Template talk:la-ndecl

I've created a new template to replace all of the noun  templates. Some examples can be seen at User:Benwing2/test-la-ndecl. I'll document it better but in general the first param is the declension class, the second param is the lemma, and the third param if necessary is the stem of third-declension nouns (same as the second argument before to various of the templates). It does a good job of autodetecting the declension subclasses, but you can also specify them explicitly using a slash and one or more subtypes, or force the autodetector not to autodetect a certain subtype by specifying a minus before the subtype (e.g.  to turn off the automatic detection of the   subtype for nouns whose lemma ends in -ius). I've also created a script to convert all the old template calls to the new format, and verified that each of them doesn't change the contents of the declension table (but in the process, a lot of likely bugs in the preceding tables have been revealed). Benwing2 (talk) 05:39, 5 July 2019 (UTC)
 * Seems reasonable. Out of curiosity, does it have a better way to deal with heteroclitic nouns like and with all the stuff la-decl-multi handles? —*i̯óh₁n̥C[5] 08:54, 6 July 2019 (UTC)
 * I am modelling it after Module:ru-noun, which is pretty much able to handle all of these things, although the syntax isn't quite as nice as Module:la-decl-multi. It would work like this (not yet implemented):
 * (a noun with two possible declensions, either Serapis -is -ium or Serapis -idis -idum)
 * (a heteroclite, modeled as a noun with two-declensions, one singulare tantum and one plurale tantum; I might need a bit of special-casing here to avoid thinking the noun has two possible lemmas)
 * (for ; the _ means "join two words with a space", the + means adjectival, the masculine gender of the noun is autodetected as the default for decl-4, the fact that the adjective is third-declension two-stem is autodetected)
 * (for ; 0 means "indeclinable")
 * (for ; the - means "join two words with a hyphen")
 * (for ; a colon followed by arbitrary text means "join two words with that text"; here, there is no text, hence the words will be joined directly)
 * (this should also work for, because adjectives have no inherent gender [I think I will implement it this way] and the gender for rēs is explicitly given as feminine)
 * (for ; the gender of rōs cannot be autodetected so it must be given, and an error will result otherwise)
 * Benwing2 (talk) 14:29, 6 July 2019 (UTC)
 * A little verbose for my taste, but whatever. Can you also handle assimilation? —*i̯óh₁n̥C[5] 19:12, 6 July 2019 (UTC)
 * If you have suggestions for making it less verbose, while still supporting cases with multiple declensions (e.g. Serapis, epulum), I can certainly implement them. As for quicquid, can't you just use overrides, rather than having a special case in the module? Benwing2 (talk) 21:29, 6 July 2019 (UTC)
 * BTW I can certainly imagine a -type syntax that autodetects. Benwing2 (talk) 21:32, 6 July 2019 (UTC)
 * One possibility: = epulum vetus pl. epulae veterēs. The other cases would be expressed like this:
 * Benwing2 (talk) 23:22, 6 July 2019 (UTC)
 * This syntax looks very good. —*i̯óh₁n̥C[5] 01:57, 7 July 2019 (UTC)
 * Benwing2 (talk) 23:22, 6 July 2019 (UTC)
 * This syntax looks very good. —*i̯óh₁n̥C[5] 01:57, 7 July 2019 (UTC)
 * Benwing2 (talk) 23:22, 6 July 2019 (UTC)
 * This syntax looks very good. —*i̯óh₁n̥C[5] 01:57, 7 July 2019 (UTC)
 * Benwing2 (talk) 23:22, 6 July 2019 (UTC)
 * This syntax looks very good. —*i̯óh₁n̥C[5] 01:57, 7 July 2019 (UTC)

Ready to convert existing noun to
I am close to being ready to convert the 40+ existing noun templates into calls to the single  template. The syntax is as documented here, and follows the general format of closely except that it requires the full lemma and autodetects most subtypes. Support for multiword and multipart nouns is present, e.g. and, and alternant support to support heteroclites (e.g. ) and nouns with multiple possible declensions (e.g. ) is also present. The only thing I haven't yet implemented is support for adjectives in multiword nouns. But that shouldn't prevent conversion of all calls to for nouns. The only thing that can't be converted at present is about 1/3 of calls to, specifically those that include declined adjectives. I have written a script to do the conversion and verified that it doesn't change the declension, by using template calls such as, and  to generate all the forms for the existing and proposed converted templates and comparing them. I'd like to convert the nouns and delete the then-unused old templates; what do people think? Benwing2 (talk) 04:37, 12 July 2019 (UTC)
 * I'm fine with this. —*i̯óh₁n̥C[5] 18:16, 12 July 2019 (UTC)

Disabling older genitive for nouns in -ium/-ius
Could there possibly be a field to mark a term as postclassical so that it disables the alternative gen. sg. -ī ending for nouns in -ium/-ius without having to key in the genitive directly? See absarius for an example where I've overridden it manually. This would make creating Medieval Latin entries a little easier. —Nizolan (talk) 19:41, 11 May 2020 (UTC)
 * Apologies for not responding earlier. There's already support for this; use  along with the declension number for masculine nouns, and   for neuter nouns. I've fixed up absarius appropriately. In general you can cancel out any autodetected subtype using a hyphen before the subtype. For example,   indicates a non-i-stem noun when the noun would be autodetected as i-stem. There's an example of using   in the  documentation in conjunction with, and the general canceling behavior of the hyphen is mentioned, but   and   aren't specifically mentioned. I will fix that and try in general to clean up the  documentation and include undocumented features. Benwing2 (talk) 17:40, 17 May 2020 (UTC)

Removing unlikely forms
In the ablative singular “Homine sapiente sapientī”, “Homine sapientī sapiente” and accusative plural “Hominēs sapientēs sapientīs”, “Hominēs sapientīs sapientēs” inflections need to be removed because using different forms in one inflection is unlikely. J3133 (talk) 14:54, 30 October 2020 (UTC)
 * Yes, I agree. I implement something similar already for Russian and Hindi, I'll see about implementing it for Latin. Benwing2 (talk) 02:58, 1 November 2020 (UTC)

Manual declensions
Can and  be made automatic? J3133 (talk) 10:45, 2 February 2021 (UTC)
 * Let me look into this. should be easier; there may not be current support for the hypothetical inflections of, I'll have to look into it. Benwing2 (talk) 06:14, 15 February 2021 (UTC)

i-stem twice
See. J3133 (talk) 06:58, 21 August 2021 (UTC)

Fifth declension
The Accusative-Singular should be written -ēm, shouldn't it? Sabinettus (talk) 01:06, 27 January 2024 (UTC)