Wiktionary talk:About Luxembourgish

Eifeler Regel

@JohnC5 How is the Eifeler Regel handled in inflection tables? Since Luxembourgish spelling follows the pronunciation in this regard, all the verb forms ending in -n have alternative forms without it. There is a note at the bottom of the table, but this seems to be only about the auxiliary verb and not the main verb. —CodeCat 01:40, 11 July 2016 (UTC)

@CodeCat: What a great question—one for which I haven't the faintest idea. My knowledge of Luxembourgish is very recently acquired and accordingly scant. I have noticed that the conjugation tables are fairly unwieldy, and I would love to redo them with Lua. Perhaps @Kolmiel, Lo Ximiendo, BigDom have more knowledge on this matter. —JohnC5 01:50, 11 July 2016 (UTC)

I can make a Luacised version, but I would also like to remake the table a bit. I don't think periphrastic forms should be in the table, since there's potentially many many combinations you can come up with. The German inflection tables put them in their own separate table, and the Dutch ones don't list them at all. So I would like to at least split them into two tables like German does. —CodeCat 01:57, 11 July 2016 (UTC)

I'd certainly support splitting them into separate tables. —JohnC5 02:05, 11 July 2016 (UTC)

Any word (be it a verb, a noun, a pronoun, a conjunction) in Luxembourgish ending in an alveolar nasal can undergo n-deletion, so potentially all these words have alternative forms that could be displayed. None of the Luxembourgish dictionaries list these as alternative forms however, simply because n-deletion is so regular. Not sure how it should be dealt with in the inflection table. BigDom 07:09, 11 July 2016 (UTC)

I'd say that in a new module we should represent the outcome of the Eifeler Regel within the paraphrastic forms and then put something like -e(n) at the ends of the form. So the 3p future perfect of doen would be *wäerten gedoen hunn → wäerte gedoen hu(nn) or wäerte gedoen hunn/hu. —JohnC5 14:56, 11 July 2016 (UTC)

My idea is for the last part, the (n) or (nn), to be a link to the form with -n, and the remainder of the word link to the form without it. So something like [[féiere]][[féieren|(n)]]. We should presumably do this for lemma forms too, in the headword: head=[[hu]](nn) to link to the n-less form. It's more compact this way than listing two forms. For the actual entries of these forms, I propose making a new form-of template for Luxembourgish that allows an extra parameter like ER=1 to indicate that the Eifeler Regel applies to this form. —CodeCat 17:00, 11 July 2016 (UTC)

I was hoping you'd go for the parenthesized link method, since that is the system we use in AG tables. I would, however, prefer [[féiere]]([[féieren|n]]) with the parentheses not in the link text instead. —JohnC5 17:36, 11 July 2016 (UTC)

A question: Does Luxembourgish have a present participle? If not, what happened to it and how is it expressed instead? —CodeCat 17:55, 11 July 2016 (UTC)

@CodeCat: No it doesn't. Through the development -nd- > -nn- it became homophonous to the infinitive and was done away with. There are a few lecialized forms (e.g. rosen). They use a relative clause instead (as in French). New lexicalized forms with the German ending -end are sometimes introduced, but only as adjectives translating German adjectives. Kolmiel (talk) 18:01, 11 July 2016 (UTC)

For example "am lafende Joer" for German "im laufenden Jahr" (in this year, in the current year). Kolmiel (talk) 18:09, 11 July 2016 (UTC)

1.) Luxembourgish inflection tables must be revised anyway. The verb one is correct but messy (as you noted), while the adjective one misses a lot of forms (strong dative, independant plural in -er, superlatives, etc.). 2.) It is not true that all words ending in -n undergo n-deletion. For example, Mann does not ever become "*Ma". It is a lexicalised thing! However, it is safe to say that final -n is always deletable as a grammatical ending. 3.) I would be in favour of doing away with periphrastic verb forms in the tables. Apart from the fact that I think them unnecessary in German as well, in Luxembourgish there's simply too much variation. For example the conditional tense (I would say) can be "ech géif soen" (lit. I gave say), "ech géing soen" (I went say), or "ech déit soen" (I did say). The "géif" one is the most common, but the other two are standard Luxembourgish and at least the former of them is very common, too. 4.) In synthetic forms, I think, the n-less forms should automically be derived as variants given next to or under the one with -n. Kolmiel (talk) 16:24, 11 July 2016 (UTC)

I wonder if for the conditionals we could list the form as "ech géif/géing/déit soen" for brevity? —JohnC5 17:36, 11 July 2016 (UTC)

Maybe. If you guys want these lists, there'll be a way of course. As I said, the amount of variation in Luxembourgish is another argument for me against them, but essentially I just find them redundant to begin with. — At any rate, priority should be to make the synthetic forms complete. As I mentioned, the adjective table is very incomplete. But the verb table is also incomplete for separable verbs. It misses subordinate-clause forms (androen > andro, andréis, andréit, androt) and the composed infinitive (> anzedroen). Kolmiel (talk) 17:58, 11 July 2016 (UTC)

New Lua-based inflection

I've created Module:lb-verbs and changed {{lb-conj-regular}} to use it. It doesn't do anything differently from the old template, so all the parameters should still work the way they used to and are still necessary as before. The table layout has been changed, though, only actual forms (words) are listed, not the periphrastic ones. We can add those back in later if there is a need/desire for them. I would suggest something compact like {{sl-conj-table}}, where only the basic formula of periphrastic forms is shown, not entire paradigms.

Now that the table is implemented in Lua, I would like to know which parts of the inflection are entirely predictable from spelling (final consonants and such), and therefore which parameters are redundant once the module is capable of inferring them itself. —CodeCat 18:57, 11 July 2016 (UTC)

If I understand you request correctly: a) Stem-final -w- becomes -f- unless followed by a vowel. b) Stem-final -d- (-dd-) becomes -t- (-tt-) unless followed by a vowel. c) After -t (whether original or from -d-) the ending -t is dropped. d) Single vowels a, i, u, o are doubled when followed by two or more consonants (e.g. molen, imperative mol, but du mools, hie moolt, etc.). However, when the a, i, u, o are preceded by another vowel it depends on the pronunciation. (An example from adjectives: naiv has the neuter naiivt, but if -ai- were a diphthong it would not be changed.) d) In regular verbs: infinitive = 1st p. sg. present = 1st and 3rd p. pl. present. And: 3rd p. sg. present = 2nd p. pl. present (Except for 1st = 3rd pl. none of these necessarily apply in irregular verbs). Kolmiel (talk) 19:20, 11 July 2016 (UTC)

That's helpful, but I was also wondering about the pref-n, pref-cons, pref-inf-n and ën parameters, which are also present on {{lb-conj-regular}}, as well as the 4th and 5th parameter which apparently are only ever "s" and "t" respectively if they are provided at all. —CodeCat 19:23, 11 July 2016 (UTC)

"ën" probably means that -ën is used instead of -en, which is regular for stems ending in -ee (leeën), but might occur in other cases as well (I don't know right now). The "pref-" ones refer to the Eifeler Regel with separable prefixes. Most or all of it would be predictable, but I don't fully understand how these parametres interoperate. The "t" parametre is the omission of -t described above under c). The "s" parametre is the omission of -s in the 2nd p. sg. which is regular after -s, -x, -z in the stem. Kolmiel (talk) 19:42, 11 July 2016 (UTC)

The "ën" parameter was used when the verb stem ends in -ee (example seeën), so this is completely predictable. The "s" parameter simply stopped the template duplicating the -s ending in the second-person singular. The "t" parameter dealt with the case Kolmiel mentioned above where the -t ending is dropped if the stem ends with -t or -d.

The other parameters dealing with "n"s were (admittedly unwieldy) ways of dealing with deleted n in separable prefixes. For example, the prefix an- undergoes n-deletion when the first consonant of the verb stem triggers the Eifeler Regel. However, this -n- reappears in the separated forms (e.g. abeezen in the infinitive but beezt an in 3.s.pres). The "pref-cons" parameter deals with when the separable prefix itself begins with a consonant that triggers Eifeler Regel (for example, matbréngen would have brénge mat in 1.s.pres rather than bréngen mat). I can't remember or figure out what "pref-inf-n" could possibly be used for. BigDom 20:01, 11 July 2016 (UTC)

The thing with separable verbs, though, is that the two parts can have more parts in between. In that case, presumably the Eifeler Regel is triggered by the word that actually follows, which may be the separable part, or it may be something else. So with matbréngen it's still possible to see bréngen mat if another word with the appropriate initial sound stands between them. So it may make more sense to show (n) even here. —CodeCat 20:05, 11 July 2016 (UTC)

I would agree with you, and that's why I didn't use that parameter even when it could have been employed (e.g. matspillen). BigDom 20:08, 11 July 2016 (UTC)

These new templates look great! I might ask, however, that if we don't add the periphrastic forms, we create an appendix that gives all of the possible constructions and link to it from the template. —JohnC5 04:38, 12 July 2016 (UTC)

Why do denken and bréngen have no -t in the 3rd singular present? They are shown on the conjugation on the "Luxogramm" site. —CodeCat 21:51, 11 July 2016 (UTC)

They should have -t there, not sure why they don't. Is it because of our template? BigDom 05:24, 12 July 2016 (UTC)

They definitely do have the -t in all 3rd-person-singular forms. At Luxogramm (at least where I looked) they don't lack it in the present, but they lack it in the subjunctive. That's also a mistake. Kolmiel (talk) 16:26, 12 July 2016 (UTC)

Luxembourgish (lb) and Central Franconian (gmw-cfr)

I saw today that CF has been set as an ancestor of Lux. I don't know when and by whom, but at any rate I think it's good and correct! Lux branched off from Moselle Franconian and the absolutely earliest term that could be set for this development is 1815, when the modern border between Luxembourg and Prussia/Germany was established. So making Lux a direct descendant of Middle High German (as we had at some point) was a major anachronism.
On the other hand, we must be aware that there is not much attestation for CF between the end of MHG and 1815. So if we put some modern CF lemma as the ancestor of a Lux lemma, this is also an anachronism. And it is even entirely wrong if we give a form that exists in the dialect of Koblenz or some other place, but is not precedent to the Lux form. Therefore I think the best and safest will be to use "gmw-cfr" only in the "dercat" template of Lux entries, but not in the text of an etymology.
I will do things that way from now on, and I wanted to put my reasoning here for future editors. (Obviously you can take a different dicision, but please do take the above into account.) 88.64.225.53 17:35, 15 January 2024 (UTC)