Panda10

Potential forms

I made some experimental versions of the verb conjugation tables to include the potential forms as well (which is an absolutely productive feature, unlike the causative, frequentative or other forms). You can change "hu-conj-ok", "hu-conj-ek" and "hu-conj-ök" to hu-conj-ok2", "hu-conj-ek2", "hu-conj-ök2" in the preview to see how they would work and what they'd look like. Suggestions are welcome. Since the "-hat/-het" form is in the second table, there is more room for the adverbial participle, so I added the "-ván/-vén".

Another important thing would be to mention the possibility of verbal prefixes being separated (especially in the subjunctive/imperative forms and after a focus or negation), but for the time being, I think we could just include a line at the top reminding readers of this fact, rather than inserting more tables and/or more forms. Adam78 (talk) 15:54, 28 January 2024 (UTC)

@Adam78 I tested a few verbs, it looks good. Suggestions:

• The verb sodor has three conjugation tables, the verb függ has two. Since the potential is the same for all variants, it would make sense to add the potential to only one of them, otherwise there will be too many tables. This means we have to keep the old tables around and maintain them the same as the new ones in case of future modifications. Perhaps adding a new parameter nopotential=y to suppress the potential table might help in these situations where only the first table needs the potential forms.
• There is a typo in the new Adverbial Participle cell for -ek and -ök verbs: the -ván suffix is added instead of -vén. E.g.: történve (történván).

Panda10 (talk) 18:02, 28 January 2024 (UTC)

@Panda10 I fixed the typo (actually a different error) on the same day. Today I also managed to implement the parameter no-potential (with a hyphen, in accordance with no-subj).

• I wonder if we can eliminate also-cond by allowing the table to display alternative conditional forms depending on another parameter, also-subj. Having also-cond seems like an extra nuisance to insert everywhere (I didn't even remember it existed), especially since these forms are possible in the same cases when the alternative subjunctive forms are possible (at least I believe so). I wonder at how many places the former is currently applied.
• You might want to check the table {{hu-conj-ok2}} etc. at genuine ik verbs, with parameters like m=y|also-subj=y|also-cond=y to find out if my solution with the note at the bottom right corner is all right. (We could have a similar note for the earlier, "non-potential" table wherever this form is displayed.)
• How can we find all the verbs with more than one conjugation tables, where the new parameter should be used in all tables except the last? Category:Hungarian verbs with alternating stems tend to have a single table each.
• Having the old tables around and maintaining them simultaneously with the new ones (forking) doesn't sound like a good idea to me. I hope that (after verifying and finalizing it) we can stick to the new one by using the parameter "no-potential" wherever necessary. Accidentally having a redundant table here or there is not as bad as constantly having to update two tables. (We'll also need to update conjugation tables of irregular verbs manually.)

Adam78 (talk) 17:21, 30 January 2024 (UTC)

@Adam78 Thanks for making the changes.

• The also-cond parameter: Currently, it is used only in illik, érzik, hallik, történik, mérséklődik. I'm not sure if it can be eliminated because for some verbs the also-subj can be used but not the also-cond (e.g. csúszik).
• The ik verbs: I tested viseltetik, vitatkozik and történik. Since the latter has the impers=y parameter, the note at the bottom (*:currently rare due to homonymy with 1st-person) may not make sense since no 1st person is displayed. You might want to remove the colon after the asterisk.
• I requested a bot for finding verbs with multiple conjugation templates. There are 30 in Category:Hungarian verbs with alternating stems and found illik by chance - that’s 31 so far.
• We agree - we don’t want two sets of templates.
• There are quite a few custom templates that will have to be updated. Besides the ones listed in {{hu-conj}} documentation, there are the ones built for a single verb such as {{hu-conj-iszik}}. An option for these could be a new version of {{hu-conj-potential-hat}}/{{hu-conj-potential-het}} which would be added manually after the hu-conj template. The only visible problem is the template header text. For example if I add {{hu-conj-potential-hat}} to ró, the text is "conjugation of ró", what we'd need is "potential conjugation of ró".
• What are you planning to do with the -hat/-het verb form entries? They contain their own conjugation template now, are you planning to delete the conjugation section? Or we keep them? What to do with new -hat/-het verb forms? Should a conjugation template be added?
• Are there any other new functionalities you added that I should test?

Panda10 (talk) 22:05, 30 January 2024 (UTC)

• also-cond: I checked csúszik and I found instances for first-person csúsznám [1] and third-person csúsznék (looking up "hiba csúsznék") [2] [3] [4], so actually it's no exception: if the also-subj can be applied, the more literary conditional forms can be formed as well. That's why I think it's redundant. In fact, both of these parameters might be redundant when m is present. For example, I checked viselkedik where only parameter "m" is given, but actually if you look up the forms viselkedném [5], viselkedjem [6], viselkedjék [7], you can see that they exist. I'm not convinced we need three parameters for this kind of conjugation. Even two might be one too many.
• I added impers=y as a condition for suppressing the existing note.
• We could have a line as an extra bottom row of the existing table about these forms, noting that they are somewhat archaic or literary. Originally I was thinking of writing that they are fairly uncommon since the 2nd half of the 20th century, but Digitális Irodalmi Akadémia has several authors who were active in the second half of the past century and a form like uralkodjam (for instance) is found there 18 times, so maybe we don't need to specify a time limit.
• Another note that could be added is some warning for the verbal prefix to be separated, but only to the lemmas of verbs that have a prefix. Maybe a bot could add a parameter to these few hundred tables if we come up with a name (e.g. prefixed=y?) and make a bot request. In fact, a more specific parameter like pref=meg would be better because it would also make it possible if we want to display separated forms, while it can be used for the time being to simply check if it has any value whatsoever or it's not defined. This piece of information could be extracted from the existing categories (like "Hungarian verbs prefixed with meg-") or the {{af}} above that generates this category.
• Category:Hungarian verbs with alternating stems has many more than 30 entries (namely, 224) but none that I checked has more than one conjugation templates. Thank you for asking a bot admin to find them. (I didn't find an option to look for >1 instances of a template at Petscan.)
• hu-conj-potential-hat and -het could be used both for irregular verbs and for -hat/-het forms (though we won't need to add more to the latter group), but I'd like to update them to the layout that we've just developed. A new parameter could be used at irregular verbs to modify the header text. What should be the name of this parameter? For example, derived=y could suppress the text "potential" in derived (-hat/-het) forms?
• I don't think I added any new functionality that we haven't covered.

Adam78 (talk) 14:40, 31 January 2024 (UTC)

@Adam78

• also-cond: Thank you for the linked examples. Since also-cond is used by only a few verbs, its function can be added to also-subj and the parameter removed. We may have to do some more research about merging this into m=y but I'm not against it.
• impers=y: I tested történik. The note is no longer there but the asterisk is still there at történhetnék. Is there a way to suppress that, too?
• I'm fine with adding an additional note about somewhat archaic or literary.
• For the prefix topic, pref=meg makes sense.
• Yes, Category:Hungarian verbs with alternating stems has 224 entries, out of which 30 has multiple templates. I went through the entire category manually.
• derived=y is fine for the parameter name.

Panda10 (talk) 17:51, 31 January 2024 (UTC)

I think I did these all except for modifying the existing two -hat/-het templates. You may want to check its functionality in various combinations again because I had a feeling on many occasions that I messed it up… even though I tried to fix it and check it afterwards. I also made minor changes in the layout. Adam78 (talk) 21:25, 2 February 2024 (UTC)

I'm sorry, there have been still some mistakes but I think I've corrected them. You can try {{hu-conj-ek2}} in megtörténik, also removing the parameters impers or intrans and especially adding also-subj=y and pref=meg or either of them. Adam78 (talk) 14:49, 3 February 2024 (UTC)

@Adam78 Please check sodor. When I switched to {{hu-conj-ok2}} in the first conjugation template, the potential table suddenly included the two additional conjugation tables that should be below the first one (they are on the lower right side). Panda10 (talk) 21:16, 3 February 2024 (UTC)

Now I think it's okay. It seems I cannot make an entire row conditional (with no trace when it's inactive), so when none of the two additional rows are present (about the prefix or about archaic forms), there will be a thin light green line at the bottom. Maybe it's not that bad. Adam78 (talk) 22:25, 3 February 2024 (UTC)

@Adam78 The extra green line looks fine but you could still add a comment about the forms in parentheses. That would solve it. Panda10 (talk) 22:30, 3 February 2024 (UTC)

Done. Adam78 (talk) 22:51, 3 February 2024 (UTC)\

Potential forms - testing

I'm adding my testing notes as I find issues. Please see elromlik. After adding pref=el, the note comes up like this: "The prefix can split from the verb, e.g. nem romlhat el or el is romlhat." Panda10 (talk) 18:46, 4 February 2024 (UTC)

@Adam78 I started a new section for testing. Panda10 (talk) 18:48, 4 February 2024 (UTC)

Thank you; I think it's fixed. Adam78 (talk) 20:26, 4 February 2024 (UTC)

@Adam78 összefügg: Adding pref=össze doesn't add the note at the bottom. Also, the pref parameter can have any incorrect value (any string). I wonder how hard it would be to check this against a list of valid verbal prefixes. Panda10 (talk) 20:48, 4 February 2024 (UTC)

I'm sorry, I forgot to update the -ök conjugation type in accordance with -ok and -ek. Now it's done. I don't think this error can occur again at other prefixed forms if it works fine with the ones we've checked, because now the templates for all three vowel types are completely identical with the exception of the linking vowel itself. Adam78 (talk) 21:08, 4 February 2024 (UTC)

@Adam78 Testing the pref parameter with a plain verb: szerkeszt - I added pref=meg and the following note appeared: "The prefix can split from the verb, e.g. nem rkeszthet meg or meg is rkeszthet." Testing the wrong prefix: elver - adding pref=össze, the note is "The prefix can split from the verb, e.g. nem het össze or össze is het." Panda10 (talk) 21:08, 4 February 2024 (UTC)

It's currently not meant to work in any other way as long as you supply a prefix in a parameter, which prefix is not actually there in the verb. The template only counts the number of characters, so if you supply "meg", it will remove the first three characters, which is "sze" in "szerkeszt" as if it were a prefix. The template is easy to make a fool of, like if you supply "ki" (or any other sequence of two characters) as the prefix for "kiált", then it will display "nem ált ki, ki is ált". Do you think the template should ignore the prefix parameter if it's not the starting few letters of the verb? (It still couldn't cope with the "kiált" type, though.) Adam78 (talk) 21:17, 4 February 2024 (UTC)

@Adam78 It's a good idea to ignore the pref parameter if is doesn't match the starting letters of the verb. So I assume you can't check the value of the pref parameter against a list of valid and allowed prefixes. I saw that you created a module. Where do you use it? I also tested leading and trailing spaces (e.g. pref= meg ), they were correctly ignored, but pref=me g created the same problem as above. I understand why (because of the counting) and I'm not sure how deeply we should go into this problem and its solution. Editors can make mistakes and it can go unnoticed. Panda10 (talk) 21:26, 4 February 2024 (UTC)

As a matter of fact, editors can make mistakes when entering the parameters for any other inflection table, for verbs, nouns, adjectives, or pronouns alike. There are nearly a hundred verb prefixes; I don't think we need to check all in the template. However, checking the first few characters shouldn't be difficult; I'll try to do that. (The module proved to be unnecessary as I delved more into the matter. Please be so kind as to delete it.) We might as well want to make it possible by means of another parameter to indicate if a verb does NOT have a separable prefix, like "ellenőriz" (though Nagyszótár does have an example where it's split; looks jocular though) as opposed to "ellenáll". Adam78 (talk) 21:40, 4 February 2024 (UTC)

I deleted the module. So what would be the other parameter: nopref=ellen? As for making mistakes, sometimes they are obvious because the template will not work at all. But if you don't want this additional check, I'm fine with it. By the way, the list of verbs with multiple templates was generated and can be seen here: Wiktionary:Grease_pit/2024/January#Bot_request:_Hungarian_verbs_with_multiple_conjugation_templates. Panda10 (talk) 21:52, 4 February 2024 (UTC)

Thank you. I wrote "checking the first few characters shouldn't be difficult", but I still cannot implement it. So we'll have to make do without this verification. :( Adam78 (talk) 08:54, 5 February 2024 (UTC)

@Adam78 Have you tried the switch function? I see you used it in {{hu-affix-group}}. Can all the valid prefixes listed in a new switch function and used as a verification? Panda10 (talk) 14:44, 5 February 2024 (UTC)

It should be possible for sure, I can create the algorithm, but I don't actually know the syntax and the inner workings of the code, the background of these languages, so I can't locate the error. You can see I created {{module:hu-prefixed}} and I modified the code accordingly, but it still won't produce the right output. Adam78 (talk) 18:17, 6 February 2024 (UTC)

@Adam78 Sorry, I didn't mean for you to create a module. The switch function is not Lua. I thought that it can be used the same way as {{hu-case}} is using {{hu-case-ending}}. A new template would contain the switch function and the list of valid prefixes. If there is a match, the prefix is returned, if there is no match, the default is returned. Maybe I'm mistaken, I don't know. I do see the error when I test the new template in megtörténik, but I have not idea how to fix it. Let's just leave the idea of verification and trust that the pref parameter will always be entered correctly. I'm really sorry you spent so much time on this. Panda10 (talk) 18:35, 6 February 2024 (UTC)

I know you didn't mean I should create a module but I didn't know how to apply a switch. The syntax of #switch: (which you brought up) is very similar to that of #if: (which I tried to use previously), and the latter did not work for some reason, that's why I tried resorting to another approach. Anyway, I restored the version that works. Adam78 (talk) 18:59, 6 February 2024 (UTC)

@Adam78 I'm in the process of writing up the next steps. And since I'm still testing, I found something at összefügg: Switching to the new template name, adding pref=össze to both templates, adding no-potential=y to the second template resulted in an extra blank line between the last conj table and Derived terms. Could you please take a look? Panda10 (talk) 22:19, 6 February 2024 (UTC)

I noticed it too, just I forgot about it. I think I've fixed it now. Adam78 (talk) 22:30, 6 February 2024 (UTC)

@Adam78 Looks good, thanks. Two more questions: We talked about adding a function to check if the pref parameter value matches the first few characters of the verb. Was it implemented? Was a nopref= parameter added for non-separable prefixes such as ellenőriz? I'm going through our entire thread to make sure there is nothing missed for the implementation and documentation. Panda10 (talk) 22:34, 6 February 2024 (UTC)

Checking the first few characters: In theory yes (I modified the code accordingly), in practice no (it didn't work, so I reverted it). The nopref parameter hasn't been added; I'll try to implement it. Adam78 (talk) 08:11, 7 February 2024 (UTC)

You can try e.g. kiabál with {{hu-conj-ok2|kiab|á|l|nopref=y}}. Adam78 (talk) 10:38, 7 February 2024 (UTC)

@Adam78 Excellent, thanks. Next set of items:

1. There is the special case of reorganizál and reprivatizál. The etymology says re- is a verbal prefix but I don't think it can be split. How should we handle them? Just a comment at re- that this is not separable?
2. I believe you added the function of the also-cond parameter to also-subj, correct? The parameter is still in the new code, though.
3. When also-subj=y is added, there is no asterisk in the regular table for conditional present indefinite 3d-person forms but there is one in the potential table along with the note at the bottom: * currently rare due to homonymy with the 1st-person.
4. I can’t figure out why the documentation page of {{hu-conj-ok}} is not showing up on the template page and why isn’t this template in any category. This is an old issue and has nothing to do with your new templates.

Panda10 (talk) 17:53, 7 February 2024 (UTC)

1. With a comment at re- and probably also in the entry of the verbs. I'm not convinced these prefixes were added within the (history of the) Hungarian language: if not, then they're just loan words from forms that already contain re- in another language (reorganize, reprivatize or their equivalents), so they're not prefixes from the point of view of Hungarian. However, if they can be added to existing Hungarian words no matter whether this form exists in the source language, then they're prefixes indeed within Hungarian. We should still mention it in the entry of such verbs that they're not separable.
2. Yes, only for safety, as long as testing is still in progress.
3. I haven't been able to reproduce this error. Please show me an example verb where it comes up. (I suppose we only use this parameter with -ik verbs, don't we?)
4. I think it's because of the excessive length of its documentation.

Adam78 (talk) 19:48, 7 February 2024 (UTC)

@Adam78

1. I agree.
2. I think also-cond should be removed from the code now while testing is in process. We want to find issues now, not after implementation.
3. I tested kopik.
4. Okay, it's possible.

Panda10 (talk) 20:04, 7 February 2024 (UTC)

2 & 3: I think they're done. Thank you for your thorough testing! Adam78 (talk) 21:02, 7 February 2024 (UTC)

Causative-factitive form for verbs

If you look at a paradigm table in Nagyszótár (e.g. emel and you click 1b), you can see that the causative-factitive form emeltet is also included in the paradigm (along with the archaic passive form, formed similarly). Magyar grammatika says the following (p. 54):

Alapvetően elsősorban a ragok és jelek előtti tőváltakozást kell figyelembe venni a tőtani besoroláskor. Kivételek az igenévképzős, a műveltető képzős, a ‑hat/​‑het képzős, az ‑ás/​‑és képzős, valamint a ‑d törtszámnév- és a ‑dik sorszámnévképzős ala­ku­la­tok, ezek ugyanis közel állnak az ige, valamint a melléknév paradigmájához, és kép­zőik automatikus képzőknek tekinthetők (minden igéhez, illetve melléknévhez hozzá­tehe­tők).

We have already included participle and infinitive forms as well as verbal nouns, we're currently including potential forms, so the only present-day form that is lacking from our verb conjugation tables is causative-factitive. I was thinking maybe we could insert it under the infinitive (?).

It cannot be added automatically, partly due to semantic reasons, and partly because even morphology doesn't always determine it unambiguously. There is a tendency though that monosyllabic stems take -At and polysyllabic stems take -tAt. The final consonant is not quite relevant: olvas-tat and mos-at, őriz-tet and hoz-at, pihen-tet and fon-at, idéz-tet and néz-et, zavar-tat and mar-at, ígér-tet and kér-et. Some monosyllabic stems still take -tAt: szok(ik), szök(ik), kel, ül, un, te(sz), ve(sz), hi(sz), hány, áz(ik). Also, some polysyllabic stems like sikong and borzong take -At. If the final consonant is t, the quality of the preceding sound is decisive (vowel: arat-tat; consonant: bont-at). Sometimes both forms are possible with different meanings (e.g. [ki]vonat vs. vontat, folyat vs. folytat, járat vs. jártat, nyomat and nyomtat). Maybe there could be one optional parameter with the causative-factitive suffix for the given verb (and possibly a third option for both) and another parameter where the full causative-factitive form can be supplied when necessary (e.g. for short stems like zörget from zörög). Adam78 (talk) 13:53, 8 February 2024 (UTC)

@Adam78 If you want to add a single cell with one maybe two values (for emel, it would be emeltet and emeltetik) then it could be added on the same place where the current potential form is since in the new system that cell was removed. If you want to add a full conjugation table for the causative form, that might be a little too much. Panda10 (talk) 14:31, 8 February 2024 (UTC)

I only thought of a single form (I'm not sure about the passive though), not another table, since it should also have its own potential table (just like the passive form). Adam78 (talk) 14:56, 8 February 2024 (UTC)

I think I've implemented it. The parameter is "caus"; the template will automatically use the short form (short stem) for -at/-et and the long stem (form) for -tat/-tet. The cell header "Causative" and its value below are displayed only if a parameter is supplied. Maybe there could be an optional "caus2" for cases like nyomat/nyomtat. Adam78 (talk) 16:25, 18 February 2024 (UTC)

Custom conjugation templates

@Adam78 I moved the new templates to Category:Hungarian verb inflection-table templates - test and I wanted to test a custom template, as well, so I created {{hu-conj-alkuszik2}} by copying the original and changing hu-conj to hu-conj2 at the top. Then I changed the template name in alkuszik but it did not work, so obviously this minor change will not be enough. Have you thought about how the custom templates will be changed? Or we just go with adding {{hu-conj-potential-hat}}/-het with an updated header? I know we talked about this. I'm thinking about the implementation but it's getting confusing. Panda10 (talk) 18:57, 8 February 2024 (UTC)

I thought we agreed on the latter option. The content of hu-conj2 (even if we call it hu-conj-alkuszik2) is not capable of handling the particular forms of this verb; that's why hu-conj-alkuszik as such was created in the first place. Adam78 (talk) 19:08, 8 February 2024 (UTC)

@Adam78 I didn't use {{hu-conj2}} instead of {{hu-conj-alkuszik}}. I copied the content of {{hu-conj-alkuszik}} into {{hu-conj-alkuszik2}} and inside this latter template, I changed the hu-conj call (first line in the template) to hu-conj2. Yes, we did agree that we will just add the potential templates after the custom templates. There are 111 verbs using these templates. But we also agreed that we will not keep two sets of templates. So I assumed that the seven templates that you modified will be replacing the old templates. These are: hu-conj, hu-conj-unified, hu-conj-unified/doWork, hu-conj-ok, hu-conj-ek, hu-conj-ök; I think hu-conj-suf and the new hu-conj-suf2 serves different purposes, so those will be kept. It seems that hu-conj has to stay along with hu-conj2. We can't implement the new templates until we clarify all issues with all existing templates. Panda10 (talk) 19:45, 8 February 2024 (UTC)

@Adam78 What if we approach this in a different way? Actually, we did talk about this before. Instead of modifying the main conjugation templates, we could just add {{hu-conj-potential-hat}} and {{hu-conj-potential-het}} to the lemmas, under the current hu-conj template if we use a new parameter that would allow an accurate table title: "potential conjugation of PAGENAME" for lemmas and "conjugation of PAGENAME" for -hat/-het forms. You mentioned adding derived=y, but this would require us to go through every single -hat/-het forms that use this template and add the new parameter. I think it would be better to use lemma=y which would leave the current -hat/-het forms intact and when we add the template to the lemmas (which would have to be done anyway), then we add this new parameter. Would you be interested in this solution? Panda10 (talk) 17:28, 11 February 2024 (UTC)

Yes, I thought of it too, I completely agree. Adam78 (talk) 23:58, 11 February 2024 (UTC)

I think I created the updated versions of all verb conjugation templates. The tables of the potential forms still need to be updated.

Some changes for van: I deleted the empty lines for definite and 2nd-object forms; I added a switch that only adds the negative adjective (lehetetlen) for van if no prefix is supplied; and I also added the parameter "no-pos-adj" for cases when the positive adjective should be suppressed—it is displayed by default (e.g. hátralehető exists; meglehető and odalehető may not exist). The negative form nincs is only displayed now if its order is changed with a prefix, but maybe this form should be included somewhere in the conjugation table for the unprefixed form as well (even if it's amply described in its Usage notes). Also, the past conditional (lett volna) is currently displayed for van and its prefixed forms; perhaps there should be a note about forming these constructions for other verbs as well.

I widened the cell of the adverbial participle whenever the causative is omitted, but the view may become dense if the causative is present. I think I added only one more feature, which I indicated briefly in the documentation of the given conjugation template. As far as I can recall, the only change seems to be the introduction of "pref". Let's keep in mind that, as an exception, {{hu-conj-suf2}} is not meant to replace {{hu-conj-suf}}, but it's additional, so the latter needs to be kept too, rather than overwritten like the other "old" templates (once they are deemed to be ready). Adam78 (talk) 21:14, 9 March 2024 (UTC)

@Adam78 Unfortunately, something came up and I won't be able to spend time on the project. Panda10 (talk) 16:41, 11 March 2024 (UTC)

All right. Best wishes. Adam78 (talk) 18:07, 11 March 2024 (UTC)

Fakabát

First of all, thanks for checking my entries, I really appreciate your help! Yet, it was weird for me that you added the template {{hu-pos-otok}} to fakabát. Technically speaking it's correct, but I can't imagine a situation where these forms would actually be used. Only the police has cops, and when they communicate they don't really use such slang terms. Drkazmer (talk) 12:40, 27 March 2024 (UTC)

@Drkazmer Theoretically, the possessive forms could be used for the second sense. In cases like this, you can always add a comment under the Declension header to clarify the usage. Panda10 (talk) 16:52, 28 March 2024 (UTC)