Template talk:descendant

Calques and semantic loans

@Victar: hello. Do you think we should consider w:semantic loans as well? I'd like to add Latin aspectus to the descendants of Russian вид (vid) (in the grammar sense), and French souris to the descendants of English mouse (the computer thingy). --Barytonesis (talk) 12:04, 24 October 2017 (UTC)

I think French souris is a calque of English mouse, so yes in that case, but I don't think the same for Latin aspectus/Russian вид (vid). Might be more of a question for the WT:BP. --Victar (talk) 14:46, 24 October 2017 (UTC)

@Victar: haven't gotten around to doing it yet. Another suggestion: could we rename the parameter {{{calq}}} to {{{cal}}}, to align it with {{cal}}? --Barytonesis (talk) 11:52, 1 November 2017 (UTC)

@Barytonesis: You can user either; both link to {{calque}}. --Victar (talk) 15:25, 1 November 2017 (UTC)

@Victar: Oh, I hadn't noticed {{calq}} was fine as well, alongside {{cal}}. But what I meant is that currently, only {{desc|xx|xxx|calq=yes}} works, while {{desc|xx|xxx|cal=yes}} gives out a module error. --Barytonesis (talk) 15:30, 1 November 2017 (UTC)

@Barytonesis: Ah, OK. I'm not a fan of |cal=, as I think it's too generic. I'd prefer |clq= if we were to shorten it. --Victar (talk) 17:18, 1 November 2017 (UTC)

@Victar: I'd be fine with that! --Barytonesis (talk) 17:40, 1 November 2017 (UTC)

@Barytonesis: I know there is a way to get a list of all transclusions of {{desc}} with |calq=, but I don't recall how. --Victar (talk) 01:13, 2 November 2017 (UTC)

@Barytonesis: |clq= has been added and made the default. --Victar (talk) 21:45, 12 November 2017 (UTC)

der=1

@Erutuon: The arrow generated by |der=1 is too similar to what we use for borrowings. Is there a better way to format them? Do we have any precedent for using this symbol before this template was created? --WikiTiki89 21:33, 7 December 2017 (UTC)

@Wikitiki89: Honestly, the symbols look visually distinct enough to me, but they aren't particularly clear about what they mean. They were created only for this template. I don't have any ideas for improvement. I just looked for two other shapes of arrow symbol in Unicode that might be better, but couldn't find any. — Eru·tuon 22:01, 7 December 2017 (UTC)

What about U+27FF LONG RIGHTWARDS SQUIGGLE ARROW? --Barytonesis (talk) 22:08, 7 December 2017 (UTC)

@Erutuon: I think you misunderstand the problem I see. It's not that you can't tell them apart, it's more that it's confusing that two similar symbols mean different things. To me, the arrow for borrowings is a natural choice, but an arrow somehow does not intuitively indicate that there is some morphological step missing. I was thinking we could give it a more explicit tag like (indirectly) or something. --WikiTiki89 22:11, 7 December 2017 (UTC)

I'm also partial to keeping it as is. I've often seen ⇒ used to denote indirect descent, so it's a very familiar symbol, at least to me. --Victar (talk) 22:18, 7 December 2017 (UTC)

Ok, so that's what I was asking about. If it's a common convention, that's fine. I've just never heard of it. Perhaps we should add hover text or a link to the symbol to clarify its meaning. We could do the same for the borrowing arrow as well. --WikiTiki89 22:20, 7 December 2017 (UTC)

Perhaps. Also, just to clarify, ⇒ was being used in the project for this purpose long before this template. This template only formalized that usage. --Victar (talk) 22:26, 7 December 2017 (UTC)

I added tooltips. Revise them if you like. — Eru·tuon 22:36, 7 December 2017 (UTC)

I have no objections. Thanks, Erutuon. --Victar (talk) 22:44, 7 December 2017 (UTC)

I'm fairly certain I was the first person to use ⇒ to mean language-internal derivation on a PIE page. I may be misremembering, but I think I came up with that format as well as the *: used for listing dialectal forms. If these were being used elsewhere, perhaps they evolved multiple times, but I'm fairly certain I started doing them independently. —*i̯óh₁nC^[5] 12:42, 8 December 2017 (UTC)

Let’s use ⥅ U+2945 RIGHTWARDS ARROW WITH PLUS BELOW with large font size. The plus shows that something additional has happened. Or if it not enough: ➕ U+2795 HEAVY PLUS SIGN – it often appears colored. Maybe the heavy plus sign as a CSS hack under or in the now implemented arrow. Palaestrator verborum (loquier) 22:25, 7 December 2017 (UTC)

I very much hate that idea. --Victar (talk) 22:26, 7 December 2017 (UTC)

@DTLHS, see discussion above. --Victar (talk) 19:59, 10 April 2018 (UTC)

I really don't give a fuck about the discussion. @Calthinus, you're the one who wanted the change. DTLHS (talk) 20:03, 10 April 2018 (UTC)

I'm very confused. What exactly is going on here and how can I help? I know I did cause a change yesterday, because I raised the issue that using just an arrow to note that a word is a loan rather than an inherited descendant is not clear enough to users of Wiktionary who are not familiar with the notation. --Calthinus (talk) 20:17, 10 April 2018 (UTC)

@Calthinus, yes, and the compromise, which you can read above, was to include a tooltip with the arrow. Anything beyond that would not have my personal support. --Victar (talk) 21:26, 10 April 2018 (UTC)

This is all very opaque for me as I don't know a lot of the referents in this discussion. Nevertheless I'd like to reiterate that I think that the most important thing should be how easy it is for readers to understand it. Not Wiktionary editors. If it is not clear that an arrow means borrowing, or anything like that, it should be changed so that it explicitly says what it means. As I have elaborated in other incarnations of this discussion, the distinction between a borrowing and an inherited word is of utmost importance in many cases, such as for Historical Phonologists. If we are using different shaped arrows to mean language internal derivation and borrowing, well that's even worse. Things should be stated explicitly, not using idiosyncratic wingdings that you have to be an editor of Wiktionary to understand. --Calthinus (talk) 14:05, 11 April 2018 (UTC)

@Calthinus, which is why we added the mouseover tooltip. Please read above. --Victar (talk) 14:48, 11 April 2018 (UTC)

The "mouseover"... isn't working for me. All I see is an arrow. Moving my mouse over it does nothing. And I've read all the stuff above; I simply don't understand all of it (example: what is the relevance of the plus signs?)-- which I am too busy for at the moment. What I observe is that on pages like beneficium all there is for French (a loan) is an arrow. Nothing else. --Calthinus (talk) 15:32, 11 April 2018 (UTC)

Null |2= with other parameters enabled

I need to be able to null out |2= whilst still being able to use the parameters, ex. {{desc|pal|-|t=high|ts=buland}} => Middle Persian: (/buland/, “high”), as seen here. Could someone help with this? @Erutuon, Crom daba, Benwing2, JohnC5? --Victar (talk) 21:05, 20 June 2018 (UTC)

It depends on the template in question. There's no general solution. BTW why do you need this? Benwing2 (talk) 21:09, 20 June 2018 (UTC)

Sorry, I realize you're talking specifically about {{desc}}, and I think param 2. The current behavior appears intentional; not sure if anyone depends on it, though. I tried specifying "" as the value and it does appear to recognize it as "unspecified" but you get a script error. Benwing2 (talk) 21:14, 20 June 2018 (UTC)

@Benwing2: {{desc}} is the template in question. We use it when all the lect forms below have the same transcription and/or definition. Right now we're writing it out manually, ex. {{desc|pal|-}} (/buland/, "high"), but that's not preferable. --Victar (talk) 21:15, 20 June 2018 (UTC)

@Benwing2: Hah, well I can tell you know one depends on the parameters being disabled with null |2= within {{desc}}. My only concern is that a change like this would need to be made closer to the root, perhaps in {{link}}. I don't really know where |2=- is being interpreted. --Victar (talk) 21:25, 20 June 2018 (UTC)

@Victar: |2=- is interpreted in the descendant() function that implements this template. If 2=- then you get no link at all according to the code, otherwise full_link() is called, and that function adds the [Term?] stuff. I'm not quite sure how you'd specify the behavior you want without it breaking something, although I can add tracking code to see where 2=- is actually used. We might be able to special-case things though if you specify 2=- along with some other params. Benwing2 (talk) 21:30, 20 June 2018 (UTC)

I made a simple elseif ts or gloss then remove the empty term using gsub but I can't seem to get it to work. Maybe someone could have look, @Benwing2, Erutuon? --Victar (talk) 15:23, 21 June 2018 (UTC)

@Erutuon, thanks for the escape characters. I'm still weirdly getting {{desc|pal|-|sc=Latn|t=big|ts=šahesprahm}} => Middle Persian: -. --Victar (talk) 18:24, 21 June 2018 (UTC)

@Victar: Fixed. It's not a durable method, though, as it relies on Module:links generating specific text when the term and link text are missing. — Eru·tuon 19:35, 21 June 2018 (UTC)

Thanks, @Erutuon. It's not the cleanest method, but better than having to manually type it in. --Victar (talk) 19:42, 21 June 2018 (UTC)

|g= from {{alter}}

@Erutuon, {{desc}} doesn't seem to be pulling |g= from {{alter}}, i.e. Ashokan Prakrit: *𑀓𑀼𑀢𑁆𑀢 m (*kutta, “dog”). Same goes for {{desctree}}. Could you have a look? --Victar (talk) 02:17, 16 August 2018 (UTC)

@Victar: Fixed. Turned out to be because of a typo. — Eru·tuon 03:54, 16 August 2018 (UTC)

@Erutuon:, as always, many thanks! --Victar (talk) 04:04, 16 August 2018 (UTC)

Descendants from other forms than the basic

How to show this? For example see Swedish mycket and Danish verden, how would you show this on Proto-Germanic *mikilaz and Proto-Germanic *weraldiz, respectively? It seems as no parameter is able to show it (der, bor, clq).Jonteemil (talk) 16:14, 1 August 2019 (UTC)

@Jonteemil: It depends on the etymology. With the case of mycket, if the neuter formation dates back to Old Norse, you could create an Old Norse and place the derivatives of that form there instead. If it just a general free formation found throughout Nordic languages, then just placing all together probably makes more sense. --{{victar|talk}} 18:06, 1 August 2019 (UTC)

As a rule, a descendant is valid if it descends from any of the forms found in the inflectional paradigm of the ancestral language. In this sense, a lemma can split into multiple lemmas through deflexion. Descendant lists should only give lemma forms, and since in Old Norse mikit is simply an inflection of mikill, it should not be listed separately among the descendants. Rather, it is a term with multiple Swedish descendant lemmas.

The same principle applies to Danish verden as well. Old Norse verǫldin is again simply an inflection of verǫld. Since the latter is the lemma, and only lemmas should be given in descendants, it is totally valid to say verden is a descendant of verǫld. —Rua (mew) 19:07, 1 August 2019 (UTC)

@Rua I agree that verden is a descendant of verǫld, but not directly. I see some issues with not having a parameter to indicate the word didn’t descend from the basic form. A use of T:desc without any parameters such as der, clq, or bor, at least to me, indicates that a word descended from its ”parent” through normal sound changes. When a reader sees that verden descendes from verǫld without any parameters used he or she may get confused and think that it is a normal sound change in Danish to add -en from Old Norse nouns, which is incorrect.Jonteemil (talk) 00:36, 2 August 2019 (UTC)

But the lemma form of many Scandinavian nouns descends from the accusative, not the nominative in -r. That -r is otherwise kept can be seen in verbs, and of course in the plurals of nouns too. So if you were to apply your principle here too, then we'd have to show the accusative form of these nouns separately for words where the modern form is from the accusative.

Moreover, you have to account for general cases where the lemma form changes from one language to the other. For Indo-European, we have chosen the third-person singular form of verbs as the lemma, but this is clearly not the lemma in all of its descendants. Proto-Slavic had an infinitive, which is the lemma, but Bulgarian and Macedonian both lack an infinitive and use another form of the lemma. Latin uses the first-person singular as the lemma, while the modern Romance languages use the infinitive. Also for Romance, the modern lemma typically descends from the Latin accusative. In all of these cases, they are simply shown as descendants of one lemma to the other, without any weird hoops.

In these cases it can be argued that the change is regular and predictable, while verden is a single anomaly. In that case I am ok with mentioning the anomaly, but I disagree that we should list more than one form of the Old Norse lemma in the descendants. Also, note that in *skōhaz, the Dutch lemma descends from the Middle Dutch plural, yet no mention of this is made here either. So this doesn't seem like the big issue you are making it out to be. —Rua (mew) 08:44, 2 August 2019 (UTC)

@Rua In the cases where the lemma form changes from a mother to a daughter language, for example Latin expulsō -> Spanish expulsar the lemma endings are regular and expected as you mention so there is no risk of confusion there. Since I yet haven’t proposed what the parameter would generate (arrow, text etc.) I don’t know why you wrote this: … then we'd have to show the accusative form of these nouns separately for words where the modern form is from the accusative.. It is not the only way of showing it. I agree with you that showing 2 or more forms of a lemma because the daughter words descended from different inflected forms may be a bit bothersome so I propose a parenthasis after (or before?) the word instead. Not sure if this is best done with a parameter or just T:qualifier though. Lastly, the fact that something exists, doesn’t imply that it isn’t ”the big issue” I’m ”making it out to be”. At least, I don’t buy that analogy🤷‍♂️.Jonteemil (talk) 22:33, 2 August 2019 (UTC)

I would just use {{qualifier}} to indicate the anomalies. An entirely new parameter for {{desc}} is overkill for these rare cases. —Rua (mew) 22:56, 2 August 2019 (UTC)