Template talk:hu-infl-nom

Parameters stem2 and spe_sg_stem

Can you get the PAGENAME in a module? If yes, stem2 and spe_sg_stem could be generated from parameter 1. Looking at the examples provided in the documentation, stem2 appears to be the lemma. If parameter 1 includes PAGENAME (e.g. háza includes ház), stem2 will not be needed in the declension, but if not (tava does not include tó) then stem2=PAGENAME. As for spe_sg_stem: It seems that the value of this parameter is the same as parameter 1 minus its last character: álmo - álm. Can you generate this in the code without asking the user to provide it? I think these two parameters should be kept on the parameter list because I am still not sure if they are needed in some special cases, but if you could apply the above logic, the user would not have to supply them. --Panda10 (talk) 00:53, 16 January 2015 (UTC)

The superessive singular stem is not always parameter 1 minus the vowel. For levél or lé it's different for example. So if we make the default as you suggested, it would just mean having to specify the parameter for a different set of nouns.

I would prefer to make the template work independent from the page name. Right now the only page where the table can be made to display right is on the page for that word. Of course that is normally fine, but what if you want to show it somewhere else? —CodeCat 01:43, 16 January 2015 (UTC)

Ok, it's fine to leave the parameters. But I don't understand the second part. You said you prefer to make the template work independent from the page name, but currently it appears to depend on it. The nominative singular does use the PAGENAME because this is what is shown when I place {{hu-infl-nom}} on a test page (and not the noun I provide as a parameter). Which you did say in your second sentence. To show the template in another page, an optional pn (pagename) parameter could be specified. --Panda10 (talk) 17:54, 16 January 2015 (UTC)

menü, eskü

Two plural forms need to be corrected (they were incorrect in the old template, as well, what's more, they are also incorrect on the Hungarian Academy site). Current (incorrect): menüken, menükhez. Correct: menükön, menükhöz. Can you take a look? --Panda10 (talk) 15:17, 16 January 2015 (UTC)

It could be fixed, but if they are also with -e- on that site, how do you know they're wrong (assuming you're not a native speaker)? —CodeCat 15:25, 16 January 2015 (UTC)

If you are not a native speaker, you may not know. You can search for the two different forms on the internet and see the difference. The site is not perfect, I've found errors on it in the past. There is actually an error reporting button. But anyway, I've just double checked the old template {{hu-decl-k-front2}}. Actually, the case suffixes were correct in the original, but I believe when you modified the old template, somehow these two suffixes were copied incorrectly. --Panda10 (talk) 15:29, 16 January 2015 (UTC)

Yes, I see it now. I'm assuming that this would also apply to nouns with a stem ending in ö, ő or ű? —CodeCat 15:33, 16 January 2015 (UTC)

Yes. --Panda10 (talk) 15:35, 16 January 2015 (UTC)

Ok, I fixed it now. —CodeCat 15:41, 16 January 2015 (UTC)

Ok, but this did not fix the old template calls, see eskü. I know that eventually you will run a bot, but still. --Panda10 (talk) 15:47, 16 January 2015 (UTC)

That's fixed too now. —CodeCat 15:51, 16 January 2015 (UTC)

Looks good. Thanks for the corrections. --Panda10 (talk) 14:11, 17 January 2015 (UTC)

Irregular nominals

Which template should be used for irregular nominals and how? If we are planning to delete the hu-decl templates, what will happen to {{hu-decl-ez}} and {{hu-decl-az}}? --Panda10 (talk) 17:57, 16 January 2015 (UTC)

I haven't done anything with them, so they would be kept as before. We might eventually find a way to integrate them, but it might also just be too much trouble to do. —CodeCat 18:11, 16 January 2015 (UTC)

Keeping them is not a problem, I just wanted to clarify which main template will not be deleted to allow for future irregular nominals. It seem that it is {{hu-decl-table}}. --Panda10 (talk) 18:19, 16 January 2015 (UTC)

Inflection table header

I'm not sure if the new addition to the inflection table header is needed. The comment about the harmony might be fine, but the other part is really confusing. E.g. stem in long/high vowel, stem in -a-. Could you remove it? No one will understand it. The stem the module is using is a constructed stem, not a grammatical one and I don't think it will be helpful to users. --Panda10 (talk) 20:46, 16 January 2015 (UTC)

I've just noticed the following header for EU: "Stem in long/high vowel, back harmony". The constructed stem for EU ends in a hyphen. How should I interpret the header? Would you mind sharing your reasoning for the data.title variable and explaining how this header would help users? I still think it should be removed and just keep the "Inflection of xyz" title. --Panda10 (talk) 14:19, 17 January 2015 (UTC)

We show the inflectional class in most other languages too, I just copied the code from other inflection modules and adjusted it. For Hungarian there are only really two main classes: those ending in a low/mid vowel (with subclasses for each vowel) and those ending in any other vowel. Showing which vowel the stem ends in is useful because the user can see it without having to expand the table. As for EU, it still applies, EU declines as though ending with -u, which is a high vowel. —CodeCat 14:23, 17 January 2015 (UTC)

I think u is a short vowel, not long. So what does 'long' represent in the title "Stem in long/high vowel, back harmony"? --Panda10 (talk) 14:43, 17 January 2015 (UTC)

I intended it to mean long or high. That is, any vowel except a, e, o, ö. If you know a better way to describe such words collectively I can change it. —CodeCat 15:46, 17 January 2015 (UTC)

Traditionally, the five inflectional classes are connected to the plural, this is how the original templates were set up. So it would make more sense to use the other data.title you specified in module hu-nominals: plural in -ak, plural in -ok, etc. --Panda10 (talk) 19:03, 17 January 2015 (UTC)

But what if the noun has no plural? —CodeCat 19:06, 17 January 2015 (UTC)

Theoretically, every noun has a plural. It may not be used as much (maybe only in literature or poems), but it still can be formed. For example, we display only the singular inflection of Budapest, but Google returns several hits for Budapestek, its plural form. So even for proper nouns, there is a plural and a Hungarian editor should be able to figure that out (and anyone who wants to add a correct inflection table will need to be able to form a plural to get the parameters right.) For the inflection of possessives, perhaps it makes less sense to talk about plurals (although forming it should not be a problem for a native speaker), but as you mentioned above, you want to show the inflectional class in the header, and the plural indicates the class. Here is another idea. Instead of "plural in -xx", another option would be Class I, Class II, Class III, Class IV, and Class V (representing -ak, ok, -ek, -ek2, -ök, respectively) plus the harmony info, where each header item would be linked to a glossary or to the module's documentation itself where each term would be explained in a simple, non-technical language, so even a non-linguist editor would be able to understand their meaning. --Panda10 (talk) 19:55, 17 January 2015 (UTC)--Panda10 (talk) 19:55, 17 January 2015 (UTC)

I really don't see how having numbered classes is any clearer than just showing the stem vowel. I also don't understand why showing -ek is better than showing -e-. —CodeCat 19:57, 17 January 2015 (UTC)

Would you mind pointing me to the source where you got this inflectional class specification? The "stem in -a-", "stem in long/high vowel" etc. terminology? Was it a recent online PhD dissertation or something else I could read in English? I was born and raised in Hungary and grammar was taught with a different terminology. What you call the "stem vowel" is the vowel belonging to the suffix and not to the stem. Stem is defined in this wiki as "the main part of an uninflected word to which affixes may be added to form inflections of the word." It doesn't sound right to me to call "the lemma + the suffix vowel" a stem. This is not about the technical aspects of your module. That seems to be fine. It's about giving incorrect and confusing information to the users (the readers, not the editors). So could you point me to your source? Hopefully, that would resolve my concerns. --Panda10 (talk) 21:16, 17 January 2015 (UTC)

The vowel belongs to the stem. It's inherent in each word and not part of the ending. What happens is that the vowel preceding the suffix gets dropped in some cases if it's a short low or mid vowel. In Old Hungarian, the vowels hadn't been lost yet, so all nouns still had a nominative singular ending in a vowel back then. Treating the vowel as part of the suffix is confusing because then you suddenly have to deal with many suffixes, when clearly the suffixes are all the same except for the vowel. —CodeCat 21:25, 17 January 2015 (UTC)

Test cases page

Would it make sense to create a test case page for the module? I've seen this for hu-IPA and found it very helpful. --Panda10 (talk) 20:47, 16 January 2015 (UTC)

What would it be used for? Could you not test things in a sandbox? —CodeCat 23:25, 16 January 2015 (UTC)

I could test individual words in a sandbox. But I would not be able to see mass errors in a series of words in case the module is modified. I am aware of the Category:Pages with module errors. Maybe that would do it. I just thought a page for unit test cases was part of the process for Lua modules. --Panda10 (talk) 00:50, 17 January 2015 (UTC)

I designed this module using existing entries as a testbed. I would make a proposed change, but instead of letting the change actually show up in the entries, I would compare the new method with the old one to see if they give the same result for all entries. That's how I know that the module should work for everything it did before. —CodeCat 00:58, 17 January 2015 (UTC)

Nouns ending in -y (a vowel), as in tory

@CodeCat Hi, User:Einstein2 ran into a problem with a Hungarian word borrowed from English, tory. It ends in y which in this case is treated as a vowel and is pronounced /i/. The module generates an error that y is not a vowel. Would you please take a look and see if such cases can be included? Thanks. --Panda10 (talk) 16:24, 26 May 2015 (UTC)

Another example: Munkácsy. Many surnames end in -y that is not part of ny, gy, ly, gy. --Panda10 (talk) 14:05, 20 July 2015 (UTC)

New parameters to display variant forms

Please add the following new parameters to allow displaying variant forms within a table cell:

1. acc2 (accusative singular variant, appears only in a single table cell) - the difference between the two variants can be:
   • a linking vowel, e.g. tradicionális: tradicionálisat / tradicionálist
   • a variant stem, e.g. borjú: borjút / borjat, bő: bőt / bővet, fiú: fiút / fiat
2. pl2 (plural variant, appears in all plural cases in the plural column) - the difference between the two variants can be:
   • a linking vowel, e.g. kapható: kaphatók / kaphatóak
   • a variant stem, e.g. falu: faluk / falvak; borjú: borjúk / borjak, szó: szók / szavak, bíró: bírók / bírák, fiú: fiúk / fiak

Thank you! --Panda10 (talk) 13:42, 20 July 2015 (UTC)

It's been a little while since I worked on this and Hungarian declension, so just to refresh. The nominative plural form of the noun is always used, unmodified, as the stem for all plural nouns, correct? Like, all other plural forms just attach endings to the nominative plural? —CodeCat 17:20, 20 July 2015 (UTC)

Yes, that's correct. --Panda10 (talk) 17:29, 20 July 2015 (UTC)

And is there a correlation between the second plural stem and the second accusative singular stem? I know that often, the vowel of the accusative singular (if there is one) is the same as that of the plural. So I'm thinking that it might apply also to these extra forms. In that case, it might be more practical to have one parameter that specifies the alternative stem for both of them at once. Extra parameters would be needed then to handle the exceptional cases where they do differ, but you probably wouldn't need them as often? —CodeCat 17:39, 20 July 2015 (UTC)

There is a correlation but the linking vowel is not always used in the accusative when the plural has variants, and vice versa. For example, many adjectives end in -ható, they have two plural variants -hatók/-hatóak but only a single accusative form without a linking vowel -hatót. Currently, the template calls are {hu-infl-nom|kapható|o} and {hu-infl-nom|kaphatóa|o|-}. So if we specify the alternative stem for both at once, how would it know that the alt stem is for the plural but not for the accusative? Regarding the variant stems: It is true that there are only a few of them. --Panda10 (talk) 18:25, 20 July 2015 (UTC)

Well, presumably the same way it's done for single-stem words: the second accusative stem parameter would be "-". So {{hu-infl-nom|kapható|stem2=kaphatóa|o|acc_vowel2=-}}? No idea if that would actually work in practice, this stuff is hard. —CodeCat 18:33, 20 July 2015 (UTC)

The parameter stem2 is already in use. Could you use pl2 and acc2? They are clearer. Actually, acc2 (if there is one) could be a full form, but in this case it might not be needed: {{hu-infl-nom|kapható|o|pl2=kaphatóa}}. Since acc2 is not specified, there is no variant accusative form. When the variant is only the accusative, the template call would be {{hu-infl-nom|tradicionálisa|o|-|acc2=tradicionálisat}} or {{hu-infl-nom|tradicionálisa|o|acc2=tradicionálist}}. If it's easier, even pl2 could be a full form. --Panda10 (talk) 19:22, 20 July 2015 (UTC)

Proper nouns ending in -ee such as Lee

All forms should contain Lee, but in the inflected singular forms there is only *Le. Only the accusative is correct. For example, the dative is incorrectly *Lenek, instead of the correct Leenek. Currently, the old template is used because the forms appear correctly there: {hu-decl-k-front1|Le|e|t}. --Panda10 (talk) 13:56, 20 July 2015 (UTC)

This really is the same problem as the one with "Tory" above. The template expects regular Hungarian spelling, which never has ee in it I believe, nor does it have y representing a vowel. So it doesn't deal with unusual or foreign ones well. I'm not sure how it could be made to handle them either, at least not without sacrificing some of the automatic things that it currently does, like shortening stems ending in non-high short vowels (which is what happens with ee > e). So perhaps for now it's best to use the old templates for these cases. —CodeCat 17:23, 20 July 2015 (UTC)

Hungarian surnames often end in y as a vowel. The -ee ending is foreign, but all foreign words are inflected in Hungarian. Some with a dash, some without. The rules are a little complex even for the native speakers, so that's why it is important to show the inflection in a dictionary. I have no problem using the old templates as long as they can handle it. None of the templates could handle the inflection of Munkácsy for the second plural form Munkácsyak (the correct superessive is Munkácsyn, but the templates generate Munkácsyon), so for now I had to remove it and only one form is displayed. --Panda10 (talk) 17:52, 20 July 2015 (UTC)

Surnames ending in -ty such as Vörösmarty

@CodeCat The module treats the -ty ending as the letter ty. However, when a surname ends in -ty, it is traditionally pronounced /ti/. The instrumental and translative cases should be Vörösmartyval and Vörösmartyvá, but they are incorrectly generated as *Vörösmarttyal and *Vörösmarttyá. The template call is {{hu-decl-k-back|Vörösmart|y}}. It really bothers me that the new powerful Lua module is not able to handle simple things that the old template could before it was connected to Lua. It seems that the old approach to the parameters made that possible. Could you replicate the old functionality somehow? --Panda10 (talk) 14:50, 21 July 2015 (UTC)

This is quite difficult because the template has no idea when to treat ty as a letter and when as two. So it's far from a "simple thing"; it requires additional knowledge about the word that the template isn't aware of. I'm not sure what the best way to handle this is. Perhaps an additional parameter that tells the template to treat the word as if it ended with a certain vowel. That would also solve the "Lee" problem. Compare {{fi-decl-parfait}} which also needs such a parameter. —CodeCat 14:56, 21 July 2015 (UTC)

The call {{hu-decl-k-back|Vörösmart|y}} provides that information. The last parameter is y separated from the t. So that's the clue that is is not the letter ty. For words ending in the letter ty as in ponty, I used {{hu-decl-ok}} in the past. So could you just update the function within the module that handle the {{hu-decl-k-back}} calls? --Panda10 (talk) 15:04, 21 July 2015 (UTC)

Ah, I see the problem now. The module just sticks the two back together again. I will need some time to work this all out though, it's a rather difficult set of problems that has turned up. —CodeCat 16:01, 21 July 2015 (UTC)

Please remember to update the functions in the module for all the templates that were originally designed for nominals ending in a vowel: {{hu-decl-k-back}}, {{hu-decl-k-front1}} and {{hu-decl-k-front2}}. --Panda10 (talk) 12:39, 22 July 2015 (UTC)

ablaka, ablakunk, kertje in the plural

Why are the plural forms of these words deactivated? They're quite possible: ablakai, ablakaink, kertjei. Adam78 (talk) 06:52, 20 May 2019 (UTC)

@Adam78: This template doesn't deal with possessive forms. I'm not sure what you mean by deactivated. The above mentioned forms all exist: ablakai, ablakaink, kertjei. Panda10 (talk) 14:16, 20 May 2019 (UTC)

I asked this question because I wanted to add the declension to java, however, I've since found the way on how to do it.

What I meant by "deactivated" was that if you check out the examples at the bottom, you'll see n=sg in the lines mentioning ablaka, ablakunk, and kertje, while in fact their plural forms shouldn't be disabled or deactivated in these cases, even if they are not formed with -k but with -i. Instead of disabling these forms altogether, it would be better to enable the template to create the forms that you and I both mentioned as their plural forms. In other words, the declension chart at ablaka etc. should include ablakai etc. too, while currently it doesn't. Adam78 (talk) 14:23, 20 May 2019 (UTC)

Unfortunately, this template doesn't have the functionality to deal with the plural form of possessives (-i instead of -k) the way you describe it. It wasn't written for possessives at all. I started using it for the declension of possessive forms because there was nothing better. If you want this functionality added, the requirements will have to be clearly presented and a request would have to be made to find a Lua expert. As for the requirements, I still don't understand how this will work. Let's take ablak as an example. Under ablakom, both the singular and the plural columns of the declension table would be filled. Singular: ablakom, ablakomat, ablakomnak, etc. Plural: ablakaim, ablakaimat, ablakaimnak, etc. However, the head word is ablakom, so why would the table contain the declension of ablakaim? And what happens to the entry ablakaim? What would be displayed in the two columns of the declension table? The same as under ablakom? For some reason I find this confusing. But I agree that the current system has room for improvement. I believe the best solution would be to have a full combined declension table under the headword ablak (as shown for ember in {{hu-possessive}}) and no declension table at any of the possessive forms such as ablakom or ablakaim. Panda10 (talk) 20:35, 20 May 2019 (UTC)

partitive 3rd-person possessive (genitive) forms of comparative/superlative adjectives

@Panda10 Many (if not all) comparative and superlative adjectives can take the possessive (genitive?) -ja/-je and -juk/-jük in a partitive sense ("the … one/s out of them") and occasionally in phrases like "the most …… of the world", e.g. a (leg)szebbje/jük, a (leg)jobbja/juk, a (leg)rosszabbja/juk, a (leg)fiatalabbja/juk, a legidősebbje/jük, a (leg)normálisabbja/juk, etc. Obviously, the full possessive table cannot be inserted to their entries because their 1st- and 2nd-person possessive forms still don't exist, so I wonder if one extra possessive row could be added to their primary inflection table with an optional parameter, above or below the existing non-attributive -é(i) rows. If we find a solution, maybe a bot admin could be asked to insert this extra bit of code to their tables. Adam78 (talk) 22:31, 13 April 2022 (UTC)

@Adam78 I've added {{hu-pos-etek|legszebb|j|jei}} to legszebb. I think those forms can exist, Google returns hits. So, good point. But we should use the existing possessive template instead of adding extra possessive rows to the main table. The only change needed is adding the possessive template to all comparative and superlative adjectives. The derived word legszebbik also needs a possessive table. Panda10 (talk) 17:29, 14 April 2022 (UTC)

@Panda10 I'm afraid this is not a good solution. The 1st- and 2nd-person possessive forms normally don't exist for adjectives; except when they also exist as nouns. A szebbem sounds rather weird, for example. If we do use the possessive table, most of its forms need to be removed (replaced with a dash). But I don't think we should use them as it creates the false impression that their possessive forms exist. There are always one or two stray results, but basically it's a systematic difference between adjectives and nouns that only nouns have possessive forms. The 3rd-person possessive is a different thing, it is not used in the normal possessive sense but in a partitive sense, and only for comparative and superlative forms. Adam78 (talk) 18:23, 14 April 2022 (UTC)

@Adam78 OK, in that case one row (labeled Partitive) should be added below the existing non-attributive -é(i) rows. It should be enabled with a parameter just for the comparative and superlative adjectives, otherwise a dash should be displayed. As you know, the accelerated form creation for the -é(i) rows is still not working properly, the same expected for the new row. Are you planning to do the changes? Panda10 (talk) 18:36, 14 April 2022 (UTC)

@Panda10 I tried to update it but it didn't work. I reverted it to the preceding version. Adam78 (talk) 20:46, 15 April 2022 (UTC)

@Adam78 It will work but you also have to add a new row to {{hu-decl-table}} and several other hu-decl templates were updated as I look at the edit summaries (usually phrased as "added new parameters for non-attributive possessive forms"). Panda10 (talk) 21:10, 15 April 2022 (UTC)

@Adam78 What was the error? The step I described above is actually needed for the old declension tables. Panda10 (talk) 22:01, 15 April 2022 (UTC)

@Panda10 I think there were four errors. You can try reverting it and checking it out. Adam78 (talk) 22:09, 15 April 2022 (UTC)

@Adam78 There are typos in two places: pp1_pp instead of pp1_pl. What does bbj stand for? In case you need it, there is a sandbox method for testing modules. Create Module:User:Adam78/hu-nominals and copy the code. A test template that calls this module has to be created, as well (hu-infl-nom-test) and placed in Category:Hungarian experimental templates. Panda10 (talk) 17:45, 16 April 2022 (UTC)

@Panda10 bbj stands for the usual element of these derivations: legszebbje, legjobbja etc. Thank you. Sooner or later I'll muster the spirit to make another attempt. :) Adam78 (talk) 20:56, 18 April 2022 (UTC)