" in them!

At first, I thought that perhaps consensus had changed and qfa was no longer used, but the list further down does have some qfa language families. So perhaps some of those should be used as examples. -- pne (talk) 07:32, 23 May 2012 (UTC)

Removing/adding families
Is this the place to bring up suggestions on adjusting the family tree structure? Sure is looking quiet in here.

I'd like to propose getting rid of the "Finno-Ugric" family. This is a contended node (much like other intermediate Uralic nodes like Ob-Ugric, Ugric, Finno-Permic, Finno-Samic, none of which appears in this list) and is, altho found widely in older literature, not accepted by current Uralistics consensus. On Wikipedia we've abandoned using it on language infoboxes etc. Cf. Finno-Ugric languages, Uralic languages for some basic points and further reading.

In practical terms, the issue is that maintaining separate categories for Finno-Ugric and Uralic reconstructions would be largely redundant, since PFU and PU are practically identical (some differences arise from varying transcription practices).

Relatedly: are there uses beyond reconstructions for having a family in the system? A lot of subfamily units seem to be missing from here, but if there's no other use, this probably won't be necessary too soon.

--Tropylium (talk) 15:07, 24 October 2012 (UTC)
 * If we did that every time a language family is redefined, we would never be done moving categories back and forth.
 * Agree on that, hence I did not push for adding any of the competing new subfamily proposals, only scaling back to what is actually stable. (Having some universally undisputed subgroups like Permic around would come handy eventually, but again that probably would be years from now at best.) --Tropylium (talk)
 * The most common use for these categories, I think, is for new users to find their language. We're not here to describe the most current theories or anything, we have Wikipedia for that. Therefore, there should be some kind of, well, stability in the category structure, so links will not turn dead over night. This is why I often stick to traditional categorizations, which are probably more well-known than the newest and most current theories. This is just my two cents of course, but I think I am the only one dealing with these categories around here. -- Liliana • 19:45, 24 October 2012 (UTC)
 * I agree with Liliana, but I think we can also mention somewhere that it may not be a real language family. --WikiTiki89 00:45, 25 October 2012 (UTC)
 * We also use families to categorise etymologies. We don't normally put etymologies directly in families, but in particular the subcategories of Category:Terms derived from Germanic languages have many entries in them (most erroneously). 01:08, 25 October 2012 (UTC)
 * Yes, etymologies is exactly the angle I'm coming from here. If you peeps think changing the system from what is prescribed by ISO 639-5 to something else would be too much of a hassle or would set a bad precedent, would it be OK to at least ignore the underlying system and start functionally using a single Uralic etymologies category rather than FU/U separately?
 * Alternately, in case you have no objections beyond "too much work": would dealing with this take admin rights or something (beyond actually deleting cat pages)? Any reason I (or anyone else wanting to do this kind of an edit) could not just myself edit in/out families? --Tropylium (talk) 12:42, 27 October 2012 (UTC)

West Atlantic
"Atlantic", a.k.a. "West Atlantic" (code: alv-wat) is now generally recognized as a geographic grouping rather than a genetic family. And unlike e.g. the Australian Aboriginal language family (aus), which is a top-level family and thus makes itself semi-useful* by collecting all the various languages of Australia which aren't in other families into one category, Atlantic isn't even a top-level group: there are valid genetic families above and below it which all of its languages already belong to. Therefore, I propose to delete it from Module:families/data, and to delete its various empty categories, such as Category:Atlantic languages and Category:Terms derived from Atlantic languages. - -sche (discuss) 20:10, 6 December 2013 (UTC) * Perhaps aus does not make itself useful enough to be kept, but that's for a separate discussion to decide. - -sche (discuss) 20:10, 6 December 2013 (UTC)
 * ✅. - -sche (discuss) 06:18, 16 December 2013 (UTC)

Australian Aboriginal (aus)
These are geographic groupings rather than genetic families. The codes should be deleted as groups (i.e. nothing should declare cai as its family, and the categories can be deleted). It might be worth keeping them as prefixes, however, and continuing to use them in that capacity (i.e. keep Chinookan as nai-ckn rather than renaming it qfa-ckn, and if a new North American language family is recognised, assign it the code nai-foo rather than qfa-foo). - -sche (discuss) 02:01, 20 December 2013 (UTC)
 * I have deleted cai, nai, sai and aus but left notes explaining that they can continue to be used as prefixes in creating exceptional language codes and family codes. - -sche (discuss) 04:57, 14 January 2014 (UTC)

Khoisan (khi)

 * There's also "cau" (Caucasian), "paa" (Papuan) and "khi" (Khoisan) that have "qfa-not" as their family. 02:47, 20 December 2013 (UTC)
 * Abstain on Caucasian. Keep Papuan for convenience — I realise that's not an entirely consistent or defensible position, but we even disambiguate some lects (which otherwise have the same names as other lects) by their membership in the Papuan family, so it seems worth keeping. Delete Khoisan. - -sche (discuss) 10:29, 21 December 2013 (UTC)
 * delete all except Khoisan -- Liliana • 09:33, 21 December 2013 (UTC)
 * Papuan and Khoisan struck as "no consensus for deletion". ("cau" not dealt with at this time because it is still in use.) - -sche (discuss) 04:57, 14 January 2014 (UTC)

Kordofanian (kdo)
There's also Kordofanian, an amalgamation of some probably-Nilo-Saharan languages and some probably-Niger-Congo languages. - -sche (discuss) 09:04, 24 December 2013 (UTC)
 * Hell no! This one should definitely be kept. -- Liliana • 09:38, 24 December 2013 (UTC)
 * Kordofanian struck as "no consensus for deletion". - -sche (discuss) 04:57, 14 January 2014 (UTC)

bih
Just a note that the only family code which was included by the ISO as of 2014 (sic), but which was and is not included by Wiktionary, is bih (Bihari languages). - -sche (discuss) 20:38, 11 August 2015 (UTC)