User talk:Metaknowledge

Definition from Wiktionary, the free dictionary
Jump to navigation Jump to search
Archive
Archives
  1. 2012                5. 2016                9. 2020
  2. 2013                6. 2017
  3. 2014                7. 2018
  4. 2015                8. 2019

reverted taxlinks[edit]

What "broken links"? DCDuring (talk) 23:11, 7 January 2021 (UTC)

Did you try clicking the link? —Μετάknowledgediscuss/deeds 23:49, 7 January 2021 (UTC)
Of course. It yields the page that allows one to search for the collocation on other pages, where one finds wikispecies:Ficus pumila, the named variety of which is apparently the source of a popular Chinese confection, with several links, four of which are now not to be counted. You should probably reverse all of the counted ones as part of your effort purify our links. Not to mention all the taxonomic names so liinked. DCDuring (talk) 03:48, 8 January 2021 (UTC)
It yields a page that does not exist. You changed a link that goes to an informative entry to a link that goes to an entry that does not exist. What do you care about more: our readers or "purifying" links? —Μετάknowledgediscuss/deeds 04:43, 8 January 2021 (UTC)
"Broken links" refers to links to target web pages that previously existed, but no longer exist. "Dead links" can refer to "broken links" or to links to web pages that don't exist. The links from all wiktionary pages with {{taxlink}} always go to a web page that exists. The landing page may not be an entry, but it is a page that enables the user to search for the collocation in the link. Almost all the time something can be found. So users can get information AND I can keep track of missing organism names, gradually adding the ones that have the most pages that would 'want' to link to them. DCDuring (talk) 16:16, 9 January 2021 (UTC)
Then replace the word "broken" with "dead" in all my previous statements. Your keeping track of missing names is honestly much less important than using a link that gives a reader information, as opposed to a link that doesn't. —Μετάknowledgediscuss/deeds 17:31, 9 January 2021 (UTC)
I think we need some version of taxlink to be used in cases where there's no entry at the page directly linked to. You have no idea of the magnitude of the problem: I've been creating such links for years on a massive scale because they're better that redlinks. It's true that the place where they land is Wikispecies' 404 page rather than an actual entry, but, as DCD says, one can search for the entry from there. I would like to suggest that we add a couple of parameters: |nows= (or a better name if we can come up with one) should provide a redlink, but with the categorization that DCD uses, the same way that {{vern}} behaves. |wp= should link to the name in the first parameter on Wikipedia if the value is "1" and link to whatever is in the wp parameter if it isn't. Chuck Entz (talk) 17:54, 9 January 2021 (UTC)
@Chuck Entz If you could specify what we need, maybe there is someone willing to take it on. As soon as modules are involved, I am lost. DCDuring (talk) 00:48, 10 January 2021 (UTC)
The links are neither dead nor broken. If your are going to be curt, at least be accurate. DCDuring (talk) 00:01, 10 January 2021 (UTC)
Right, distinguishing different kinds of links that don't lead to information is the real problem here. My mistake. —Μετάknowledgediscuss/deeds 01:10, 10 January 2021 (UTC)

Blocking of User:Returning2stadia[edit]

Hi. Why did you infinitely block to Returning2stadia? What nonsense/gibberish things did Returning2stadia do? Regards. --Vivaelcelta (talk) 01:17, 16 January 2021 (UTC)

User:AryamanA/WonderfoolΜετάknowledgediscuss/deeds 07:46, 16 January 2021 (UTC)

Attestability of some Indian terms[edit]

Hello. You asked at puttun "are you sure the stuff you're adding is attested?". That one might be a bit borderline. I do check for some basic level of existence. I've also occasionally added terms that mostly seem to appear in italics (i.e. not properly adopted as a loan word) which I suppose I didn't do in the past. My current thinking is (i) if it's from a language that doesn't use Latin script, but can be found in Latin script, italicised or not, that suggests some kind of level of adoption; (ii) I'm not doing any real damage because this stuff can always be RFVed and perhaps it will fail and get deleted. That's okay. That's why RFV exists. But it also might pass RFV, and if I hadn't added it then perhaps nobody else would have done, and we'd be missing it. Short answer: I'm still quite conservative about what I add, so probably no big deal. Equinox 14:11, 21 January 2021 (UTC)

@Equinox: I disagree on (i), because italicisation seems like a good rule of thumb, and on (ii), because RFV has enough work to do as it is and doesn't need people who know better adding to the problem. But I hope you're right about this... —Μετάknowledgediscuss/deeds 17:05, 21 January 2021 (UTC)
Haha. No panic. I am a small-scale creator. Sometimes they are things I found in OED's "new words" list and I think "oh yeah we should have those!". I don't automatically blindly add them though. I wish you could see my Wiktionary folders and subfolders. I can get on here any day and start tackling Blotto's chemical list. But I also have a zillion things I won't finish before I die. Also my employer occasionally expects me to do some work. Equinox 20:11, 21 January 2021 (UTC)
You should do a livestream on Discord showing your folders and editing workflow. I might even come and watch it. —Μετάknowledgediscuss/deeds 22:19, 21 January 2021 (UTC)
I would have to put a rigid firewall in place between Wiktionary stuff and...other things I visit in my browser. Equinox 12:05, 22 January 2021 (UTC)
Of course. It wouldn't do for us to find out that you secretly edit Encyclopaedia Dramatica in your free time. —Μετάknowledgediscuss/deeds 16:54, 22 January 2021 (UTC)
I'm the Turing of my day. Doing unacceptable things with a computer. Oh wait, he was doing unacceptable things and using a computer. Cancel that, Turing was better. (How do you feel about Turing? I don't know, I've never Tured.) Equinox 06:33, 20 June 2021 (UTC)
Love you too, Eq. —Μετάknowledgediscuss/deeds 06:55, 20 June 2021 (UTC)

Here's my current list of Wiktionary work folders, to reward your curiosity. Some have languished ignored for years though.

  • A browser bookmarks folder for possible words/lists I stumble on; hovers around 100 entries; many get offloaded to my user page
  • Chambers 1908 (only ~300 entries left to look at, but long ignored as they are the long or awkward ones); plus special app for navigating these
  • "EWDC" generator code and archives (English Word Defining Club, currently dormant)
  • "Lucky dip" (mostly difficult science words and some lists from Ungoliant, those mostly netslang and historical places and weapons etc.)
  • Newspeak audio (I basically voice-recorded this entire 1984 jargon dictionary, a long project! ~4 hours still to review; most is done)
  • OED new words (lists of their additions 4 times a year since 2000, usually without definitions; I only bother with some of these)
  • Quercus words antiblued (remnants from a good list in Quercus Solaris' userspace)
  • Sowpods (a huge old Scrabble word list)
  • And various odd files: animal verbs; blends; dump processing ideas; maths terms ending in "-variant"; "more organic compounds" (the list that never ends, because whenever you google one you find ten more); John Camden Hotten's old Slang Dictionary (one big HTML file); and video game terms from Wikipedia. Equinox 16:19, 27 January 2021 (UTC)
Update: "Newspeak" is down to about 2 hours. dear god. Equinox 07:06, 20 June 2021 (UTC)

Request- Deletion of User Pages[edit]

Hey Metaknowledge! For several months I have been thinking about "quitting" Wiktionary, and I am proud to say that I have finally determined that I am done with Wiktionary, Wikipedia, Wiki Commons, etc. It's not you, it's me! I just wanted to make some headway for others to build on and I did do that. I would like to find some way to clear/delete my user pages, talk pages, archive talk pages, and sandboxes on English Wikipedia, Mandarin Chinese Wikipedia, English Wiktionary and Wikimedia Commons (which are pretty embarrassing as a whole). Do you know someone who can do this for me? I have no idea if I might "come back", but I just think ending now is better than continuing and likely getting blocked in the near future for aberrant behavior or annoyingness, to which I am prone. Thanks for your guidance and help over the course of my editing. --Geographyinitiative (talk) 18:27, 1 February 2021 (UTC)

You can use {{delete}} on any personal user pages and an admin will delete them. However, user talk pages should be preserved so there is a record of those conversations. —Μετάknowledgediscuss/deeds 18:36, 1 February 2021 (UTC)
"I have finally determined that I am done with Wiktionary". I'm not supposed to say this but there is a really easy way to get permabanned, which is making a legal threat. I haven't gone so far myself yet but some day when I decide I am done, I will probably do that. You could merely write on Jimbo's talk page "I am going to sue you in the courts of England and Wales, for having a sucky haircut" and they will descend on you like flies, and block you forever. Most sins can be forgiven (Catholicism) but legal threats cannot. Equinox 07:26, 20 June 2021 (UTC)

lad-proper noun[edit]

At Espanya and איספאנײה‎, do you know that it's female? It's definitely grammatically consistent but I can't find any proof that it is. —Justin (koavf)TCM 07:44, 2 February 2021 (UTC)

Ladino is grammatically almost identical to Old Spanish. But if you need proof, all you have to do is do a quick Google search. Within a minute, I found this poem that uses feminine agreement. —Μετάknowledgediscuss/deeds 18:30, 2 February 2021 (UTC)

زمین اور آسمان [edit]

Hi, I saw your message on the edit history! I've stopped romanising the َو sounds as 'au', to make it more closer to the Urdu spelling. I know it's different to the Indic-based translations, but I feel the Indic transliterations methods aren't really for Urdu. I'm also doing the same for ی, for instance ایک which is written as "aik" in Roman Urdu, more often than "ek", only in the initial form, I should mention.
-Taimoor Ahmed(گل بات؟) 19:10, 4 February 2021 (UTC)

@Taimoorahmed11, AryamanA: We aim to present a consistent, professional product, and that means you can't just change your romanisation scheme because you feel like it. All entries should follow WT:UR TR, and if you disagree with any part of it, that needs to be a discussion with the Hindi/Urdu-editing community, rather than something you do on your own. —Μετάknowledgediscuss/deeds 19:16, 4 February 2021 (UTC)
@Metaknowledge I wholeheartedly understand, but WT:UR TR needs to be updated as it doesn't explain Urdu spelling in detail so I assumed it wasn't being followed, hence why I didn't take it into consideration.
-Taimoor Ahmed(گل بات؟) 19:27, 4 February 2021 (UTC)
@Taimoorahmed11: You knew you were changing the romanisation scheme, so you knew you were introducing undocumented inconsistency. That's what we want to avoid. If the page needs to be updated, then there should be a discussion involving all relevant editors. —Μετάknowledgediscuss/deeds 19:39, 4 February 2021 (UTC)
@Metaknowledge In that case, I'll revert my edits of ao and ai (if you could also, if you find any?), and will, later, propose them on the WT:UR TR page first as well as to update the page as whole.
-Taimoor Ahmed(گل بات؟) 19:51, 4 February 2021 (UTC)

Is it a whoozit or a thingamajig?[edit]

Could you check my edits to cavosteliid? It looked like Semper's definition was too narrow, but anything that involves nomenclature of slime molds tends to be as amorphous and hard to pin down as the ...uhh... whatever-they-are s themselves. I just don't have the background to do more than guess. Also note what I did with the link targets for the templates. Chuck Entz (talk) 23:33, 5 February 2021 (UTC)

@Chuck Entz: No idea what your second def was supposed to be, but the way it was worded (and misspelt) made it wrong. Your UN looks good, though, and should probably be templatised. We also need a better category for these guys... —Μετάknowledgediscuss/deeds 01:51, 6 February 2021 (UTC)

Outdated template and link[edit]

Hi, please solve this problem. In Help:Namespace there is a link to an outdated page in meta. --BoldLuis (talk) 17:36, 2 March 2021 (UTC)

It was an obsolete page with no information, so I just converted it into a redirect. —Μετάknowledgediscuss/deeds 17:44, 2 March 2021 (UTC)

Accel for Nepali[edit]

Could you fix this? Can you explain why don't Conjugations/inflections of Nepali entries support Accelerated link (green links)? See गर्नु. Kushalpok01 (talk) 04:02, 7 March 2021 (UTC)

Not sure, but assuming you put the code to generate green links in the module, you might need to add a Nepali subpage to Module:accel. —Μετάknowledgediscuss/deeds 07:30, 7 March 2021 (UTC)

Emptying categories[edit]

Did you notice that you did this at Yuganda? —Justin (koavf)TCM 03:00, 10 March 2021 (UTC)

Yep. That's not an appropriate use for those categories. —Μετάknowledgediscuss/deeds 03:05, 10 March 2021 (UTC)
How so? "Yuganda" is a Dinka term related to Uganada. —Justin (koavf)TCM 03:06, 10 March 2021 (UTC)
No, it's a term for Uganda, and it is in the proper category (for countries of Africa). It would be useless to categorise this word thus, and it would be the only word in the category, rendering it doubly useless. —Μετάknowledgediscuss/deeds 03:09, 10 March 2021 (UTC)
"Uganda" is a word related to the topic of Uganda. Please discuss this if you want to remove these categories en masse and get consensus or point me to an existing consensus. Additionally, I'm sure that there are in principle plenty of other words related to Uganda in these languages, even if we don't have the categories yet. —Justin (koavf)TCM 03:10, 10 March 2021 (UTC)
I don't want to remove them en masse; I want to remove them where merited. I want useful categories, not a maximal number of categories with one entry each. Honestly, I find this whole matter a waste of time. —Μετάknowledgediscuss/deeds 03:16, 10 March 2021 (UTC)
Which ones are or aren't useful? Again, this seems arbitrary unless you have some criteria that we've previously discussed. —Justin (koavf)TCM 03:38, 10 March 2021 (UTC)
As I said, one entry each for a category. I would like to stop discussing this now, because I actually prefer to have you revert me and leave your silly categories than have to keep being distracted from actually working on entries. —Μετάknowledgediscuss/deeds 03:41, 10 March 2021 (UTC)

Igbo verbs not in their root forms[edit]

Some of the Igbo verbs are not in their root forms, like gee ntị which is in the imperative and ịgba ọsọ which is in the infinitive. Is this fine for now? Umu igbo (talk) 13:02, 10 March 2021 (UTC)

@Umu igbo: No, we ought to fix them... the issue is that I never really learned how Igbo verbs work, so either I'll have to do some reading, or you'll have to explain it to me! —Μετάknowledgediscuss/deeds 19:35, 10 March 2021 (UTC)
Well, I'm not sure if I can explain how the infinitives work without making this a really long post, but in this case, the infinitive in ịgba ọsọ's case is marked by the prefix í- which also undergoes ATR harmony depending on the ATR quality of the root.
In this case, the root would be gbá.
In simple verb roots (that is, verbs with only one monosyllabic element in them), the imperative is usually marked with a high toned harmonizing open vowel suffix. In the case of the verb root gè, this is the suffix -é, and thus the imperative form should be gèé.
You'll notice that both entries are accompanied by a nominal. Basically all verbs in Igbo require a verb complement - a nominal that complements it. Some verbs are more semantically dependent on their VC than others. gbá and gè are two examples of verbs that require their verb complement in order to retain the sense implied when the VC is present. As far as I know, gè, for instance, doesn't mean anything unless accompanied by ntị, and gbá's meaning changes drastically depending on what its VC is.
I'm not really good at explaining stuff, but I hope this helps at least a little! Umu igbo (talk) 20:10, 10 March 2021 (UTC)
@Umu igbo: Okay, let's move them all to their roots, but it's now clear to me that we'll need an inflection-table template for verbs {{ig-conj}}, and it'll need to be done in Lua. This might be beyond my ability, so if you use Discord, that's where I'll be getting help with this. —Μετάknowledgediscuss/deeds 23:53, 10 March 2021 (UTC)
Oh ok. I don't know anything about Lua, though, so I can't really be of any assistance with that. I was only able to make the number list module because the one for Yoruba was really similar to it. Do you have anyone in mind that could help with this? Umu igbo (talk) 00:48, 11 March 2021 (UTC)
@Umu igbo, Smashhoof: I guess I'm still kind of on the fence about what counts as lexical derivation and what counts as an inflected form. I was reading Emenanjọ, and it made me think that the agent noun, noun of result, etc better belong as derived terms, leaving just the infinitive and the participle as inflected forms, and those can comfortably fit in {{ig-verb}}. The problem is that I still need Lua help in order to actually execute that. Smashhoof, I pinged you because I reckon this should be an easy one for you — would you mind helping out? Here are the forms:
  • Infinitive: prefix i before +ATR, before –ATR (tone of prefix opposite to tone of stem)
  • Participle: prefix e before +ATR, a before –ATR (tone of prefix opposite to tone of stem)
Note: +ATR vowels are <e i o u> and –ATR vowels are <a ị ọ ụ>, and there are two tones, H (marked with acute accent) and L (not marked). —Μετάknowledgediscuss/deeds 03:15, 11 March 2021 (UTC)
Maybe @Rua might be interested- she's worked with (somewhat) related languages and she certainly knows Lua as well as anybody here. Chuck Entz (talk) 04:10, 11 March 2021 (UTC)
@Umu igbo, Metaknowledge I can certainly make a module for that, next time I have some free time. Or if Rua wants to do it, feel free. Whoever gets to it first I guess. Umu igbo, it would be helpful if you can make a list of example roots and their inflections, to use as test cases for the module. Smashhoof (Talk · Contributions) 05:06, 11 March 2021 (UTC)
Sure, I can definitely do that.
-kpọ́ (base form, "call")
ị́kpọ̄ (infinitive)
prefix i- always high tone; tone of basic verb stem like this becomes mid if high, doesn't change if low
àkpọ́ (participle)
this form can combine with auxiliary verbs, the most popular of which are (marks progressive) and (marks future)
kpọ̀ọ́ (positive imperative)
ákpọ̄lá (negative imperative)
suffix is -le after -ATR and -la after -ATR.
kpọ́ọ́ (subjunctive, some verbs don't take the final vowel suffix, like -hụ́, whose subjunctive form is hụ́)
kpọ́ọ́ (sequential, same exception as above.)
Some verbs don't take a final vowel suffix for the positive imperative, subjunctive, and sequential. One such verb is -hụ́, whose positive imperative, subjunctive and sequential are hụ́. The verbs that are exceptions follow the same pattern. edit: Made an error here, the imperative isn't a mid-tone Umu igbo (talk) 09:28, 12 March 2021 (UTC)
kpọ̀ (simple present should work)
kpọ̀rọ̀ (see note below, probably just call it "simple past" or "-rV form" though)
Verb root tone becomes low, takes -rV suffix. Onicha uses -lu/-lụ/-li/-lị (-lu used after o and u; lụ after a, e, ọ, and ụ; li after i; and lị after ị).
The suffix is used with action verbs to mark what my Igbo teachers called the simple past tense. With stative verbs it's tricky; some already take the suffix in their present tense forms and thus can't use them in the past tense (like -nwé ("to have") in "Ḿ nwèrè íhé" which means "I have something"). Others readily use it (like -bụ́ in "Ḿ bụ̀rụ̀ ónyé ísī" which means "I was the boss").
Conjugation: ákpọ̀rọ̀ m̀ ("I called"), ị́ kpọ̀rọ̀ ("You called"), ọ́ kpọ̀rọ̀ ("He/she/it called"), ànyị́ kpọ̀rọ̀ ("We called"), ụ́nụ̀ kpọ̀rọ̀ ("You all called"), há kpọ̀rọ̀ ("They called").
kpọ́ghị̄ (negative, doesn't translate with a present meaning.)
Suffix is -ghi before +ATR, -ghị before -ATR. Person conjugation below. Note that NPs take the plural pronoun form too. In compound verbs, the first element of the verb stem is the one that gets affected by tone the way -kpọ́ does below.
Conjugation: àkpọ́ghị̄ ḿ ("I didn't call"), ị̀ kpọ́ghị̄ ("You didn't call"), ọ̀ kpọ́ghị̄ ("He/she/it didn't call"), ànyị́ ákpọ̄ghị́ ("We didn't call"), ụ́nụ̀ ákpọ̄ghị́ ("You all didn't call"), há ákpọ̄ghị́ ("They didn't call").
kpọ́ọ́lá (perfective)
Suffix is -le after +ATR and -la after -ATR. Takes vowel prefix when the subject is an NP or plural pronoun. The tone the vowel prefix takes depends on the tone of the first element of the verb root and the subject. When the first element of the verb root is low, my dialect takes a high prefix. Vowel prefix is e- before +ATR and a- before -ATR. The conjugation list below should show the different tones the prefix takes when the first element is high.
Conjugation: ákpọ̄ọ́lá m̄ ("I have called"), ị́ kpọ̄ọ́lá ("You have called"), ọ́ kpọ̄ọ́lá ("He/she/it has called"), ànyị́ ākpọ́ọ́lá ("We have called"), ụ́nụ̀ àkpọ́ọ́lá ("You all have called"), há ākpọ́ọ́lá ("They have called"), Tòóchī ákpọ́ọ́lá ("Tochi has called").
This is where I'd personally draw the line for inflectional forms, the rest are just basic suffixing as far as I'm aware. I only put the conjugations because I didn't know if you'd want them or not: I could just write the different forms that a particular inflection can take in the entries themselves. Umu igbo (talk) 11:05, 11 March 2021 (UTC) EDIT: Just a note, but the -la/-le suffix depending on where you are doesn't take vowel harmony. Umu igbo (talk) 09:01, 25 March 2021 (UTC) EDIT 2: When speaking, I don't apply vowel harmony to the suffix, but the grammars say that vowel harmony is applied, so idk which one you want to do.
@Smashhoof, I don't think there's any point in waiting for Rua (and I frankly don't know why she was even pinged). It looks like my first intuition was right, and we do have enough forms to fill a nicely sized {{ig-conj}}, so whenever you have time, it would be much appreciated. —Μετάknowledgediscuss/deeds 22:54, 15 March 2021 (UTC)

flood flag[edit]

Hello meta, could I get a flood flag to do some de-IPA cleanup? There are around 4000 entries to fix, replacing no longer used /ʀ/ with /ʁ/ in German entries. (See last few edits for a sample). de.wikt did a similar thing in 2018, and most entries which need changing here were presumably copied from there, before late 2018. – Jberkel 13:38, 16 March 2021 (UTC)

@Jberkel: You now have the flood flag for 24 hours. Let me know if you're done and need it removed, or if you're unable to complete the task in that time. —Μετάknowledgediscuss/deeds 15:46, 16 March 2021 (UTC)
Done, thanks. – Jberkel 19:44, 16 March 2021 (UTC)

headword-line template[edit]

Hi. You said, all entries need to have a headword-line template, which I see you've been removing. Do you mean like {{en-noun}}? What would be appropriate? kwami (talk) 05:07, 23 March 2021 (UTC)

@Kwamikagami: I provided the appropriate template ({{head|mul|numeral}}) in the edit where I pinged you. Take a look at my other formatting changes in that edit while you're at it... I see now that you've edited a bunch of entries with these formatting problems, so it would be appreciated if you could clean them up as well. Let me know if you're unclear on any of the changes. —Μετάknowledgediscuss/deeds 05:11, 23 March 2021 (UTC)

collectivization[edit]

Why did you cancel the edit? Gnosandes (talk) 19:25, 27 April 2021 (UTC)

You changed the meaning. Stop making politically biased edits or you will be blocked again. —Μετάknowledgediscuss/deeds 19:30, 27 April 2021 (UTC)
I changed the meaning to neutral. Can you quote such an edit of mine for today and explain why it is political and biased? Gnosandes (talk) 19:34, 27 April 2021 (UTC)
Read Wikipedia, it says so: The process by which farmland is aggregated is called collectivization. Gnosandes (talk) 19:38, 27 April 2021 (UTC)
You have now been blocked for your edits by Surjection. —Μετάknowledgediscuss/deeds 19:51, 27 April 2021 (UTC)

Gammaridea[edit]

Where is Gammaridea redefined? Who has accepted it? DCDuring (talk) 14:55, 30 April 2021 (UTC)

You're right, I was relying on my outdated knowledge. I guess Lowry and Myers really left it for dead, so I'll reinstate the wording. —Μετάknowledgediscuss/deeds 16:46, 30 April 2021 (UTC)

WhatLinksHere cleanup[edit]

Hi. You deleted a page that had a lot of incoming links. When you get a moment, could you sort them out please? Special:WhatLinksHere/Appendix:List_of_protologisms/third_person_singular_gender_neutral_pronouns. Equinox 01:27, 22 May 2021 (UTC)

Forgot to deal with this earlier, but thanks for letting me know. —Μετάknowledgediscuss/deeds 06:38, 13 June 2021 (UTC)
Thanks. Remind me later if I have to prove you aren't perfect. Equinox 02:07, 7 July 2021 (UTC)

Hi i may need some help![edit]

Hey, I don't know if you know how to translate proper nouns, such as names for different languages, but im trying to find my name in japanese but as such is not there, only the words ( nouns ) what can i do about this, as i need to know what my name is in japanese, thankyou for your time! EzeeWiki (talk) 03:07, 6 June 2021 (UTC)

If your name is Ezee, there isn't going to be an attested translation of that, just a transcription in katakana. —Μετάknowledgediscuss/deeds 06:38, 13 June 2021 (UTC)

Hausa pronunciation module[edit]

Hi Metaknowledge! Before programming the Hausa pronunciation module I need some clarifications on Hausa phonology and on what you wish for this module. I apologize for the length of this message and possibly silly questions, as well as for the long answer you'll have to reply.

  1. What about Ajami? Should Ajami entries be deleted ? Else the module should provide pronunciation for them too. Should the module handle this writing system ?
  2. For the representation of tones, what do you prefer : accents (àâá), numbers (like in sinological IPA, a⁵⁴a³²), IPA signs (a˥a˥˩ etc; for low tone ˨; for falling tone a lot ˥˩, ˥˨, ˥˧, ˥˦, ˦˩, ˦˨, ˦˧, ˧˩, ˧˨, ˨˩; for high tone ˦) or another thing?
  3. Along with Kano dialect (Kananci), Daura's one (Dauranci) is looked upon as standard (peculiarly Standard Kano Hausa), they are spoken in Nigeria. Both are Eastern Hausa, as well as Bauchi's (Bausanci), Katagum's (Gudduranci) and Hadejiya's (Hadejanci). Western Hausa dialects are Sokoto's (Sakkwatanci, also known as Classical Hausa), Katsina's (Katsinanci, transitional with Eastern dialects), Arewanci (Gobir's (Gobirawa), Adar's (Adarawa), Kebbi's (Kebbawa) and Zamfara (Zamfarawa)) and Kurhwayanci (in Kurfey, Niger). There are also Northern Hausa in Arewa (Arewaci) and Southern Hausa in Zazzau (Zazzaganci). Hausa is spoken in (Southern) Niger, in Ghana (Ghanaian Hausa or Gānanci, having [c] > [t͡ʃ] and [ɟ] > [d͡ʒ]) and in (Southeastern) Sudan (and Cameroon, Chad, etc., dialects under Arabic influence).
  4. What about dialects, which ones to include, and what names to display? They are numerous. If you refuse every dialect, how to know the preferred pronunciation of <j>, <ƙ>, <ts> and <ʼy>? What's more, there are dialects without tone distinction. Kurhwayanci, Damagaram and Aderawa are non-tonal dialects, having pitch accent instead.
  5. Glottal stops are written as <i> except at the beginning of a word where it's unwritten. Is that true? How to differentiate the glottal stop and the vowel ?
  6. Is initial /r/ rolled: [ɽ] > [r] \ #_?
  7. Is <f> realized as [ɸ] or [f]?
  8. According to maguzawa.dyndns.ws found in the refs of w:Boko alphabet, there are other tones in Hausa: mid and falling (from high, mid, or low, all to low). Is that true, and if yes, is it in standards or widely spoken dialects?
  9. Non-native speakers can merge implosives and ejectives with their regular counterparts, lose vowel length, and change or lose tones. There's an ejective /t͡ʃʼ/ <tsh> or <ch'>, not in standard Kano dialect. I doubt whether those facts are relevant, though one third of Hausa speakers aren't native.
  10. Before /a/, palatal, velar and labiovelar stops occur so there's no processing to do. Before front vowels /e/ and /i/, non-labial velar > palatal. Before back vowels /o/ and /u/, non-labial velar > labiovelar. This means that apparently, non-labial velars can occur only before /a/ and that some spellings would be equivalent in pronunciation: <kwo> and <ko>, or <kye> and <ke>. Furthermore, can it be /kj, gj, kʼj/ instead of truly palatal stops?
  11. Is <'y> as currently a palatalized glottal stop [ʔʲ], or a creaky-voiced palatal approximant [j̰]?
  12. There are four diphthongs: <ai>, <au>, <iu>, <ui>. Are they realized as true diphthongs (ie. one vowel who varies), or as separate vowels (with hiatus) or as a sequence vowel + glide (since the four end in <i> or <u>)? Are short /i/ and /u/, close ([i] & [u]), centralized ([ɨ] & [ʉ]), or near-close ([ɪ] & [ʊ])? Is short /a/ a near-open central [ɐ]? What is meant by neutralization of medial consonants, do they become central or mid, and if yes, to what extent and in what precise phonetic context?

I've read the introduction of Newman's (2007) A Hausa-English Dictionary (ie. until page 1 where the very dictionary starts). I've picked out these facts, according to him, and brought questions I'm wondering:

  1. Hooked <ɓ> and <ɗ> are laryngealized implosives ([ɓˤ] and [ɗˤ]?), not just implosives. Likewise, hooked <ƙ> and <ts> are glottalized ejectives ([kʼˀ/kˮ/k̰ʼ] and [(t)sʼˀ/(t)sˮ/(t̰)s̰ʼ]?), not just ejectives. The transcription I propound here assumes that both features are realized simultaneously, but it shall vary insofar as it's just on the onset (e.g. [ˀkʼ]) or it's a sequence (e.g. [ʔɓ], with glottal closure).
  2. <'y> is a laryngealized semi-vowel /jˤ/, which makes a third transcription for this digram.
  3. <j> is affricate [d͡ʒ] in Standard Hausa and Eastern Hausa, but fricative [ʒ] in Northwest Hausa. Therefore, I guess if you decide to have only a standard pronunciation, [d͡ʒ] should be displayed.
  4. /p/ exist as a consonant, because of some English loanwords, so I guess I've to add it to the module as a consonant.
  5. In some Hausa dialects /f/ is pronounced [p]. /f/ is actually surfaces as [f], not as [ɸ] as suggested above.
  6. There are also a set of palatal and one of labiovelar: <gw> [ɡʷ], <gy> [ɟ], <kw> [kʷ], <ky> [c], hooked <ƙw> [kʷʼ], and hooked <ƙy> [cʼ].
  7. <ts> is affricate [t͡sʼ], at least in Standard Kano Hausa. Therefore there are dialects wherein it's a fricative [sʼ].
  8. Glottal stop should always be included in spelling, like <ɗab'ī̀> "printing", <jāmi'ā̀> "university", <'amai> "vomit". BTW, you know that, once this module achieved, adding pronunciation won't be automatic, that is, adding barely {{ha-IPA}} on every entry, manually or bot-made, won't suffice. You will systematically need to fill one parameter with the word, including tones, vowel length, glottal stops, <r/r̃> distinction, and possibly other information omitted in the pagename.
  9. This distinction is not in Ajami, so likewise, if you wish that this module works for Ajami entries, a parameter would be required to indicate the presence or absence of a feature, if it's not better actually to put the Boko equivalent as a respelling for the template.
  10. <r> is a retroflex flap /ɽ/, whereas Hausa entries with a pronunciation currently show <r> as a simple flap /ɾ/. Initial <r> isn't necessarily rolled, like <rānā> "sun", starting with the flap. Also, the <r/r̃> distinction (or better, the [ɽ/r] one) doesn't exist in every dialect.
  11. Geminate consonants exist in Hausa. Is it systematic, that is, do [C¹C¹] > [C¹ː] necessarily happen?
  12. All tones can apparently exist on short and long vowels, but macron isn't written for a the falling tone on a long vowel, so there are: simple vowels (no diacritics); with grave; with circumflex; with macron; and with macron plus grave. Effectively the falling tone is less common, but, well, on second thought, how it is possible to know whether a vowel is long or short with the falling tone? Vowels with falling tone in open syllables become long (without coda, ie. in syllable ending: [V̂] > [V̂ː] \ _CV, _#). Is it better to indicate it manually or to let the module generate this, if it's easily predictable without exception?
  13. In Northwest dialects, the final consonant geminates after the plural suffix -ai: [C] > [Cː] \ _ái#. There's an assimilation, [r] > [d] \ [_ d], which is the sole given by Newman. In total, there are only two rules applied at the scale of the language, and 9-10 across dialects (which is fairly reasonable, given the number of dialects). I'm glad there are very few sound laws involved. This fact can seem negligible but it will keep the module rather simple, compact. Malku H₂n̥rés (talk) 13:24, 13 June 2021 (UTC)

Okay, thank you for taking this on. An important note: I am not too fond of the idea of making a phonetic transcription; I would prefer just to do phonemic, perhaps with a little more detail than usual. The reason is that vowel quality does change with length and position, but I know I speak with an accent, and I don't think I could reliably represent what the vowels should exactly be in IPA. Anyway, starting with your first section:

  1. The module shouldn't have to handle ajami; those entries should be soft redirects to Latin script anyway.
  2. I use accents.
  3. At least to start, with I don't want to bother with the (often poorly described) dialects. We should just display (Standard Hausa) and base it on the speech of Kano.
  4. See above.
  5. No, glottal stops are written with <'> in medial position. You are correct that all seemingly vowel-initial words do have an initial glottal stop (and whether it is phonemic could be disputed, but I would advocate including it in the phonemic transcription).
  6. No, initial r can be contrastively rolled or not. There are no native proper nouns with initial <r>, so initial <R> should be interpreted as /r/. (That is, we do not write <R̃> on Wiktionary). However, there should be a way to override this parameter; a couple very rare exceptions exist.
  7. The choice is somewhat arbitrary in a phonemic transcription. On Wiktionary, we tend to depart from linguistics norms to opt for the more accurate sound, so /ɸ/.
  8. Standard Hausa only has high, low, and falling.
  9. Non-native speech is always unstable anyway; I don't think we have any need to cover it.
  10. That's right. They are truly palatal.
  11. I would notate it with the latter for modern speech. I'm no phonetician, though.
  12. I was taught that there are only two diphthongs, /ai/ and /au/. In any case, shouldn't be a problem for phonemic transcription.

And now your second section:

  1. Kind of an overcomplication; not relevant to phonemic transcription.
  2. As above.
  3. Yeah, this does test my commitment to only showing Standard Hausa, because we have a bunch of entries that are Nigerien Hausa, and of course they simply aren't used in Standard Hausa at all, only in the Northwest dialects. But the problem is that I have not spoken with Nigeriens, and I don't know the other features of Northwest Hausa that we'd have to code!
  4. Yes, although even if you didn't add it, any unfamiliar symbol would simply be copied in the IPA, so it would have worked out.
  5. No, it's [ɸ] in Standard Hausa.
  6. All those are possible before /a/.
  7. I would say it varies idiolectally, even in Kano. For a phonemic transcription, I would use /sʼ/ or maybe /(t)sʼ/.
  8. No, you misunderstood — when a word is seemingly vowel-initial, the module should infer an initial glottal stop. It will not be written. (And yes, obviously, but a bot can add it based on the headword line).
  9. Yes, as I said above.
  10. You're right, /ɽ/ is a preferable transcription, and we should be using that instead.
  11. I would treat them as two consonants rather than a single geminated one, as that's what Hausa phonotactics suggest.
  12. A vowel with falling tone is always long unless there's a coda, so it's predictable.
  13. Those changes are represented at the orthographic level. Northwest Hausa is not usually written, though.

Let me know when you want testcases. —Μετάknowledgediscuss/deeds 19:56, 13 June 2021 (UTC)

In any case, one phonemic transcription in mandatory to start. One or several phonetic transcription would be really helpful to know how it is actually realized, and they would be drawn from the phonemic one, whence its necessity. That is, if the vowel position is affected by its phonetic context, I believe it's better to have a phonetic transcription indicating it so that the user can pronounce the word as native speakers would do. As for their exact pronunciation, it may be possible to get a native speaker (of Standard Kano Hausa preferably) to record files. So here I assumed only the phonemic and the standard phonetic transcriptions are displayed by the module, but keeping the possibility of adding dialects later. The order is different, having mixed the answers of both sections (Newman and non-Newman), thus it's by topic.

  1. <ɓ> /ɓ/ [ɓ], <ɗ> /ɗ/ [ɗ], <ƙ> /kʼ/ [kʼ], <ts> /(t)sʼ/ [(t)sʼ], <'y> /j̰/ [j̰], <j> /(d)ʒ/ [d͡ʒ], <f> /f/ [ɸ], <Ø-/'> /ʔ/ [ʔ], <r> /ɽ/ [ɽ] is convenient to you (with the phonetic transcription being the one of the standard, that's why phonemic and phonetic transcription are mostly lookalikes here).
    1. /(d)ʒ/ is more parsimonious, because the phonemic pronunciation stands for all dialects, and the phonetic ones respect the dialects, here the standard, so [d͡ʒ] is displayed for this one. /f/ is enough for a phonemic transcription (no contrast [f]/[ɸ] exists for obvious reasons). However I have no knowledge of earlier stages of Hausa, so as to know whether [d͡ʒ] and [ɸ] are with original sound or dialectal innovations.
    2. Glottal stops are written as <'> except at the beginning of a word where it's unwritten, implicit. Therefore they should always be included in spelling when not initial, for the module can't predict it. You wonder whether or not it's phonemic. It depends on the phonotactics: if the syllable structure requires an initial consonant (and apparently it does), so yes it is phonemic. If it's impossible to imagine a syllable beginning with a syllable then it's phonetic since it's just to help pronouncing. Neither in Wikipedia nor in Newman I've found the syllable structure, but from what I've seen it seems it is CV(V)(C), with the second vowel being for the diphthongs (because the patterns in the module will read two vowels even if it's one phonemically). Another way is to know whether all (or most) dialects have this systematic glottal stop to prevent vowels from getting an initial position, and if yes, it's phonemic.
      1. You haven't heard of <iu> and <ui>. The aim of the question was to get firstly the vowel structure in a syllable, and secondly a narrow phonetic transcription for the standard according as they are true diphthongs or not: [ai]&[au], [a.i]&[a.u], [ai̯]&[au̯], [aj]&[aw], [aɪ]&[aʊ], [aɪ̯]&[aʊ̯], [a͡i]&[a͡u], [a͜i]&[a͜u], and so on. You haven't replied to the rest of this question actually. We aren't compelled to have the narrowest transcription, we can satisfy with something a bit more simple and the possibility to improve it later.
    3. <'y> would be rather a creaky-voiced palatal approximant [j̰] than a palatalized glottal stop [ʔʲ]. Given the orthography, the former is more intuitive with <ƴ> (a yod, given <y>, with a different voice, given the hook) and the latter with <'y> (a glottal stop <'>, with palatalization <-y> like on the digraphs <ky>, <ƙy> and <gy>). For phonemic transcription it's just a sign after all, so it doesn't matter that much and I align on your choice, but for phonetic purpose, both seem illogical: a glottal stop undergoing a second (simultaneous) articulation seems strange to me, and having creaky voice for a single phone too (unless it stands for the same mechanism as ejectives and implosives so that the system is phonemically coherent).
    4. <ts> is phonetically affricate [t͡sʼ] in Standard Kano Hausa according to Newman, however you would prefer rather the fricative /sʼ/ for the phonemic one. On the contrary, I would prefer /(t)sʼ/ for the phonemic one since it can stand for at a time [t͡sʼ] and [sʼ]. But if the actual realization varies inside Kananci, what should be displayed for the phonetic transcription of the Standard Hausa? [(t)sʼ]?
    5. Not showing dialects, the fact that /d(ʒ)/ is [ʒ] in some is likewise inconsistent. One question though: isn't it contradictory with having dialects labels for glosses? I mean, having an entry with (Niger) and displaying the Kano pronunciation... The best would be to display only the phonemic transcription as something neutral, and later its actual one if (Northwest Hausa) or (Niger) or the like is added to the module.
  2. I admit I didn't clearly understand what Newman meant with "laryngealized implosives" and "glottalized ejectives" because it appears as redundant and overladen. Without mentioning the second articulation, the question of their link (simultaneity, onset, sequence) becomes inconsistent.
  3. Accents are by far the best representation for tones in a phonemic transcription but it can be imprecise in a phonetic one. But using numbers or the IPA signs may be too narrow for the current purpose, thus I agree for accents even in the phonetic transcription of the standard. Likewise, for now we forget the "toneless" dialects, assuming it's useless (the user just need to remove accents). Standard Hausa has 3 tones, and for the moment for the moment we are also ignoring dialects with extra tones.
  4. Initial <r> and <r̃>, can exist, representing their habitual pronunciation, but <R̃> is replaced by <R>, still representing /r/. The precise function of the supplementary parameter you wish is to indicate that <R> is indeed the regular /ɽ/.
  5. Non-native speakers are ignored.
  6. The ejective /t͡ʃʼ/ <tsh> or <ch'>, not in Kananci, is also ignored.
  7. Non-labial velars can occur only before /a/, otherwise it's assimilated to labiovelars or palatal, which are really palatal.
  8. The question of /p/ as a consonant was mostly to add it to the module as a consonant so that if a sound law involving any consonant apply, /p/ would count. On Module:tzm-IPA it's essential, here it would be limited to knowing is a syllable is open or close (if the next characters match a string with every consonant) for falling tone vowel length. It was also subtly meant to ask whether /p/ is actually [p], and because you haven't said anything about that, I guess it's true.
  9. Geminate consonants exist in Hausa, but you prefer to transcribe them as in Italian, that is /VC¹.C¹V/, where the geminated consonant spans on either side of the syllable break (unless you know a counterexample where both parts are in the same syllable).
  10. The length of vowels with falling tone is predictable so the module will handle this.
  11. Ajami entries are soft redirects, and the pronunciation being on the Boko entry, there's no need to put it on the Ajami entry and therefore the module won't handle this script.
  12. I agree to display Standard Hausa. Is it peculiarly Kananci, or does it apply to Dauranci too? For dialects, maybe it's too narrow to name them after one city without knowing their precise phonology. However, don't you think it's possible to rather show tendencies, that is, to display "(Southern) Niger", "Nigeria", "Ghana" or "Eastern/Western/Northern/Southern Hausa", so that it's enough broad to be relevant, at a time displaying major changes in pronunciation and making possible to find common features? I just ask which ones could be added in later updates of the module.
  13. Two additions. First, Hausa has an official alphabet in Niger having <fy> as a digraph: what's its value, assuming it would be revelant if the module displayed this dialect? Do we have entries with it currently or can we ignore this for now? Second, I guess that /a/ can surface as a central [ä], since it's a "symmetric" 5 vowel system, it varies with central [ɐ] and non-labial velars don't assimilate before it.

In order to start programming, I just need that you answer precisely or agree to the facts. Thank you in advance! Malku H₂n̥rés (talk) 12:10, 22 June 2021 (UTC)

  1. Agreed.
    1. Agreed. Newman and Schuh have some work on Hausa' phonological history, but I haven't read it either.
    2. They are always written in non-initial position, don't worry. I guess I'm still not on board with a narrow transcription yet — maybe I should work on getting audio recordings first.
      1. I checked, and Newman confirms that only <ai> and <au> are diphthongs in Standard Hausa, as I thought, with */ui/ and */iu/ having monophthongised. I think that /ai/ and /au/ are good for broad transcription (without any diacriticking), but some care would have to be given a narrow transcription for /ai/, which can surface differently in different environments.
    3. Don't judge things based on orthography. It may seem illogical, but I do think it usually comes out as creaky, and it doesn't really align with the rest of the system anyway. Language doesn't always work the way you think it ought to!
    4. Yes, I think /(t)sʼ/ and [(t)sʼ] are preferable on further reflection.
    5. You mean /(d)ʒ/, of course. Yes, a Niger parameter seems like a good idea.
  2. OK.
  3. Agreed.
  4. Agreed.
  5. Agreed.
  6. No, we should include it. It can't hurt, after all.
  7. Agreed.
  8. Yes, it's [p]. (And in some dialects, so is /f/.)
  9. Exactly. On that topic, don't forget that geminated <sh> is written as <ssh>, geminated <ts> as <tts>, etc for all digraphs.
  10. Agreed.
  11. Agreed.
  12. I'm very hesitant to add those updates, as I don't have enough knowledge about them. The Standard is specifically based on Kano.
  13. Yes, it's palatalised and it can turn up in Standard Hausa. Sure, but I have to admit that I'm not confident about assigning narrow IPA values to the short vowels.

Μετάknowledgediscuss/deeds 19:19, 22 June 2021 (UTC)

As I'm programming the module, two last questions come to me:

  1. is it <fy> /fʲ/ [fʲ], as you said it's palatalized?
  2. <iu> and <ui> are monophthongized towards what? What's their phonetic transcription? I checked in the 1,195 Hausa lemmas, and you are totally right, there are numerous occurrences of <ai> and <au> but not a single of <ui> nor <iu>. Should I rather ignore them?

Malku H₂n̥rés (talk) 15:56, 29 June 2021 (UTC)

  1. @Malku H₂n̥rés: Yes. As for */iu/ and */ui/, you must've missed my asterisks. They monophthongised in the past, hence not existing in Hausa any longer. —Μετάknowledgediscuss/deeds 22:46, 29 June 2021 (UTC)

Hi Metaknowledge! I've just finished the module (Module:ha-IPA) and created the infrastructure for testcases (Module:ha-IPA/testcases).

  1. Please provide testcases so that every feature is tested (for some reason the text is not shown, but it's enough to see the expected and actual pronunciations).
  2. For the pronunciation of initial <R>, the solution I've found is to use lowercase <r> for /ɽ/ instead of uppercase <R>, which gives /r/. It's easier if you just put in lowercase for the rare exceptions than adding a parameter for that.
  3. There's a parameter for Niger pronunciation working like that |niger=1, which currently just removes the Standard Kano Hausa pronunciation. You can put several spellings in order to generate several pronunciations if needed.
  4. For the diphthongs, the tone is written only on the <a> (<ai, ài, âi, au, àu, âu>), and there's never macron. How is tone distributed in the diphthongs? Does it spread from /a/ to /i/ and /u/, or should /i/ and /u/ be toneless? Are those diphthong always short, even with a falling tone?
  5. Does <tsh> geminate as <ttsh>, <ch'> as <cch'>, <'y> as <''y>? I'm not sure, I just followed your rule.
  6. What about affricates /(d)ʒ, t͡ʃ, (t)s/ when geminated (ejictivity not shown here)? Is it simply /(d)ʒ.(d)ʒ, t͡ʃ.t͡ʃ, (t)s.(t)s/ (as currently by default), or /d.d͡ʒ, t.t͡ʃ, t.t͡s/ or /s.s/, or something else? In other words, does gemination turn /(d)ʒ, (t)s/ into true affricates, and is the stop the sole part before syllable break with the entire affricate after? And eventually, does it happen (if it does) at the phonemic scale or only in Standard Kano Hausa?
  7. I've found a chapter about Hausa in the handbook of the IPA: [1]. Should I use it?

Thanks in advance for your answers. Malku H₂n̥rés (talk) 16:48, 8 July 2021 (UTC)

Great work! 1. Haven't got time for testcases right now, but why don't you add a bunch and then I'll loom over them and see what's missing. 2. That's a good solution. 3. That works for now. 4. The tone is distributed, but we don't have to worry about that — just follow the orthography. 5. Those are either nonexistent or never geminated in Standard Hausa, so I have never seen them. 6. This is a bit tough. On a phonetic level, I think it's just the stop that gets doubled, but on a phonemic level, Hausa doesn't care (e.g. dialects that have [ʒ] and those with [dʒ] will treat it identically in terms of morphophonology). That would lead me to the (somewhat ugly) current default. 7. Yes, those are reliable authors. —Μετάknowledgediscuss/deeds 20:20, 8 July 2021 (UTC)

Reconstruction:Proto-Bantu/ngàdí[edit]

i believe the Luganda word for blood fits well as a descendant from proto bantu Wojak6 (talk) 06:50, 17 June 2021 (UTC)

@Wojak6: What word? The word I'm seeing is omusaayi, which is unrelated. —Μετάknowledgediscuss/deeds 17:06, 17 June 2021 (UTC)

Regarding comprehensive etymologies[edit]

Good evening! :^)

You reverted one of my edits on acappellakor (which removed a lot more improvements than just the etymology by the way..) because my etymology was too correct .. I mean too long! With the reasoning that it did not improve anything, and that the full etymology could be found by clicking further on each word in the etymology, I would argue this is false for two reasons:

1. If you want to get to the "end" of a word (aka back to the first root-word), you'd most likely have to click to a ton of pages with simple etymologies, only showing the one preceding word. I don't see how it is useful to have to click through or open 15 new tabs only to get to the source of a word, when you could have all the information on one page, which really does not take up much space.
2. In this case, it does not even work, if you click on "a cappella" in the etymology, it leads you to Norwegian, where you can click to get to the Italian entry, here there is no etymology, and the journey ends. In my long etymology, it went all the way back to Proto-Indo-European, and you could see the full journey from the root word to the modern Norwegian Bokmål term, this is now lost.

Therefore, I am asking pretty please, if I could create another long etymology (which I will try to simplify and shorten down, as well as add some spaces to make it much cleaner than it was, but still correctly categorizing the entry down to Proto-Indo-European, as the root is now missing), what do you think about this? Thanks! Supevan (talk) 17:27, 19 June 2021 (UTC)

As I said in my edit summary: This kind of over-the-top etymology does not improve entries. Instead, it makes them messier and liable to fall out of sync. If readers want to know the etymology of the component forms, they can click through. Remember, the word itself does not go back to PIE! Its components were just componded together. This being the English Wiktionary, we tend to give extra lenience to English entries, but in general, we should give the minimum necessary, and trust that our readers know how to click a link. —Μετάknowledgediscuss/deeds 23:25, 19 June 2021 (UTC)
Right, agree to disagree then. If that is the case, it would make no sense to have categories such as the Proto-Indo-European roots, I mean there are no Norwegian words which stem directly from PIE without any words/languages in between, so in theory they would all have to be deleted (if we only added the one preceding word in the etymology). Only Italic, Ancient Greek and a handful of other languages would actually use these PIE root categories, but plenty of languages are using them now. There were a bunch of Norwegian words already using the PIE-root categories before I started editing, and it makes little sense to only categorise some random words, which is why I am creating these etymologies to categorise all of them.
We could be adding just the straight up categories instead, but then you'd have to map out the full etymology anyway, so might as well put it there, because if you want to double-check the categories you would have to follow the trail again every single time through a bunch of other entries (often finding entries with dead ends / empty etymologies!), instead of just having everything on one page where it is all automatically categorised.
I will leave one more example before I give up, look at the entry I created for Norwegian accessoir, imagine if this entry was "improved", and it only said "From French accessoire.", click on that word and you're already at a dead end, is this really an improvement? Supevan (talk) 06:21, 20 June 2021 (UTC)
I don't really care about the root categories one way or the other. But yes, your etymology at accessoir was a mess that didn't help the entry. I mean, you even detailed the origin back to PIE of a suffix that doesn't even appear in the French word that the Norwegian is borrowed from! Maximalism isn't a virtue. You've got skills and knowledge, use them to create more entries instead of copy-pasting etymologies. —Μετάknowledgediscuss/deeds 06:55, 20 June 2021 (UTC)

small oops[edit]

Right message, wrong heading. Good catch, thanks. I'm reposting with a correction. There's no harassment involved, just a word to the wise. Please report me if you think I'm wrong and let Hazarasp delete the post if he sees fit. Cheers. --Kent Dominic (talk) 17:05, 3 July 2021 (UTC)

I'm not an admin. I also don't know whether deleting a post from another user's talk page is a blockable offense. If I were an admin. and someone asked me to block you for doing just that, I'd probably let you off with a stern warning against taking it upon yourself to disrupt what might or not be disruptive or harassing behavior in someone else's house. If some user posts what you consider a prolix load of disruptive harassment onto my talk page, kindly leave it there for me to deal with. Cheers. --Kent Dominic (talk) 11:13, 4 July 2021 (UTC)

The only potentially blockable offence I see is your behaviour. —Μετάknowledgediscuss/deeds 19:33, 4 July 2021 (UTC)
I need your advice. What should I do about an editor who made an unannotated, seemingly spiteful edit reversion that was contrary to an RFD consensus and disruptive of a long overdue compromise reached by the main principals of a dispute concerning the wording of a particular sense? I don't want to cause trouble for the editor. I'd just like him warned to lay off. If you don't reply, I'll just drop it entirely. I've got better things to do than to hold grudges. Cheers. --Kent Dominic (talk) 22:02, 4 July 2021 (UTC)
I could complain about the rather tilted description of events you've provided here (for one, said editor is of the "principals" in said dispute, so a compromise cannot exclude him). But let me just say this, Kent: Let it go, as I've done. Hazarasp (parlement · werkis) 23:30, 4 July 2021 (UTC)
Read between the lines, Hazarasp. Said editor wasn't you (i.e. all of your reversions were annotated), and said editor was an interloper who contributed no comments whatsoever to the RFD thread nor to the other talk pages regarding the mentioned sense. Follow Metaknowledge's cue in seeing my request for advice in the light it was intended: a rhetorically cynical footnote about hypocrisy, not a bona fide request concerning a one-off tantrum with no effect beyond a momentary disruption that really didn't faze anyone, least of all me. I'm pretty sure the editor himself or herself knew, either in hindsight or immediately or upon making the reversion, that it was an impulsive lack of discretion on the part of someone who's never once otherwise engaged in divisive behavior. --Kent Dominic (talk) 06:42, 5 July 2021 (UTC)
Just got a bit paranoid and neglected to use my better judg(e)ment given the events of the last few days. Hazarasp (parlement · werkis) 07:09, 5 July 2021 (UTC)

You have to supply the source word for a borrowing[edit]

(At kofia): thanks for the tip. I will try to remember this. The problem is that if I don't know the source language I can't always be sure of the exact spelling. For example we might borrow a word from Maori or Hawaiian but drop the macron (or whatever they call that accent). If I write the word per se in the template, it will satisfy the template, but I will think "I've just said something about Maori or Hawaiian that I'm not sure about". So of course I have the option of adding no etymology at all. But if I know the source language... what is the solution, Meta? I stay awake nights crying about this. Equinox 02:05, 7 July 2021 (UTC)

Bad excuse. The Swahili entry was right there on the same page! You can ping me for all the other stuff; unlike you, I actually check notifications. —Μετάknowledgediscuss/deeds 03:01, 7 July 2021 (UTC)
Why you so cruel Meta :( :( emoji Equinox 03:24, 7 July 2021 (UTC)
Also, if I spell it the same way, AND it's on the same page, I know from experience you will come and spank my arse for not getting the long accent right in Latin or something. I can't win. OK in future I will copy the word and we'll see if you can leave me alone or not. Equinox 03:25, 7 July 2021 (UTC)
No, really! Just ping me and I'll check it. —Μετάknowledgediscuss/deeds 03:44, 7 July 2021 (UTC)
Somebody teach me to ping so I can subsequently never do it! — [ זכריה קהת ] Zack. — 21:34, 8 July 2021 (UTC)

Chadic subclassification[edit]

Here languages codes are sorted alphabetically. It's a tree to fill. If it bothers you, I can remove the {{desc}} template to leave only the languages codes, and/or put them in line rather than in a list.

I can keep on helping for the subclassification itself. For sources I know Newman (1977) Chadic Classification and Reconstruction [2] and possibly Numeral Systems of the World's Languages [3]. Malku H₂n̥rés (talk) 08:19, 12 July 2021 (UTC)

@Malku H₂n̥rés: I specifically asked for alphabetisation by language code because the point of this was to make it easier for me to roll out the changes! Is this complete? It doesn't seem like that many codes. Also, is it okay if we autocollapse the list so others don't have to scroll through it? —Μετάknowledgediscuss/deeds 17:01, 12 July 2021 (UTC)

Is it good now? It's collapsible, language codes are sorted alphabetically, and I added the 5 missing codes, now it's really complete. Malku H₂n̥rés (talk) 09:00, 13 July 2021 (UTC)

@Malku H₂n̥rés: You still haven't quite got the idea (it's organised by branch, when I wanted it sorted alphabetically overall to make it easier to edit the data modules, which are organised alphabetically), but I suppose this isn't that much worse. I'm going to change the code for Central Chadic from cdc-cbm to cdc-cen first, so we can reflect the name we actually use. —Μετάknowledgediscuss/deeds 16:34, 13 July 2021 (UTC)

Sub-standard Attestation[edit]

Hey Metaknowledge: I have been able to find Wiktionary-grade durably archived attestation (meeting the other requirements too) for many English language proper noun placenames and their alternative forms and synonyms. However, there are some words that I haven't gotten to first base with despite repeated searches, like Damiku or Pangudang. Another example is the partially-attested Darya Boyi, where the entry has only two durably archived cites (seemingly). I want to start putting seemingly unattestable entries into rfv, and any of their less than three attestations onto the relevant 'Citations' pages for future editors to use if these words ever become more common. That is to say, I want to move from a more inclusionist mindset to a more deletionist mindset, for the purpose of bringing everything I am looking at up to Wiktionary standards. Let me know if you have any thoughts or suggestions. --Geographyinitiative (talk) 12:03, 12 July 2021 (UTC)

@Geographyinitiative: Inclusionism is perfectly fine, as long as you follow the rules. Your recent habit of sending obviously attested words to RFV is just as much a waste of others' time as the opposite extreme. It looks to me like Darya Boyi passes, no? —Μετάknowledgediscuss/deeds 17:03, 12 July 2021 (UTC)

I don't think so. There are only two durably archived cites and hence it fails Wiktionary:ATTEST Prem Rec Media requirement unless there's another source. Geographyinitiative (talk) 18:23, 12 July 2021 (UTC)

@Geographyinitiative: It would be helpful to remove citations that are not durably attested (or move them to the Citations tab under a header explicitly saying they are not durable). —Μετάknowledgediscuss/deeds 19:03, 12 July 2021 (UTC)
Hey, if I remember correctly you are at least tangentially interested in dinosaur names related to Chinese characters- here's a children's book that uses a lot of scientific names that look like they come from Chinese characters- at least forty--[4] --Geographyinitiative (talk) 20:39, 21 July 2021 (UTC)
I definitely don't need a children's book for that! I'm interested in taxonomic etymology in general, but there's a tremendous backlog around here. —Μετάknowledgediscuss/deeds 22:18, 21 July 2021 (UTC)

Moroccan Amazigh audio files[edit]

Referring to our recent talk (see additions for practical use of files), I propose the following points to be discussed somewhere (here ? on Meta ? on Commons ?).

To differentiate orthography and pronunciation, I propose:

  1. All the Moroccan Amazigh odio files should be labelled with Zgh-, the only official Tifinagh-based orthography, which can be authenticated in published IRCAM documents.
  2. Pronunciation would also refer to a sub-language (shi, tzm...), the one of the speaker, by using a second language code (Zgh-shi, Zgh-tzm). Mixed labels (Zgh-shi-tzm) should be avoided.
  3. Pronunciation characterizing transition dialects should only be labelled with Zgh- (without additional code), if that pronunciation suits with the Standard orthography (main caracters of Moroccan Amazigh). But if a speaker of a transitional dialect pronounces just like "standard" pronunciation of a codified sub-language, it can be labelled with the sub-language code.
  4. If there is a noticeable difference (between pronunciation and phonetic value of letters in standard orthography), and that the recorded pronunciation can be noted with one of the additional characters of Moroccan Amazigh, language code should be Zgh-al (Standard Moroccan Amazigh, alternative) (like here.

Sincerely yours.

--Lucyin (talk) 10:59, 25 July 2021 (UTC)

I'm not sure I see the point of any of these reforms except for your 2nd point (that mixed labels should be avoided). Why does it matter than ZGH is the only official orthography? These are audio files, which have nothing to do with written norms, and ZGH is not a real spoken language (which is why we are working to eliminate that code from English Wiktionary). —Μετάknowledgediscuss/deeds 22:27, 25 July 2021 (UTC)

遣冊[edit]

Current practice is we use {{n-g}} when the definition is an explanation of the sense rather than a direct translation. Does that make sense? ---> Tooironic (talk) 11:36, 27 July 2021 (UTC)

@Tooironic: Current practice is actually that we use {{n-g}} when the definition is not a gloss. My usual rule of thumb is that if you could substitute the phrase we give for the word itself when translating, then it's a gloss. —Μετάknowledgediscuss/deeds 15:58, 27 July 2021 (UTC)

Toki Pona "n"[edit]

You removed the article Appendix:Toki Pona/n with the reason "not official, not attested". However, as of July 21th, 2021 Sonja Lang released a new book called Toki Pona Dictionary where she emphasizes 17 non-pu words as "essential" or "core" vocabulary, because these words are used frequently enough by the Toki Pona community according to her research. As of a week ago these new "nimi ku suli" are considered official or semi-official by many Toki Pona speakers. "n" is one of these words, so I believe it should be included in the Toki Pona appendix. (In case you want an overview, you can find a list of these 17 words at this site, and I also added them to Appendix:Toki Pona#Word list.) —Rajzin 02:32, 28 July 2021 (UTC)

@Rajzin: I wasn't aware of this change. However, given that it only happened a week ago, I would tend to doubt that those words meet our criteria for attestation, which require three independent uses spanning over a year in durably archived media (like books). —Μετάknowledgediscuss/deeds 02:34, 28 July 2021 (UTC)