Wiktionary:Beer parlour/2015/January

discussion rooms: Tea room • Etym. scr. • Info desk • Beer parlour • Grease pit ← December 2014 · January 2015 · February 2015 → · (current)

---

Wiktionary:Translation requests

I think, with 38½ sections per month on average and little interest in archiving, this page is a good candidate for conversion to the monthly pages system. If we still want to keep it, anyway... — Keφr 22:33, 1 January 2015 (UTC)

Or it could be purged more often. — Ungoliant ^(falai) 22:36, 1 January 2015 (UTC)

Yes, but who wants to do it? — Keφr 22:48, 1 January 2015 (UTC)

Nevermind. I thought it had the same “archiving” system of the feedback page. — Ungoliant ^(falai) 23:25, 1 January 2015 (UTC)

I prefer the purging method, but the auto-archiving method is easier! Renard Migrant (talk) 13:42, 2 January 2015 (UTC)

Maybe have some new words of the day this year

Does anyone fancy updating the word of the day a bit more? I noticed we had a lot of repeats last year and already this year. I'd be grateful to anyone who put any work into it. Renard Migrant (talk) 13:08, 2 January 2015 (UTC)

I’ve briefly considered it before, but I’ve worried that, if I were to accidentally feature a word that was already Word‐of‐the‐Day elsewhere, it could spell serious trouble for the project. --Romanophile (talk) 13:11, 2 January 2015 (UTC)

To blow out of proportion, Template:pos verb: To overreact to or overstate; to treat too seriously or be overly concerned with. — Keφr 13:27, 2 January 2015 (UTC)

Quite an understatement. I do not think we had any new word of the day at all this year. And I think you can guess at least one reason why if you look up how these are set up compared to FWOTDs. — Keφr 13:27, 2 January 2015 (UTC)

From above "if I were to accidentally feature a word that was already Word‐of‐the‐Day elsewhere, it could spell serious trouble for the project." Seriously? Is this a joke? Renard Migrant (talk) 13:28, 2 January 2015 (UTC)

Couldn’t somebody sue the Wikimedia foundation if it looked like we copied their Word‐of‐the‐Day? See also: ‘Avoid other WOTDs - We want to feature words that haven't been WOTDs for other dictionaries, partly to highlight unique terms that make Wiktionary so special, partly to avoid complaints (by preventing the possibility entirely) that WOTDs were "stolen" from other dictionaries.’ --Romanophile (talk) 13:37, 2 January 2015 (UTC)

I'm not an intellectual property lawyer, but no I don't think so! If we copied the definitions or copied large strings of words of the day from another websites, then yes. But nobody can copyright the use of a word on its own. @BD2412 would be the person to ask (he actually is an intellectual property lawyer) but I can't imagine I'm wrong on this. Renard Migrant (talk) 13:41, 2 January 2015 (UTC)

You're not wrong on this. Cheers! bd2412 T 14:29, 2 January 2015 (UTC)

Certainly one reason we haven't any new WOTDs yet this year is that this year is only two days old. —Aɴɢʀ (talk) 14:02, 2 January 2015 (UTC)

While not ideal, perhaps it would be better to cycle through our (non-offensive) entries one by one, regardless of what the words are, than to repeat a few hand-picked ones forever. Equinox ◑ 14:08, 2 January 2015 (UTC)

How about extrabellum? Cheers again! bd2412 T 14:37, 2 January 2015 (UTC)

I don't like the idea of cycling through all our entries regardless of quality. At WT:FWOTD there is the requirement that a candidate word must have pronunciation information and at least one citation; I think that English WOTDs ought to be held to at least that standard if not a higher one. —Aɴɢʀ (talk) 15:17, 2 January 2015 (UTC)

I was just going to propose something similar. — Keφr 15:30, 2 January 2015 (UTC)

How hard can it be to find words that meet the criteria? Or, better yet, to take words that are close to that and bring them up to par? bd2412 T 15:39, 2 January 2015 (UTC)

That depends on the word, but usually relatively easy, methinks. But I guess it has to be blessed as policy somehow. Not necessarily a full vote. — Keφr 15:52, 2 January 2015 (UTC)

The problem is not the lack of words to feature (the nominations page has enough interesting words to last for months), it’s the lack of someone to update the templates. — Ungoliant ^(falai) 17:20, 2 January 2015 (UTC)

Deprecating "Acronym", "Initialism" and "Abbreviation" headers

Previous discussions: Template talk:abbreviation-old#RFD discussion, Wiktionary:Beer parlour/2014/July#Template:abbreviation-old

Some time ago it was proposed that these "part-of-speech" headers be deprecated (in favour of real part-of-speech headers like noun, adjective, interjection, etc.). I support the idea, but to be honest, I cannot recall a wider discussion about it. Meanwhile people are still adding new "initialism" entries, and there is nothing to point them to about it. If we formally deprecate those headers, we should remove them from WT:NEC, deprecate and track {{head|...|abbreviation/acronym/initialism}}, and probably set up an edit filter. Do we do it? — Keφr 18:10, 2 January 2015 (UTC)

I get the impression, of the people who've expressed an opinion, that everyone's broadly in favour of it (i.e. deprecating), but implementing it would be horrendously difficult. Because they all need to be sorted by hand, it can't be automated. Users are still adding them in good faith and it's better to have them under an initialism header than not at all. Renard Migrant (talk) 18:13, 2 January 2015 (UTC)

Like I said, we could set up an edit filter which would catch that and show an explanatory message. — Keφr 18:16, 2 January 2015 (UTC)

Sorry I missed that bit. Support. Renard Migrant (talk) 19:11, 2 January 2015 (UTC)

Support. — Ungoliant ^(falai) 19:05, 2 January 2015 (UTC)

I support it in principle, but I've found that there are some cases where it's not easy to define it with any other part of speech. I can't think of any example right now but I know there are some. —CodeCat 19:19, 2 January 2015 (UTC)

@CodeCat: I don't know if this is what you meant, but there are various kinds of abbreviations of phrases for which Phrase doesn't seem like the right header, eg, YOLO. DCDuring TALK 19:41, 12 January 2015 (UTC)

Yes, that would be a good example. —CodeCat 19:53, 12 January 2015 (UTC)

YOLO seems somewhat like an interjection to me. And the more I think about IANAL, the more it seems like an interjection too. (Compare QED.) — Keφr 20:56, 12 January 2015 (UTC)

@Kephir: Just because we have numerous erroneous uses of Interjection as PoS header doesn't mean that we should err yet again. Consider this definition of interjection from MW Onlne:

"an ejaculatory utterance usually lacking grammatical connection: as

a : a word or phrase used in exclamation (as Heavens! Dear me!)

b : a cry or inarticulate utterance (as Alas! ouch! phooey! ugh!) expressing an emotion"

Their definition of ejaculatory refers to ejaculation "something ejaculated; especially a short sudden emotional utterance"

Their other definitions are even less applicable.

YOLO, QED, and IANAL represent full sentences without marked emotional content. DCDuring TALK 21:06, 12 January 2015 (UTC)

What is the term for a phrase that has a finite verb as its head? —CodeCat 21:07, 12 January 2015 (UTC)

@CodeCat: Do you mean an absolute? That also strikes me as too technical for a PoS header, even if it were in WT:ELE (and for any definiens, except possibly some more technical style/grammar/linguistics terms). DCDuring TALK 22:39, 12 January 2015 (UTC)

No, that's not it. It would have to include the terms that the abbreviations above stand for. All I can think of is "sentence", but they're not necessarily always used as stand-alone sentences (they can be, though). —CodeCat 22:53, 12 January 2015 (UTC)

Wikipedia says at w:Finite verb: "A finite verb is a form of a verb that has a subject (expressed or implied) and can function as the root of an independent clause; an independent clause can, in turn, stand alone as a complete sentence." So I think "independent clause" would be most fitting as the part of speech for these and similar terms, whether abbreviated or not. It may not be the easiest to understand for readers, but any alternative would be ambiguous, so we don't have much choice. —CodeCat 22:58, 12 January 2015 (UTC)

Whatever grammatical term we use is likely to be a problem because the term will be technical or it will clash with the basic notion that something spelled without spaces is a word (not a clause or phrase), or both. We have entries for MWEs that have the PoS header "Phrase" that are technically not phrases, but I don't think that the technical requirement that a "real" phrase be a constituent bothers very many ordinary users.

For now I am willing to plow ahead on the entries that have clear-cut PoS headers to replace these and wait for lightning to strike in the form of some conceptual breakthrough. Or for English speakers and their dictionaries to decide that interjections don't have to be emotional outbursts or that phrases don't have to have spaces between the component words. DCDuring TALK 23:19, 12 January 2015 (UTC)

I definitely support it in principle, but it is somewhat tedious to implement.

It would be handy to have a template to automatically provide the pronunciation for initialisms so that the pronunciation information implicit in the initialism header would not be lost due to the additional effort to provide a proper IPA pronunciation. DCDuring TALK 20:14, 2 January 2015 (UTC)

I did make one some time ago, but it was before Lua so it was a bit cumbersome to use, and never saw much use. I don't remember if I deleted it or not. —CodeCat 20:31, 2 January 2015 (UTC)

Tidiness strikes again? Even a poor template could be used to test the concept and discover missing features that might be useful. DCDuring TALK 20:52, 2 January 2015 (UTC)

Would that be {{IPA letters}}? DCDuring TALK 21:02, 2 January 2015 (UTC)

{{IPA letters|I|B|M|lang=en}} ⇒ DCDuring TALK 21:05, 2 January 2015 (UTC)

It would be handy if it worked directly on {{PAGENAME}}, perhaps with {{PAGENAME}} subst:'d. DCDuring TALK 21:23, 2 January 2015 (UTC)

You would have to use Lua for that. — Keφr 21:26, 2 January 2015 (UTC)

I hope you mean you in the sense of "one". DCDuring TALK 23:06, 2 January 2015 (UTC)

I support it. Some pages already use the right header (e.g. UFO). The example of fr.wikt shows that it's possible to do without these headers. Lmaltier (talk) 17:08, 3 January 2015 (UTC)

Filter created: Special:AbuseFilter/42. — Keφr 19:17, 12 January 2015 (UTC)

Thanks. Once we clean up the instances of {{en-acronym}} and {{acronym-old}} and any associated Acronym L3 headers, which usually need Pronunciation sections too, we will have {{initialism-old}} and {{en-initialism}} to work on, adding {{IPA letters}} under Pronunciation. DCDuring TALK 22:50, 12 January 2015 (UTC)

Note the filter has not been enabled yet. Also, NEC has been modified not to offer the headers discussed here. — Keφr 18:23, 14 January 2015 (UTC)

Can we insert proto‐words into translation tables?

If I desired to add Proto‐Germanic to a translation table for cow, could I get away with it? --Romanophile (talk) 01:26, 3 January 2015 (UTC)

No, as far as I know, the common agreement is that only attested terms can be in translation tables. —CodeCat 01:48, 3 January 2015 (UTC)

I for one do not want to see protolanguages in translation tables. —Aɴɢʀ (talk) 08:58, 3 January 2015 (UTC)

How about no links to unattested form in translation tables. Renard Migrant (talk) 15:09, 3 January 2015 (UTC)

I wouldn’t mind reconstructed terms in translation sections. I bet there are more people who want to know what’s water in PIE or PG than people who want to know what it is in Minica Huitoto, Northern Emberá or Lijili. — Ungoliant ^(falai) 16:19, 3 January 2015 (UTC)

The information is of interest, but it doesn't belong in a translation table with attested words. Maybe in a "See also" section? Chuck Entz (talk) 17:28, 3 January 2015 (UTC)

I can't imagine anyone wanting a translation that isn't attested. That is, a translation which according to evidence, may never have been used by anyone, ever. Renard Migrant (talk) 19:52, 3 January 2015 (UTC)

I would not be opposed to things like Appendix:English–Proto-Indo-European glossary, Appendix:English–Proto-Semitic glossary, Appendix:English–Proto-Algonquian glossary and the like, though, just like proto-language word are in Appendix space rather than mainspace. —Aɴɢʀ (talk) 21:07, 3 January 2015 (UTC)

As a translation- no. The point is that someone might want to know about the history of a particular concept within a given language family. Right now, that requires either fishing through translations in descendant languages looking for references in the etymologies, or browsing through the appendix to find the entry. That's why I suggested putting it in "See also"- there's no implication beyond there being some connection that might be of interest. Of course, it could easily be overdone, but that just calls for restraint, not a categorical exclusion. Chuck Entz (talk) 21:18, 3 January 2015 (UTC)

Make Wikisaurus a separate wiki

Hopefully this doesn't come across as blasphemous, but it seems to me that Wikisaurus should be made into a separate Wikimedia wiki project instead of existing as a sub-namespace within Wiktionary. The way this project is currently, it is quite problematic to search for thesaurus entries on words, because they're mixed in with the dictionary entries, and "Wikisaurus:" works like a kludgy title prefix on every single thesaurus entry. Also, it's awkward within a Wikimedia project that every single linked thesaurus word on another thesaurus article entry must be wrapped in a "ws" template. It seems that ideally, it would be desirable for plain word links within Wikisaurus to link within Wikisaurus, and if we wanted any links between Wiktionary and Wikisaurus then that should use proper inter-wiki links. Also, the Wiktionary articles already have static synonym lists in them; wouldn't it be desirable to have this be dynamically generated cross-wiki from Wikisaurus? Also, wouldn't it be desirable to have an automatic reference link for every word entry between these two projects, similar to how we have template-based automatic links between Wikipedia and Wiktionary? It seems very substandard to keep Wikisaurus as simply a static sub-namespace, and I feel like it needs to be developed into something much more tailored to working like a thesaurus site. — This unsigned comment was added by Wykypydya (talk • contribs) at 11:23, 3 January 2015 (UTC).

Burn the witch!

No, seriously. It would complicate things for the sake of… what exactly? Most links in Wikisaurus do, in fact, point to regular entries. — Keφr 12:01, 3 January 2015 (UTC)

No, I think that this belongs to the Wiktionary project. Here, words can be found through the search box, through categories and through thesaurus pages, these 3 methods should be kept, all of them are very useful, depending of what the user needs. A thesaurus page can provide thousands of words (not only synonyms), its contents cannot be moved to normal pages. But I would rename Wikisaurus to Thesaurus, this would be less kludgy... But something should be done about phrasebooks entries: they should be grouped into topical pages, just as in all paper phrasebooks, or they could be moved to another project (Wikiversity?) Lmaltier (talk) 17:22, 3 January 2015 (UTC)

Renaming rhyme pages vote

Could you please post your abstains (or other votes, as applicable) to Wiktionary:Votes/2014-09/Renaming rhyme pages so that the vote shows explicitly editor indifference, or maybe even gets a clearer outcome? Thank you. --Dan Polansky (talk) 11:50, 4 January 2015 (UTC)

Category:Terms by their individual characters by language automation in Lua

Happy new year to all. I've just adapted my "rare letters" identifier Lua module here to deploy this automatic categorization like on the French Wiktionary. This example works, but we'd rather invoke this module from the existing templates, like Template:en-noun. JackPotte (talk) 16:14, 4 January 2015 (UTC)

To clarify what was said at Module talk:languages, the idea is to include a list of the letters for a language in our data modules. That way, any entry containing a character not appearing in there can automatically be added to a category. The question, though, is which letters should be considered unusual in a given language. Some languages don't natively use letters like q or x, but still have them in many loanwords. —CodeCat 16:30, 4 January 2015 (UTC)

Also an implementation detail: While it seems sensible at first glance to list the letters only in one case form, case conversion is actually language-specific. A good example is Turkish: it has separate dotted İ i and dotless I ı. Using the "standard" case conversion would give incorrect results for Turkish. —CodeCat 16:33, 4 January 2015 (UTC)

For example I've let "é" into this letter list for French (Module:Rare letters/data, which can be filled with upper or lower letters) in spite of the detection of nearly 150,000 entries. JackPotte (talk) 16:36, 4 January 2015 (UTC)

I don't think that's the right approach, though. Unicode contains thousands of characters. We want to list the characters that are not rare, so that any that are not in the list are categorised. —CodeCat 16:38, 4 January 2015 (UTC)

In a second time we could gather in one category the letters which are unknown by the script. JackPotte (talk) 19:25, 4 January 2015 (UTC)

Right now, basically all the French letters (and English- the list for both is identical) with diacritics and all the digraphs are listed in (Module:Rare letters/data as "rare". That's not just isolated silliness: it's created redlinks to Category:French terms spelled with À‏‎ and Category:French terms spelled with Ç‏‎ from its use in just one entry. If this template were implemented in all French entries with the current data, a substantial part of all the French entries would be in one or more of these categories, and many of the categories would be so huge as to be pretty much useless. Please don't add this to any more entries without a consensus as to which characters should be included in the "rare" lists. Chuck Entz (talk) 04:11, 6 January 2015 (UTC)

Actually we are now all able to measure objectively and precisely the numbers of these presumed rare letters. For example "é" represents 341 342 of the 1 366 145 French articles, so 25%! Consequently we can now remove it from our categories if the consensual criteria becomes 1% at the light of these global digits. JackPotte (talk) 19:35, 6 January 2015 (UTC)

I think we need to discuss why we want such categories: remember that categories aren't for classification, but for navigation. The question you have to ask is "why would someone want a list of entries with a given letter in it?". I would say that relative frequency doesn't make a such a list of interest: even the rarest letters in the English alphabet such as j,q and x are found in lots of ordinary words- it's only unusual contexts like word-initial x or q not followed by a vowel that are of interest. French letters in the normal French alphabet aren't worthy of categorization- who wants to look through a huge list of French entries with ç in them? It's only letters that aren't part of normal French, and that make a word unusual by their presence, that should be categorized. Chuck Entz (talk) 15:03, 8 January 2015 (UTC)

Lemma dilemma

My issues are more involved than simply a lemma, but the title sounded so cool.

I've got some experience on Wikipedia, but I'm working on my first Wiktionary entry. It is the Indonesian word pengairan which is not yet complete. A number of things came to my mind as I attempted to make this word part of Wiktionary.

1. What is a lemma? According to Wiktionary:Lemmas, "When a word has multiple distinct forms, the lemma is the main entry at which the definitions, etymology, inflections and such are placed." Also, "For nouns, the lemma is normally the form that is used as the singular subject of an intransitive verb." Pengairan meets these criteria except for the etymology, since that information belongs with the root air, the prefix peng- and the suffix -an used to form the word. However, if we look at the second Wiktionary definition of lemma, it says, "The canonical form of an inflected word; ie the form usually found in dictionaries." Here is where the conflict arises. Pengairan will certainly be found in Indonesian dictionaries. However, it will be listed under a for air, not under p. This is true of all Indonesian words formed by a base and one or more prefixes and/or suffixes. By way of example, the official dictionary of the language printed by the Indonesian government has a listing for air. It gives two definitions. These are followed by six well-known proverbs in which the word is used with each proverb explained. Next, there are 171 idiomatic phrases that begin with the word air. Finally, there are six words derived from air listed: mengairi, pengairan, berair, perairan, berpengairan and keairan. None of these words can be found in the dictionary under m, p, b or k. Indonesian dictionaries treat words like this in a manner similar to the treatment English dictionaries afford inflections. In that regard, these words aren't lemmas, since the entire concept of a lemma is different. Were a dictionary of the Indonesian language to be written (for printing on paper) using the methodology employed in writing English dictionaries, these words would certainly have separate entries. For that reason, I've placed pengairan into the category for Indonesian lemmas. However, I would like to know what others think. Should the lemma category for non English words be for things that are lemmas as determined in the way they are in English, or should the foreign language's concept of a lemma be used?
2. Other than pronouns, Indonesian nouns are pluralized by doubling them. So, the plural of pengairan is pengairan-pengairan. Alternatively, the word para can be placed in front of the noun to indicate plurality as para pengairan. I created an entry for pengairan-pengairan. However, should we consider something else? Every countable Indonesian noun other than pronouns would have an entry that simply doubles the word. This seems to be a large number of entries that may be better handled another way. It might make sense to create an entry for pengairan-pengairan and have it redirect to pengairan. On the pengairan entry, the plural could be shown as an inflection but not linked. Folks should be mindful that plural forms are only used in Indonesian when they are needed to disambiguate. For instance, the word for dog is anjing, and the word for cat is kucing. People say, "She has two anjing and three kucing," not "two anjing-anjing and three kucing-kucing," since the context makes it obvious that the word is plural. This means the plural inflection has less importance in Indonesian than it does in English, and doesn't appear often.
3. Aside from plurality, other inflections apply to many Indonesian words. Nouns do not have possessive forms. Instead, they just follow the nouns they own. For example, "anak" means child, and "anjing anak" means the child's dog. However, Indonesian nouns do have possessed forms if they are owned by singular pronouns. These are the same for every Indonesian noun that can be owned. So, "anjingku" means my dog, "anjingmu" means your (singular) dog, and "anjingnya" usually means his, her or its dog. The -nya suffix can also act as a definite article. So, "anjingnya" can also mean the dog. Either way, it adds specificity to the base word. There are no irregular forms of these. They apply to every Indonesian noun. Should separate entries be created for pengairanku, pengairanmu and pengairannya, since each of these is a separate word?
4. Another inflection issue in Indonesian is the suffix -kah. If pengairan were to appear as the first word of a question, it could be pengairankah. This alerts the reader or listener that a question is coming without changing the meaning of the word. It is only used where the writer or speaker thinks it is necessary. Often, this would be at the start of a lengthy question. Usually, voice inflection alone is sufficient to make a sentence an obvious question to a listener. Question marks are used to end written questions as in English. So, -kah is omitted far more often than it is used. Should a separate entry be created for pengairankah, since it is a separate word? A vast number of Indonesian words can be used to start questions meaning each of these would have an entry for the -kah inflected form.
5. Finally, there is also an inflected form with a -lah suffix. This is most commonly used to soften a message. For example "duduk" means to sit. It can stand alone as a sentence in the imperative mood. However, by itself, "Duduk," can sound like a command given to a child. English speakers inviting someone into their homes might say, "Please, sit," as they welcome their guests. Indonesians can say "Silakan duduk," which means the same thing or simply "Duduklah," which isn't the same as saying please but is also not a stern command; it is polite. The suffix -lah can also be used to emphasize a word and will usually appear as the first word of the sentence. For example, "Pengairanlah bisa memecahkan masalah kami," is the same as saying, "IRRIGATION can solve our problem." The suffix -lah serves to highlight pengairan. This usage is fairly common and can apply to words across different parts of speech. Does pengairanlah merit its own entry in Wiktionary?

Of course, these questions are not really all about pengairan. They are about how these peculiarities in the Indonesian language should be handled in Wiktionary. Perhaps similar issues have arisen with other languages (or maybe these issues with Indonesian have been discussed before). I'd be delighted to read through such prior discussions if someone could point me in the right direction. Thank you for your patience, if you made it all the way here, and thank you for any thoughtful reply you provide. Taxman1913 (talk) 20:26, 6 January 2015 (UTC)

1. Why is pengairan not considered a lemma by other dictionaries? Being morphologically predictable is one thing, but morphology alone is not enough to make a word. For example, in English, paintify is morphologically and semantically transparent (the noun paint + -ify), but that doesn't mean that people actually use it. Considering pengairan a non-lemma form of air would only make sense if any noun has a peng- -an form.
2. I would say that the Indonesian plural shouldn't be included, as it is fully predictable, has no exceptions, and is written in a form that allows its parts to be recognised easily. This is similar to, for example, the English negative form of a verb, which just consists of a form of do paired with not and the infinitive. There's no objection to including it in an entry without a link, but that would only be useful for absolute beginners. I know no Indonesian but I still know that doubled nouns are plural.
3. I think the possessive forms of nouns should be included, as unlike with plurals, it's not obvious what parts pengairanku is made up of at first glance. Likewise for the definite form. Irregularity doesn't really matter, compare for example Esperanto where all words are regular, but we still include inflections for it. I think we also have possessive entries for Hungarian, Finnish and Turkish? It's not clear where to draw the line, though.
4. If -kah can be attached to any word, then it behaves similar to the Latin suffixes -que, -ve and -ne, the English -'s, or the Finnish -kin and -kaan. We don't include entries for words which have these suffixes. I'm not sure why not, exactly, but it could just be a purely practical consideration: if they can be attached to any word, then you'd have to double, triple or quadruple the number of entries (lemma and non-lemma) for that language, without a very clear benefit. See also the deletion discussion for satisne.
5. This seems more or less the same as above; if it can be attached to any word, it's probably better not to have entries for it.

—CodeCat 20:53, 6 January 2015 (UTC)

Note that some suffixes are systematic in English~, too (e.g. -like can be attached to any noun, according to my Pocket Oxford Dictionary), but this is not a reason to omit these words provided that they are attested. In my opinion, requiring attestations if sufficient, and this requirement would reduce drastically the number of potential entries. Lmaltier (talk) 21:26, 6 January 2015 (UTC)

The point about Indonesian lemmas is that given the second definition of lemma right here on Wiktionary, pengairan doesn't look like one, because you will not find it under "p" in an Indonesian dictionary. It will be under "a" for air. The Indonesian government's official dictionary defines the word lema (Indonesian for lemma) as (1) Kata atau frasa masukan di kamus di luar definisi atau penjelasan lain yang deberikan di entri; (2) butir masukan; entri. I'll translate that as (1) The word or phrase of a dictionary entry outside the definition or other explanation given in the entry; (2) An entry. Masukan (as used here) and entri are synonyms. The definition of masukan is not helpful as it is frequently used outside linguistics contexts. Entri is defined as (Linguistics) (1) Kata atau frasa di kamus beserta penjelasan maknanya dengan tambahan pejelasan berupa kelas kata, lafal, etimologi, contoh pemakaian, dan sebagainya; (2) lema. I'll translate that as (1) A word or phrase in a dictionary along with the explanation including additions to the explanation in the form of part of speech, pronunciation, etymology, usage examples, etc.; (2) lemma. So the first definition of lema could lead to a conclusion that air is a lema while its derived words are not, since they are presented in the explanation and not outside it. However, with lema and entri being used to define one another in secondary definitions, we are left wondering whether lema could also include the entirety of a dictionary entry as entri clearly does. Whether pengairan is deemed a lemma by Indonesian linguists has no bearing on whether it merits its own entry in Indonesian dictionaries. It is a derived term and listed under its base word which makes it look like something other than a lemma according to the definition of a lemma on Wiktionary. Taxman1913 (talk) 21:56, 6 January 2015 (UTC)

CodeCat, your point about pengairan being a lemma unless all Indonesian nouns have peng-*-an forms is an excellent one. The prefix peng- is a form of the prefix pe- which can also take the form of pem- or pen- depending upon the first letter of the base word. Both nouns and verbs can have one or both of these affixes added to them. For example, main means to play, pemain is player, mainan is toy, and pemainan is game. While these affixes are common and can be used with both nouns and verbs, Indonesian people aren't free to just make up all kinds of words by using them. According to the government's dictionary, pengair and airan are not words. Similarly, a word I used above, lema, has no derived forms resulting in pelema, lemaan or pelemaan. So, it is clear that words formed in this manner have unique characteristics that lend much credence to the argument that they are truly lemmas. Taxman1913 (talk) 22:19, 6 January 2015 (UTC)

CodeCat, your reasoning about Indonesian plurals makes sense. Unless someone presents a compelling reason why an entry for each of them is needed, I'm going to delete the one I created. Taxman1913 (talk) 22:22, 6 January 2015 (UTC)

I think attestation of the specific possessed inflections of Indonesian nouns as suggested by Lmaltier would drastically reduce the number of entries. How often will someone use the phrase "my irrigation"? While the construction of pengairanku is obvious to me, I can see CodeCat's point that most users of English Wiktionary not familiar with Indonesian word contruction will disagree. So, if attested, the word has a place here. Taxman1913 (talk) 22:32, 6 January 2015 (UTC)

Someone may not normally say "my irrigation", but Wiktionary entries don't always bother to indicate which parts of a word's inflections don't exist, especially if there are many of them. For example, it's clearly possible for one form of a Latin verb, out of the dozens it has, to happen to not ever be used (at least that we can tell). But it wouldn't make much sense to require excluding this in each case individually; it would become a nightmare to manage it. So for the sake of convenience, we sometimes list possible forms, even if they may not actually be used very much, or at all. This is not a hard rule of course, because in some cases we do exclude certain parts of the inflection, like comparatives for adjectives that are not comparable. For Indonesian nouns, I suppose we could include an approach that's somewhat similar, with "ownable" being like "comparable": there would be ownable and unownable nouns.

But we have to be careful, as someone might not normally use a form, but they still exist in case anyone wants to. For example, rain and its translations in many languages are often considered to be "impersonal" verbs, having only a third-person singular form (or equivalent), as there is not normally any specific thing that rains. But in figurative meanings or poetry, you might indeed find we rained. So we shouldn't be too quick to say that an Indonesian noun is unownable, because even though it's not "logical", there may be cases where someone has used those forms anyway. —CodeCat 22:43, 6 January 2015 (UTC)

I wouldn't consider rain an impersonal verb unless you promise never to rain on my parade. Taxman1913 (talk) 22:51, 6 January 2015 (UTC)

CodeCat, I agree with you about -kah and -lah forms. They are even more common than the Latin -que and the English -'s, since those are for nouns. These two can be attached to just about anything including prepositions and adverbs. But your mention of -que and -'s makes me think once again about -ku, -mu and -nya. Perhaps these are overkill as well. Even though pengairanku might not look obvious, are good quality entries for both pengairan and -ku sufficient? Even if attested, perhaps they're unnecessary. I could probably find an example of dog's somewhere. But we don't need a definition for it on Wiktionary. Taxman1913 (talk) 22:46, 6 January 2015 (UTC)

I'm going to work on creating an Indonesian noun template and circulate it for commentary. Taxman1913 (talk) 22:53, 6 January 2015 (UTC)

Actually, Latin -que can be attached to anything as well. It's often found attached to conjunctions for example, or indeed any other word that appears first in a clause.

I understand your point about the possessives, and there isn't really a clear line we can draw to say "this is ok to include" and "this is not ok", or at least not without just making up an arbitrary reason. The best I can come up with is this: Possessives only modify the meaning the noun they attach to, so they are "word-level" suffixes. But -kah and -lah are "clause-level" suffixes, they change the sense of whole sentences at a time. At least, that's how I understand it, please correct me if I'm wrong. So with that distinction, we could decide on a rule that words with clause-level suffixes/enclitics don't get their own entry.

We can also think of practical reasons. For possessives, we know that only nouns have them, and only some nouns at that (going by the "my irrigation" example). But the other suffixes can be attached to anything, and I'm guessing that they can even attach on top of other suffixes like possessives, and perhaps even on top of each other (is -lahkah possible, or something similar?). The number of possible combinations quickly becomes unmanageable. So we could decide to limit it to possessives, just to avoid creating a huge mess of possible combinations. —CodeCat 22:57, 6 January 2015 (UTC)

The Latin I studied for four years in high schoool is obviously rusty after 30 years. I only recalled the usage of -que as a conjunction for nouns until consulting the definition. Hopefully, Fr. Tighe and Mr. Scott can find it in their hearts to forgive me.

Your word-level/clause-level theory makes sense, and it points me more firmly into the direction of not including an entry for Indonesian possessed forms. You are correct that -kah and -lah change the sentence without changing the meaning of the word to which they attach. The change is subtle; I would say they only change the tone of the sentence, not its basic meaning. I've spent about 85% of the past four years living in Indonesia and have not come across -kah and -lah used in combination, but it is easy to imagine someone wanting to do that. The logical form would be -kahlah as in "Pengairankahlah belum digunakan di sini?" Which translates as "Hasn't IRRIGATION been used here yet?" This accomplishes sending out the question alert with the opening word of the sentence while also highlighting that word. So, -kahlah would be another possibility that could be affixed to nearly all Indonesian words including those to which -ku, -mu or -nya is already attached.

Turning back to the Indonesian possessed forms, the suffixes seem to me to also change the meaning of the sentence rather than the word. The mechanics are closer to the Latin -que than to the English -'s in this regard. Consider:

• dog's = belonging to the dog; this word is no longer purely a noun.
• puerque = and the boy; the meaning of puer is unchanged, and -que ends up replacing what could be written as a separate word in the sentence as in et puer.
• anjingku = my dog; the meaning of anjing is unchanged, and -ku ends up replacing what could be written as a separate word in the sentence as in anjing aku.
• anjingmu = your (singular) dog; the meaning of anjing is unchanged, and -mu ends up replacing what could be written as a separate word in the sentence as in anjing kamu.
• anjingnya = his dog; the meaning of anjing is unchanged, and -nya ends up replacing what could be written as a separate word in the sentence as in anjing dia.

Illustrating it in this way leads me to conclude that the Indonesian possessed forms are mechanically identical to Latin words constructed with -que and do not need separate dictionary entries.

This is also true when -nya is used as a definite article.

• anjingnya = the dog as opposed to dog; the meaning of anjing is unchanged, and -nya has its own meaning that just happens to be attached to the word.

The definite article is only used in Indonesian for disambiguation. Since the construction is the same as the third person singular possessed inflection, it often is impossible to tell what was in the speaker's mind when it was used as in "Gadis takut karena anjingnya." This could be "This girl is afraid of the dog," or "The girl is afraid of his dog," when translating to English. Only context can provide the correct answer. It really doesn't matter much. We can understand that there is a specific dog that makes the girl afraid. If we said, "Gadis takut karena anjing," we would understand that the girl is afraid of all dogs. No matter what, anjing remains dog and doesn't become somethng else with -nya affixed to it.

I'm in favor of no separate entries for Indonesian words with -kah, -lah, -ku, -mu or -nya suffixes. I also feel that separate entries are not needed for Indonesian plurals that are simply doubled. Hopefully, I can create a template that will present this in a clear way. I'm not sure there are nouns that cannot be owned. For instance, we might think arithmetic cannot be owned until someone says, "Your arithmetic is flawed." Nevertheless, I'll build an option into the template to make nouns unownable. Taxman1913 (talk) 07:46, 7 January 2015 (UTC)

But we do include possessives for Hungarian and Turkish nouns. Compare hal or hal for example. I don't think the Indonesian possessives work any different from that. So at least there is precedent for including them. I disagree with you that they work more like Latin -que, because like I said the possessives only modify a single word. "my" clearly belongs to "dog" and can't somehow be applied to the whole sentence. Your example of "puerque" shows that -que can also apply to a single word, but I think that's a wrong way of seeing it. -que joins two words, or two clauses, and therefore applies to more than just a single noun. Furthermore, you can imagine saying "a big dog and a small cat"... which word does -que attach to then? I would guess that it attaches to "small", which clearly shows that it's not about the noun, but about the whole phrase, and that it always attaches to the first word in the phrase. How would this work for Indonesian possessives? —CodeCat 15:12, 7 January 2015 (UTC)

Since Latin word order is free, -que attaches to whichever word is first: canis magnus felisque parvus or magnus canis parvusque felis, which just proves your point. —Aɴɢʀ (talk) 15:52, 7 January 2015 (UTC)

Here are a few Indonesian phrases for us to kick around:

• a big dog and a small cat = seekor anjing besar dan seekor kucing kecil
• my big dog and my small cat = anjing besar dan kucing kecilku
• your big dog and her small cat = anjing besarmu dan kucing kecilnya
• a big dog and my small cat = seekor anjing besar dan kucing kecilku
• his big dog and a small cat = anjing besarnya dan seekor kucing kecil

We can see from the above that when an adjective modifies the possessed noun, the possessive suffix usually attaches to the adjective which ordinarily follows the noun. Some adjectives (usually those indicating quantity) normally appear in front of the nouns they modify. Here are more illustrative phrases:

• many dogs = banyak anjing
• my many dogs = banyak anjingku

Here the possessive suffix attaches to the noun, because it is the final part of the phrase that is owned. This is a consistent characteristic. Similary, we consistently apply this in English in the opposite direction. We say, "my many dogs," not "many my dogs" unless we are in a particularly poetic mood.

Indonesian adjectives modifying two nouns joined by a conjunction typically refer to both nouns. This leads to more illustrations:

• blue pants and blue shirt = celana dan kemeja biru
• my blue pants and my blue shirt = celana dan kemeja biruku

Here we see -ku attached to an adjective that simultaneously modifies two nouns and indicates that the entire phrase is possessed.

I look forward to reading thoughts on these. Taxman1913 (talk) 18:53, 7 January 2015 (UTC)

These examples have convinced me that the Indonesian possessive suffixes are clitics and operate on a phrase-level rather than on a word-level. The crucial part, to me, is that -ku attaches to the last word, irrespective of what kind of a word it is. The suffix is evidently parsed as including the entire preceding noun phrase. So I don't think we should have entries for possessives. Is it the same for the definite suffix -nya? —CodeCat 19:01, 7 January 2015 (UTC)

I'm curious though, how you would write something like "a dog and a small cat" or "dogs and my cats"? You'd have to clarify in this case that the adjective or possessive applies only to the last noun phrase. —CodeCat 19:04, 7 January 2015 (UTC)

The definite article suffix -nya works the same way as the third person singular possessed -nya. It wraps around an entire phrase where applicable.

These three sufffixes show up in other places as well. They act in a manner similar to contractions at the end of some transitive verbs and prepositions. Here are some illustrations:

• bagi = for
• bagimu = for you, effectively a contraction of bagi kamu
• melihat = to see
• melihatku = to see me, effectively a contraction of melihat aku

This also affects a vast numer of words. Any thoughts? Taxman1913 (talk) 19:35, 7 January 2015 (UTC)

My first thoughts on your questions were

• seekor kucing kecil dan seekor anjing
• kucing-kucingku dan anjing-anjing

Reversing the word order solves the problem. That's what I would do to be sure I was understood. My wife, a native speaker, says that if the context is known, an adjective or a possessive suffix could apply to only the second of two nouns linked with a conjunction. Conversely, she suggests that redundant use of adjectives and/or possessive suffixes may be warranted where the listener might incorrectly assume they only apply to the second of two nouns linked by a conjunction. She offers the following:

• seekor anjing dan seekor kucing kecil
  • The presence of seekor as an indefinite article is sufficient to tip off the reader/listener that kecil only modifies kucing. Articles are frequently omitted in Indonesian in places where they would certainly be used in English. Indonesian indefinite articles are cumbersome. They vary depending upon the type of noun. The most common indefinite article is sebuah. Others include sebiji, sebatang and sebutir. Seekor is used for animals. Ekor means tail. They are all multisyllabic and force you to take a pause. In my wife's opinion, this is enough to separate the two linked nouns.

For "dogs and my cats," my wife agrees that reversing the word order is the best way to ensure you're understood. You could just go with anjing-anjing dan kucing-kucingku if the context makes it obvious that only the cats are yours. I should point out that if the listener already knows there is more than one dog and more than one cat, the nouns wouldn't be doubled. Taxman1913 (talk) 20:10, 7 January 2015 (UTC)

The same ambiguity issue can arise in English with the adjective at the front. "I went on vacation and spent the week drinking hot tea and beer." Alternatively, "I went on vacation and spent the week drinking cold beer and tea." Both of these are ambiguous. We can make the sentence more clear by using an adjective to modify each direct object. The same can be done in Indonesian to produce more clarity. Taxman1913 (talk) 20:44, 7 January 2015 (UTC)

My two cents:

With agglutinative languages, there are many components, which are attached to words. It's not really necessary to create all forms, IMHO, even if they are spelled without a space. Japanese, Korean have a lot of forms, which are attached, e.g. 커피하고 머핀 (keopihago meopin) "coffee and muffin" where 하고 (hago) is attached to 커피 (keopi). Indonesian doesn't seem to have a lot such forms and possessive forms are quite predictable. Arabic also has enclitic possessive suffixes - my, your, his, etc. It's part of the language grammar. E.g. بَيْتِي (baytī) means "my house" = بَيْت (bayt) + enclitic ـِي (-ī) but I don't think we need such entries. Arabic doesn't need definite entries like الْبَيْت (al-bayt) "the house", even if the definite article is attached to the beginning of a word.

There is no consistency in how some languages are treated. Scandinavian languages include definite forms of nouns but Bulgarian/Macedonian don't. Albanian definite nouns change their forms, e.g. feminine indefinite -ë changes to -a, so it would probably make sense to include them but redirects or definite forms in the header would suffice.

The case with "pengairan" is trickier. It reminds me of Arabic roots. Best Arabic dictionaries store information by the root consonants and having all words separately sometimes makes dictionaries too large but it is sometimes hard to determine the root consonants, e.g. فَتَاة (fatāh) is derived from ف ت ي (f-t-y) or ف ت و (f-t-w). So, of course, it's better to have separate entries.

Wiktionary doesn't boast good coverage of Indonesian. So it's better to focus on lemmas, not predictable forms, IMO.--Anatoli T. ^{(обсудить}/^вклад) 23:04, 7 January 2015 (UTC)

Two cents? That was worth at least a dollar, Anatoli. Thank you for such a thoughtful and thorough analysis. It seems there are no objections to excluding -ku, -mu, -nya, -kah and -lah forms of Indonesian nouns from having their own entries. That would apply to adjectives as well, since the analysis of the issue would be identical. Further, plurals of Indonesian nouns do not need entries. Would it be useful or advisable to have an infobox for Indonesian nouns and adjectives alerting readers to the possibility that the word may appear with one or more of the aforementioned suffixes attached? If so, is the best place for this with the headword or in the area where declensions would go? Taxman1913 (talk) 19:47, 8 January 2015 (UTC)

I tried to make pengairan-pengairan completely blank assuming that would inspire a bot (or a human) to delete it. The word cannot be attested and, therefore, doesn't meet the criteria for inclusion. The system thought I was vandalizing and wouldn't let me make it completely blank. It insisted I leave a language and part of speech on the page which I did. It suggested I contact an administrator if I think my edit was constructive. One would think something so sophisticated to protect against vandalism would provide a link to let an administrator know what you tried to do. But that isn't the case. Taxman1913 (talk) 20:00, 8 January 2015 (UTC)

Your edit showed up in Special:AbuseLog. Though I think only I look at it semi-regularly (mostly watching out for spam filter false positives). If you had used {{delete}} or {{rfd}}, the filter would not have been triggered. Whether the request would have been granted is another matter of course. — Keφr 20:04, 8 January 2015 (UTC)

Thank you, Keφr. Taxman1913 (talk) 20:36, 8 January 2015 (UTC)

Rename tr= parameter to xlit= or similar

There are quite a few entries where tr= as been interpreted as meaning "translation". This is not very surprising. To prevent this kind of misunderstanding and confusion, I'm proposing to rename this parameter to xlit=, or something else that is less liable to be confused. —CodeCat 18:31, 7 January 2015 (UTC)

On the French version, we use "R=" like "Romanization". JackPotte (talk) 19:50, 7 January 2015 (UTC)

It should be easy to detect when someone is using it for translation just from the characters, no? (or with a language that should never be transliterated). DTLHS (talk) 19:54, 7 January 2015 (UTC)

PS: Romanization stands for an hyperonym of transcription and transliteration. JackPotte (talk) 19:58, 7 January 2015 (UTC)

|r= looks good. It's not misleading, also it's shorter. "Romanization" is also more accurate to describe what we pass to the parameter. --Z 22:36, 7 January 2015 (UTC)

• Don't rename. Rare confusion is not a sufficient reason for this kind of deviation from the previous practice. --Dan Polansky (talk) 22:39, 7 January 2015 (UTC)
  • And established practice (that is, the inertia of established editors), is not an argument against improvement. —CodeCat 22:50, 7 January 2015 (UTC)
    • I also don't think a change is necessary. It seems like a solution in search of a problem. —Aɴɢʀ (talk) 22:52, 7 January 2015 (UTC)
    • Established practice is an argument against low value added change, whether considered an improvement or not. Changes have costs, and these have to be weighed against minor or even hypothetical benefits. I ask the reader to check various historical revisions of our mainspace pages and see how changes made by CodeCat had drastically reduced legibility of these revisions, where the legibility is reduced by missing templates and by various module errors produced by templates. It is a pitiable state of affairs to look at. --Dan Polansky (talk) 22:55, 7 January 2015 (UTC)
      • Writing a dictionary in plain text was a terrible idea from the start and it still is, which is why we need all the templates to compensate, and why our templates and other technical infrastructure need so many revisions. I can't help it if Wiktionary was badly designed from the start. That's no reason not to try to fix it up some. I will continue to try to improve Wiktionary to the best of my ability, and provide a counterbalance to those who do otherwise, as long as I am active on this project. —CodeCat 22:59, 7 January 2015 (UTC)
        • I could not disagree more. I love wikitext, I love Wiktionary design and I dread the day on which the likes of you are going to remove the wikitext from the editor, and lock it down behind a WYSIWYG interface. --Dan Polansky (talk) 23:06, 7 January 2015 (UTC)
          • Sadly, that day won't come. My hope is to make Wiktionary semantically parseable, ideally something that a script could convert to another format like XML. In other words, to make Wiktionary's code about content only, and not about presentation. —CodeCat 23:10, 7 January 2015 (UTC)
            • "Established practice (that is, the inertia of established editors), is not an argument against improvement". It should be in template-space, CodeCat. Moving, renaming, and adding or removing parameters confuses the shit out of editors. You and Kephir take it far too lightly. Purplebackpack89 23:13, 7 January 2015 (UTC)
          • Love makes you blind, apparently. Wikitext has quite obvious limitations and drawbacks when it comes to creating a dictionary. If it were not for the comprehensive template infrastructure, maintaining category lists would be burdensome, and categories would often fall out of sync with the rest of entry content (just look at pages "tagged but not listed" on RFX). If it were not for standardised headers, distinguishing between definitions, etymologies and synonym lists would be much harder, if not impossible. Typing wikitext manually is error-prone; if it were not for sanity checks in templates and modules, the errors might linger in entries indefinitely. Remembering how every single template should be used is tiresome; even I cannot be bothered to do it, and I would be happy if I could just assume every template works the same. There are tons of boilerplate in each entry — entries being free-form wikitext pretty much forces us to have it; the alternative is complete chaos. Our translation lists would most probably be much smaller if we did not have WT:EDIT. If it were not for edit filters, the flood of spam and vandalism might be unmanageable. And each of these solutions I just mentioned is a brittle workaround, and entries still fall between the cracks anyway. Splitting entries into per-language pages gets proposed every once in a while, and is probably the most desirable feature of all, but if we were actually to do it, it would be a nightmare from both technical and copyright standpoints. Wikitext talk pages are obnoxious for those who know how to use them and confusing for those who do not. Free-form wikitext is an awful foundation on which to build a dictionary. Now waiting for the inevitable liberum veto thought-terminating cliché of "I disagree, just because". — Keφr 10:13, 10 January 2015 (UTC)
            • If someone needs any more proof that writing a dictionary in wikitext is a complete disaster, look at Special:Contributions/188.79.152.174. — Keφr 20:40, 13 January 2015 (UTC)

Digression by Ungoliant and Kephir

If it confuses editors, there’s nothing you need to be worried about. — Ungoliant (falai) 00:04, 8 January 2015 (UTC) Is that supposed to be some thinly-veiled insult about my edit pie? If so, User:Ungoliant MMDCCLXIV, it's neither true nor germane. The fact is, I do create entries, I do use templates, and some of them have been moved or merged multiple times since I got here. It's utterly confusing, and an example of the perfect being the enemy of the good. If you want a template with a particular name that does a particular thing, I'm fine with you creating it, but I'm not fine with you moving around other widely-used templates because you want them to have a different name. Purplebackpack89 16:20, 8 January 2015 (UTC) Sow the wind, reap the whirlwind. You throw thinly-veiled insults at CodeCat and Kephir whenever they try to make any proposal, as you did in this discussion. — Ungoliant (falai) 17:20, 8 January 2015 (UTC) I think you're mixing disgust with moving of templates with insults. Purplebackpack89 17:38, 8 January 2015 (UTC) Since you mentioned "edit pie"… Damned lies.  user year of first edit main namespace edits project namespace edits user talk edits thread edits number of blocks Silent Sam 2011 7 983 (98.3%) 1 (0%) 2 (0%) 0 (0%) 0 Ivan Štambuk (sysop) 2007 93 751 (89%) 2 146 (2%) 2 166 (2.1%) 22 (0%) 8 ZxxZxxZ (sysop) 2007 11 964 (78.3%) 347 (2.3%) 392 (2.6%) 140 (0.9%) 0 Atitarev (sysop) 2007 178 512 (90.3%) 5 782 (2.9%) 3 987 (2%) 167 (0.1%) 0 SemperBlotto (bureaucrat) 2005 319 702 (92.6%) 11 542 (3.3%) 6 151 (1.8%) 44 (0%) 6 kc_kennylau (sysop) 2012 10 613 (75%) 577 (4.1%) 483 (3.4%) 23 (0.2%) 0 BD2412 (sysop) 2005 89 732 (83.2%) 5 839 (5.4%) 1 355 (1.3%) 0 (0%) 4 Dan Polansky (patroller) 2006 78 028 (77.7%) 6 098 (6.1%) 2 324 (2.3%) 69 (0.1%) 12 CodeCat (sysop) 2007 106 442 (57.3%) 11 312 (6.1%) 3 665 (2%) 1 612 (0.9%) 2 Romanophile 2011 15 242 (79.8%) 1 449 (7.6%) 1 015 (5.3%) 13 (0.1%) 10 Stephen G. Brown (bureaucrat) 2004 132 470 (84%) 12 173 (7.7%) 3 388 (2.1%) 17 (0%) 3 Ungoliant MMDCCLXIV (sysop) 2011 50 004 (85.6%) 4 603 (7.9%) 939 (1.6%) 15 (0%) 0 Kephir (sysop) 2012 12 066 (53.9%) 1 919 (8.6%) 374 (1.7%) 204 (0.9%) 1 Yair rand (sysop) 2009 6 613 (32.9%) 2 847 (14.2%) 665 (3.3%) 219 (1.1%) 0 Liliana-60 (sysop) 2008 9 277 (46.7%) 5 303 (26.7%) 720 (3.6%) 55 (0.3%) 9 Ruakh (bureaucrat) 2005 26 974 (49.7%) 15 017 (27.6%) 3 928 (7.2%) 134 (0.2%) 4 Purplebackpack89 (rollbacker) 2010 1 277 (48.2%) 851 (32.1%) 309 (11.7%) 30 (1.1%) 7 Sorted by increasing percentage of project-namespace edits. (Apologies if anyone feels hurt by omission, but the bloody statistics page sometimes would just not load. And I think this list is long enough to illustrate my point.) Like all statistics, they should not be taken very seriously, but this does not mean they are worthless. Not every edit is born equal; edit warring and minor template adjustments may inflate edit count in certain namespaces. Some edits create errors, others correct them; others are vandalism reversions; yet others merely add requests; the most valuable ones add entries with full inflection, pronunciation and citations in one go. I consider 1 mainspace edit of Silent Sam's to be worth at least 20 of mine, for one. Some blocks have been handed by Wonderfool, others for relatively minor infractions, so block count should not be considered that much of a reliable counter of times the community has been pissed off. Year of first edit is not that good indicator of how long a user has been here; and since accounts and longer periods of inactivity are not that rare, an "edits per day" statistic would not necessarily be very helpful. And despite these arguable flaws, the discrepancies sort of speak for itself. As for individual anomalies: Yair rand's comparatively low percentage of mainspace edits may be explained by his 6295 edits to Category namespace (31.3% of all edits), so it cannot be considered lack of productivity. While Ruakh has been relatively highly active in meta-processes, it has never been a reason for blocking him (my subjective judgement is that he generally worked to contain drama); his blocks were not very serious anyway. I cannot immediately find much to say in defence of Liliana, though. Food for thought. — Keφr 19:48, 8 January 2015 (UTC) Your focus on percentages is misleading. If you sorted it the proper way; by total contributions, you'd see I actually lag far behind you and other editors. It's also meaningless, because a person is entitled to whatever edit pie he or she chooses. Having 70% mainspace edits instead of 60% mainspace edits does not make you a more worthy editor any more than inventing a tool gives you an excuse to personally attack editors. Editors are entitled to comment in Wiktionary-space as much as they want regardless of their number of contributions in mainspace. Purplebackpack89 22:05, 8 January 2015 (UTC) Nice job breaking the layout, for one. For two, I hope you noticed I did a lot of category edits back when the new {{topic cat}} template was introduced. Or did you omit that significant fact on purpose? -- Liliana • 22:20, 8 January 2015 (UTC)

Unfortunately, anything you propose tends to attract a disproportionate amount of vehemence and bitterness from certain quarters, which is getting very old. Still, on the merits, I'm very leery of changing something that's second nature to anyone who's edited much in non-Latin scripts just because of a small percentage that are confused (there are probably more who mistakenly use the language code "dk" for Danish, but no one is arguing we change that). I think the potential to alienate large numbers of our hardest-to-find type of editors by forcing them to relearn a basic part of their editing routine outweighs solving such a minor problem. Chuck Entz (talk) 03:37, 8 January 2015 (UTC)

Oppose. It’s not that common, in my experience. — Ungoliant ^(falai) 17:20, 8 January 2015 (UTC)

More pathological need to change things from CodeCat. Could you change other parts of your life and leave us alone? If the need is so strong you must change all parts of your life at all time, have you considered talking to a doctor about it? No this isn't sarcasm. Renard Migrant (talk) 17:28, 8 January 2015 (UTC)

Why don't you just go fuck yourself? That's a change I could agree with. —CodeCat 18:02, 8 January 2015 (UTC)

Could you answer the questions? Renard Migrant (talk) 22:28, 8 January 2015 (UTC)

• Oppose: Because of what I said about moving stuff confusing people. Purplebackpack89 17:38, 8 January 2015 (UTC)
• Count me as opposing. I am not convinced that the benefits outweigh the costs. — Keφr 20:05, 8 January 2015 (UTC)
• Oppose Though the recommendation CodeCat proposes would have been a good idea long ago, there is no evidence that this is a significant problem. In fact we have a completely unsubstantiated claim that it is sometimes a problem without a single instance. If, 1., it isn't worth collecting the evidence that it is a problem and , 2., most contributors to this discussion don't think it is a significant problem, then we should not waste more time on this. DCDuring TALK 22:30, 8 January 2015 (UTC)

Project Wiktionary Meets Matica Srpska

We would like to inform you about the project we are starting. The project's aim is to increase support for Open Knowledge / Free Content movement through establishing long term strategic partnership with the venerable cultural institution Matica srpska and to increase the quality, accuracy and volume of Wiktionaries through digitization of two dictionaries, while developing a potential model for future development of cooperation across Wiktionaries through targeted mobilization of the communities.

There are two activities within the project that we need your support for. First, we are preparing a list of lexicographical terms (it would contain approximately 100 terms) that needs to be translated into as many languages as possible, in order to ensure further work on the project and to create the foundations for other lexicographical projects in the future. For this task, we would use a separate application, but all of the terms would be inserted into Wiktionaries, as well (making approximately 10,000 new entries per Wiktionary, counting that the terminology would be translated into 100 languages). That would also serve as the preparation for translating the Serbian Ornithological dictionary, which is the second activity.

The Serbian ornithological dictionary encompasses all local names of birds living on the Serbian speaking territory. All names are specified under the appropriate Latin name in accordance with the contemporary classification system. This creates the opportunity to translate it easily to various languages, since the basic list of terms is in Latin and it is fairly small (370 species of birds).

We would like to try and motivate as many Wiktionary communities as possible to participate in translating these two dictionaries, especially since the benefit for each particular Wiktionary would be great - for example, if we succeed to motivate people from 100 Wiktionaries to participate, the amount of primary entries to these 100 Wiktionaries would be 3,700,000 (37,000 per Wiktionary). If we succeed to motivate just 20 Wiktionaries, the amount of entries to these 20 Wikitionaries would be 148,000 basic entries. Of course, these entries would be incorporated into the respective Wiktionaries according to the interest and the rules of each community.

With this project, we are opening cooperation with the venerable Serbian cultural institution Matica srpska [2] and we believe that this partnership will have major impact on future cooperation between Wikimedia organizations and similar institutions in Slavic countries. If this cooperation could be relevant to any other partnership you are trying to establish in your country or globally, we would be more than willing to share our knowledge and contacts.

Besides support in translation, we are open for Wikimedia volunteers to participate more substantially and thus build knowledge inside of the community on how to deal with this kind of data. For example, if you are willing to join the core team and help us in communication with the Wiktionary communities of the languages which you are speaking, please contact Milica (milica.gudovic@yahoo.com) or Milos (millosh@gmail.com) via email. The same goes if you are willing to contribute by coding in Python and/or PHP.

Please join us on project's discussion page or send us an e-mail. If you are willing to participate but are not sure in your knowledge of English please check the list of languages organizational team is speaking - there is a chance we can communicate in your native language.

We are very excited about this projects and hope that you will be part of it as well!

Looking forward to hear from you!

--Godzzzilica (talk) 15:55, 8 January 2015 (UTC)

How does our merger of Serbian, Croatian, Bosnian and Montenegrin into Serbo-Croatian affect the project? How does Matica Srpska feel about that? — Ungoliant ^(falai) 17:23, 8 January 2015 (UTC)

Serbian lexicography usually treats all those languages as one, with mentioning that something is a Croatian variant (for example, attorney is "advokat" in Serbian, while "odvjetnik" in Croatian; capital monolingual dictionary of Serbian language would have entry "odvjetnik", while mentioning that it's "hrv. for advokat"). This topic is much more problematic in Croatia than in Serbia. --Millosh (talk) 20:32, 8 January 2015 (UTC)

Said so, the best idea is not to insist on their formal position. If you don't insist, that would be treated as editorial choice of the English Wiktionary, not under the jurisdiction of Matica srpska. If you insist, they would likely give a kind of negative comment. --Millosh (talk) 20:32, 8 January 2015 (UTC)

Linguistically speaking, some things could be fixed. For example, одвјетник (Cyrillic spelling of odvjetnik) is a kind of neologism. It's strictly Croatian word and contemporary Croatian is written strictly in Latin alphabet; it could be transcribed for the linguistic and similar purposes, but if in Cyrillic alphabet, it's translated to адвокат. --Millosh (talk) 20:32, 8 January 2015 (UTC)

Oh, I've read everything in relation to the decision made few years ago now. The community didn't make a formal decision in relation to that issue. So, brace yourselves! This could become significant issue on English Wiktionary (again) after this project gets public attention. Note that this is likely an exclusive practice of English (and Serbo-Croatian) Wiktionary. --Millosh (talk) 21:45, 8 January 2015 (UTC)

Topical context labels

diff. Judging by that edit, User:Jamesjiao probably feels that context labels are meant to show limited usage or to disambiguate. Since this definition doesn't need disambiguating and is in general use, he removed the label. But the fact that the label was there in the first place means that someone else thought differently. That other person presumably thought that labels could also used merely to give extra information about what general semantic field or topic a term pertains to, even if it's not a term specific to it.

These two approaches have been at odds with each other for a while now, but I don't know if there has been a discussion on it recently-ish. If there has been, it was either inconclusive or not widely advertised so I missed it. So my question is, should context labels be used to indicate topics or other information, when they give no further on the definition itself that is not already apparent from that definition? That is, should labels be always restrictive/defining or can they also be non-restrictive? —CodeCat 00:02, 9 January 2015 (UTC)

I feel that in some cases people use labels because they are too lazy too add the categories manually. — Ungoliant ^(falai) 00:33, 9 January 2015 (UTC)

Well, it's true that some other dictionaries do add topical labels. So it's probably inspired by that. That doesn't mean I understand why those other dictionaries do it, though. —CodeCat 00:39, 9 January 2015 (UTC)

I see the logic of both kinds of labels, but the restrictive usage context seems to be an essential function for a reference that purports to be about words. We apply neither type of label consistently, even in a single PoS section, eg, car#Noun. I don't think that we have even well characterized what we mean by a usage context. Is it a group of people with shared vocabulary, eg, soldiers? Is it a situation, eg, military service? Is it a kind of technical expertise, eg, modern firearms design or use? DCDuring TALK 01:27, 9 January 2015 (UTC)

The less said about the arbitrariness of topical categories the better. DCDuring TALK 01:27, 9 January 2015 (UTC)

I personally feel that geology as a science is quite a specialised field that has its own set of jargons, which this term shouldn't be a member of, imho. One reason I removed this label is because I noticed that the term is already found in this category Category:nl:Landforms, which better suits non-jargon terms like this one. Jamesjiao → ^{T ◊ C} 01:34, 9 January 2015 (UTC)

One trouble with categories is that they are not visibly connected to a particular sense of a polysemic word. DCDuring TALK 02:32, 9 January 2015 (UTC)

The discussed edit (diff) complies with Wiktionary:Votes/pl-2009-03/Context labels in ELE v2. It also matches our dog entry, which does not use "zoology" label before "A mammal, Canis lupus familiaris, that has been domesticated for thousands of years, of highly variable appearance due to human breeding" definition. --Dan Polansky (talk) 18:13, 9 January 2015 (UTC)

Word-mining at Wikipedia

A development at Wikipedia may be of interest to Wiktionary editors. Over at w:Wikipedia:Typo Team/moss a new project is finding all words occurring within the English-language Wikipedia that are not mentioned in any Wikipedia article title, nor in any Wiktionary entry or Wikispecies entry. One of the aims is to identify and fix errors in Wikipedia articles; another is to identify and fix missing entries at Wiktionary and Wikispecies.

You are invited to review the current list and perhaps be inspired to create missing Wiktionary entries. -- John of Reading (talk) 14:22, 9 January 2015 (UTC)

• I've added a few. SemperBlotto (talk) 14:59, 9 January 2015 (UTC)
• The list is useful, but does it distinguish the different languages on Wiktionary? A word may have an entry on Wiktionary, but not in English. —CodeCat 15:08, 9 January 2015 (UTC)
  • No, it doesn't do that yet. -- John of Reading (talk) 16:05, 9 January 2015 (UTC)
    • Still, neat project! :) - -sche (discuss) 23:31, 12 January 2015 (UTC)

RFV - period for closing nominations

Wiktionary:Requests for verification/Header currently says:

Closing a request: After a discussion has sat for more than a month without being "cited", or after a discussion has been "cited" for more than a week without challenge, the discussion may be closed. ...

The same page says that a discussion can be archived a week after closure.

This double one-week period seems excessive. I propose this change:

Closing a request: After a discussion has sat for more than a month without being "cited", or after a discussion has been "cited" for more than a week without challenge, the discussion may be closed. ...

Thus, once an entry is cited, the discussion should be able to be closed immediately, after which there will still be one week for objections before the discussion is archived. Thoughts? --Dan Polansky (talk) 13:48, 10 January 2015 (UTC)

• If the sole reason for change is "This double one-week period seems excessive", why would we bother? In particular it doesn't seem excessive to me. DCDuring TALK 14:18, 10 January 2015 (UTC)
  • As per the header, when the entry is cited, one first needs to wait one week before closing the nomination, and them one week before archiving the nomination. You rarely close or archive RFV nominations, which may help explain that this does not seem excessive to you. To me, waiting one week after an entry has been cited before removing the nomination from RFV page (via archiving) seems sufficient, especially when the closing person is different from the person who added the quotations. --Dan Polansky (talk) 14:28, 10 January 2015 (UTC)

Since struck headers can be unstruck, there's no reason not to close immediately if the citations are apparently watertight. It would be bad faith if there are less than three certain citations. However if a mistake is made by striking a header for a term that isn't certainly cited, we can just unstrike it. In fact, we do. Renard Migrant (talk) 15:53, 10 January 2015 (UTC)

acronym, initialism, abbreviation template use

Looking at Category:English acronyms I note many entries that seem unpronounceable as words or, at least, not obviously pronounceable. I hope we have agreement that for purposes of our definitions that, for English L2 sections at least:

1. an acronym is an abbreviation formed from the initial elements (or just initials?) of the component words of the term it represents pronounced as a word (uses {{acronym of}})
2. an initialism is an abbreviation formed from the initials (or initial elements?) of the component words of the term it represents pronounced as the component letters of the initialism (uses {{initialism of}})
3. an abbreviation is anything not certainly or probably either an acronym or initialism (uses {{abbreviation of}}).

If something is pronounced either as an initialism or as an acronym in the senses above we should have a pronunciation section and not use {{acronym of}} or {{initialism of}}. We should probably require a pronunciation section for everything claiming to be an acronym.

Is the above a reasonable summary of existing preferences? It certainly is not carried out. I would suggest that, if there is agreement, we should perform some cleanup, especially on the entries in Category:English acronyms. DCDuring TALK 14:12, 10 January 2015 (UTC)

I was doing a course a few months ago where the tutor referred to the 'acronym' DLA which of course, isn't an acronym. It's even used erroneously to mean initialism, so yes, correct it as much as possible. Renard Migrant (talk) 22:53, 11 January 2015 (UTC)

I'd like more people to agree that this is the correct approach so that I don't duplicate the mistake being perpetrated with respect to {{eye dialect of}}. (See WT:TR#gub'mint.) DCDuring TALK 23:15, 11 January 2015 (UTC)

I just looked at an extensive discussion of use of the term initialism on Wikipedia. It was a debate between advocates of "precision" versus advocates of accessibility to a general population of readers. DCDuring TALK 16:28, 12 January 2015 (UTC)

I suppose some of the good news is that {{acronym-old}} remains on only some 200 entries. I don't know how many L2 sections have Acronym as a PoS header. I miss Ullmann's runs that produced lists of non-ELE L3, L4, L5 headers. DCDuring TALK 16:53, 12 January 2015 (UTC)

KassadBot used to keep records of headers. Renard Migrant (talk) 18:00, 12 January 2015 (UTC)

I can see the category populated by {{rfc-header}}, but not something comparable to User:Robert_Ullmann/L3#Invalid_L3_headers. He seems to have been way more diligent than anyone since. DCDuring TALK 19:18, 12 January 2015 (UTC)

Someone — I think it was User:DTLHS — made a WT:TODO list last year of all headers used on Wiktionary, which they and CodeCat and I were using to find and fix misspelt headers like "===Etymologoy===", but which also caught correctly-spelt but nonstandard headers. Perhaps they could generate a new list for you. - -sche (discuss) 23:41, 12 January 2015 (UTC)

You'd be amazed if you fix say, 5 per day, how quickly a list of 200 goes down to zero. I seem to think I fixed 700 Swedish entries with a declension table not under a header doing 10 per day. Renard Migrant (talk) 16:32, 13 January 2015 (UTC)

I've added all the entries I can find with the offending headers to Category:Entries with non-standard headers. Renard Migrant (talk) 00:24, 16 January 2015 (UTC)

Template:zh-see

User:Wyang and User:Mar vin kaiser seem to be implementing Wiktionary:Votes/pl-2014-12/Making simplified Chinese soft-redirect to traditional Chinese which not only doesn't close for another two weeks, it's on course to fail by a clear margin. What to do, nothing? Blocks? Mass deletions? Renard Migrant (talk) 23:17, 11 January 2015 (UTC)

What are you talking about? The discussion has reached consensus among Chinese-language editors. As I said before on my talk page, there is no point for a vote when all Chinese-language editors support the proposal. The vote is a means for a bunch of utter standers-by to dictate what chores others should do. Wyang (talk) 23:22, 11 January 2015 (UTC)

The objective of the project is a dictionary to be used by readers. What is the proportion of Chinese-language editors among readers? It's negligible. What editors should do should be dictated by what users need, like, use. Lmaltier (talk) 22:04, 14 January 2015 (UTC)

What Wyang said - we have reached an agreement. The vote doesn't have any rationale because it wasn't created by Chinese editors. There are legitimate concerns, though. Eventually and ideally, the simplified entries should display the same information as the traditional but the contents should be stored in one place. Having entries, which are out of sync, is a big problem. (Correctly formatted) traditional entries use both traditional and simplified - in synonyms, usage examples, etc, so there is no discrimination. --Anatoli T. ^{(обсудить}/^вклад) 23:28, 11 January 2015 (UTC)

In any case, how does deleting simplified entries with {{zh-see}} help the project!? --Anatoli T. ^{(обсудить}/^вклад) 23:30, 11 January 2015 (UTC)

Right you are. Thanks for bring this to my attention. Renard Migrant (talk) 23:38, 11 January 2015 (UTC)

Then, we should nominate Wiktionary:Votes/pl-2014-12/Making simplified Chinese soft-redirect to traditional Chinese for deletion perhaps? Renard Migrant (talk) 18:04, 12 January 2015 (UTC)

I am confused. What are you trying to say? — Keφr 18:05, 12 January 2015 (UTC)

That perhaps we should nominate Wiktionary:Votes/pl-2014-12/Making simplified Chinese soft-redirect to traditional Chinese for deletion. Renard Migrant (talk) 18:13, 12 January 2015 (UTC)

…because? — Keφr 18:40, 12 January 2015 (UTC)

A RFDO on the vote is unlikely to pass the 2/3 threshold, and seems pointless. The vote is fine. The voters have plenty of room for posting their rationale directly in the vote, or linking from the vote to locations where they posted their reasoning; they can also post "as per <person>". On another point, "consensus among Chinese-language editors" is not enough, IMHO, in part since the Chinese entries also serve readers, and since multiple of the editors who helped create Chinese entries are no longer around to oppose. On a related note, in Wiktionary:Votes/pl-2014-12/Making simplified Chinese soft-redirect to traditional Chinese, among the supporters I currently see two editors with anything like significant contribution to Chinese entries: Anatoli T., and Wyang. --Dan Polansky (talk) 18:26, 12 January 2015 (UTC)

If the vote fails, which seems likely, then we maintain the status quo, which is to use {{zh-see}} anyway. Renard Migrant (talk) 18:55, 12 January 2015 (UTC)

Are you trying to solve a problem? If so, what is the problem statement and what are the tentative solutions that you offer? --Dan Polansky (talk) 19:06, 12 January 2015 (UTC)

No. Renard Migrant (talk) 19:11, 12 January 2015 (UTC)

There is no point in keeping the vote open now that it is being ignored; this is a perfectly reasonable observation. Kaixinguo (talk) 19:14, 12 January 2015 (UTC)

The vote is not being ignored: people keep posting to it. Via the vote, we are getting increasingly better evidence and information about the actual scope of support and opposition; we would not have that without the vote. Furthermore, your defeatist stance is dangerous, since it encourages certain types of editors to think like this: "go ahead, ignore any vote or opposition, and the opposers will just give up". For those who support, posting "support per <person>" costs close to nothing, likewise for those who oppose, so I see little point in trying to make prophesies about the outcome. A better course of action is figure out the right stance, and let it known in the vote. --Dan Polansky (talk) 19:39, 12 January 2015 (UTC)

Merging Finno-Volgaic, Finno-Samic, Finno-Permic and Finno-Ugric into Uralic

The internal subgroup of the Uralic family is not usually settled upon, and many competing ideas still exist, some with greater or lesser support. The first two families in the header (or three, if Wikipedia is to be believed) are still in dispute. Finno-Ugric has much more support, but as a reconstructed language it's virtually identical to Uralic. In fact, even Finno-Permic does not differ significantly from Uralic as a whole at least phonologically. The tradition among many Uralic linguists is to label reconstructions based on the branches it's attested in. So for example, an etymology would label a reconstruction Finno-Ugric if it's not found in the Samoyedic languages.

At Wiktionary talk:About Proto-Uralic, User:Tropylium (who I definitely trust on Uralic linguistics) suggested that we should eliminate the various subgroups and their associated proto-languages as separate languages on Wiktionary, and merge them into Uralic. Treating each language or branch separately is not practical for Wiktionary, where we would end up with a long string of identical words in etymologies, like the example Tropylium gives for Finnish kala. It gets even worse if we could potentially create three or more separate entries for the various proto-languages in the chain, all of which would have identical forms for almost any word. It's much more practical, as in the current Finnish entry, to jump straight from Finnic to Uralic.

So the proposal is to treat Proto-Finno-Ugric, and perhaps also Proto-Finno-Permic, as simply dialects of Proto-Uralic (which they were in reality, in all likelyhood). Entries would be moved and etymologies adjusted accordingly, so that the categories for terms derived these languages would end up either empty, or subsumed under Uralic just like we do with Category:Terms derived from Anglo-Norman or Category:Terms derived from Late Latin. References to Finno-Volgaic and Finno-Samic would be removed altogether.

I just want to note that this is explicitly not a statement that these are not valid subgroups or languages. This is just a matter of practicality, just like when we choose to merge any other group of languages. —CodeCat 22:14, 13 January 2015 (UTC)

I support merging Finno-Volgaic and Finno-Ugric into Uralic. Ever since Tropylium pointed out the issue at Wiktionary talk:Families#Removing.2Fadding_families (in 2012), Finno-Ugric has been on a sticky note in the back of my mind as one of the places where our classification of linguistic families needed to be updated; kudos to you for pushing for action on the matter. Those unfamiliar with this language family can take a look at the first two sentences of w:Finno-Ugric languages and the first three short paragraphs of w:Finno-Volgaic languages, which sum up how out-of-date the linguistics behind those groupings is. As for Finno-Samic (and for that matter Finno-Volgaic): we never gave it a code in the first place, did we? So there's nothing to merge. - -sche (discuss) 04:22, 14 January 2015 (UTC)

I am obviously enough in support of this procedure. Though as noted at Wiktionary talk:About Proto-Uralic, it might be a decent idea to not completely merge "Finno-Ugric" and "Finno-Permic", but to relabel them dialects of Proto-Uralic instead.

User:Liliana-60 also mentioned back at Wiktionary talk:Families that aside from etymological appendix work, families are useful for users to find languages. If a user wants to find information about the other Finno-Ugric languages, can we leave redirects so that they will end up at the right address? --Tropylium (talk) 22:34, 14 January 2015 (UTC)

Merge direction

A subtopic: in what direction should a merger be done? The term "Finno-Ugric" has a long history of being used not only for Uralic minus Samoyedic, but also less formally for the family as a whole. It is also much more commonly used than "Uralic": the term "Finno-Ugric languages" gets some 800k Ghits, "Uralic languages" only 80k. So it might be worth considering to abolish the label "Uralic" instead. This would however not allow a dialect status for Finno-Ugric. --Tropylium (talk) 22:34, 14 January 2015 (UTC)

If we use Uralic, then there is no ambiguity over what we mean. With Finno-Ugric that doesn't apply. So we should use Uralic I think. —CodeCat 01:11, 15 January 2015 (UTC)

Yeah, use Uralic and list "Finno-Ugric" as an alternate name. (As an aside, maybe we should revive the idea of splitting the names= field into canonical and alternate name fields.) - -sche (discuss) 03:34, 16 January 2015 (UTC)

If there are no objections or comments, I'll start merging soon. —CodeCat 22:38, 26 January 2015 (UTC)

I've now deleted the "fiu-pro" code from the main language modules, and moved it to the etymology language module. This means that it's no longer valid to link to a Proto-Finno-Ugric entry, but you can still specify it as a source language in an etymology (like "Late Latin" and similar). There will probably be a lot of module errors from entries that currently still link to Proto-Finno-Ugric terms, I'll be running a bot regularly to change these to link to Proto-Uralic. If you find one and can't stand to leave it for the bot, you can fix it yourself by replacing {{m|fiu-pro|...}} with {{m|urj-pro|...}}. —CodeCat 15:29, 28 January 2015 (UTC)

Shall we also merge Category:Terms derived from Finno-Ugric languages and Category:Terms derived from Finno-Permic languages (and related target-lang-specific ones) with Category:Terms derived from Uralic languages? --Tropylium (talk) 23:18, 28 January 2015 (UTC)

I kept the categories separate for now because I wasn't sure if we still wanted to allow these languages to be mentioned in etymologies, even if we treat them as Proto-Uralic dialects. —CodeCat 23:36, 28 January 2015 (UTC)

Aren't "etymology language" exceptions encoded separately from the language family tree (which still seems to feature fiu)? --Tropylium (talk) 05:44, 31 January 2015 (UTC)

Yes, but I think you're confusing two things. There are both the language fiu-pro and the family fiu. The language has been merged completely so it's now a dialect of urj-pro, but the family still exists. —CodeCat 14:19, 31 January 2015 (UTC)

That's exactly what I was saying: shouldn't we also merge the families? Or do you think this would take further discussion yet? --Tropylium (talk) 17:17, 31 January 2015 (UTC)

I have been, but it takes longer for the database to catch up because all the etymology categories need to be updated too. —CodeCat 18:38, 31 January 2015 (UTC)

Admin userboxes

I know that userboxes are usually not permitted, but why hasn't anyone made an admin userbox yet for Wiktionary? I think it'd be useful to be able to look at someone's userpage and know straight off the bat, without having to go through all the trouble of looking through the large list, to see if someone is an administrator, crat, rollbacker, sysop, or whatever. (All due respect, I refer to all 3, admins, crats, and sysops, all as admins, because its too complicated to differentiate). If this isn't a good idea, shouldn't we at least have a category for admin users that shows up immediately at the bottom of every admin's user page? Has there been a discussion on this yet, or a resolution? NativeCat ^{drop by and say Hi!} 00:14, 9 January 2015 (UTC)

I think it’s a good idea. — Ungoliant ^(falai) 00:16, 9 January 2015 (UTC)

I might not particularly mind them, but I worry it may encourage what on TOW is called hat collecting. Admin userboxes can be put on pages fraudulently (i.e. without user being an admin), though this is usually not a problem — the userbox may contain a link to a page which verifies that an account indeed has admin rights. On the flip side, however, (as I expect, at least) no admin will be forced to put those userboxes on their pages, so it may fail to achieve the result you have in mind. — Keφr 10:39, 9 January 2015 (UTC)

omg, like I suspected, people do not know about user scripts' existence :/

@NativeCat insert this line in your common.js

mw.loader.load('//en.wikipedia.org/w/index.php?title=User:PleaseStand/userinfo.js&action=raw&ctype=text/javascript');

--Dixtosa (talk) 13:31, 15 February 2015 (UTC)

I will be completely unavailable for the next week.

Please have the dictionary finished by the time I get back. This includes all the foreign languages. Cheers! bd2412 T 05:53, 14 January 2015 (UTC)

Haha, sorry, your wish cannot be fulfilled. NativeCat ^{drop by and say Hi!} 13:26, 14 January 2015 (UTC)

@NativeCat: Well, not with that kind of attitude. —Justin (koavf)❤T☮C☺M☯ 17:00, 14 January 2015 (UTC)

Haha, a dictionary can never be completely completed. Who disagrees? NativeCat ^{drop by and say Hi!} 21:27, 14 January 2015 (UTC)

(That's the joke.) --Catsidhe ^{(verba, facta)} 21:40, 14 January 2015 (UTC)

Jokes aside, targets for languages could and should be set for each language. How about 20,000 lemma entries for most frequent words for a language? Inflection (if necessary for a particular language), pronunciation and etymology being desired but not mandatory. If this can be achieved, then a language could be considered to have a good coverage in our dictionary. How many foreign languages would fit these arbitrary criteria at Wiktionary? Twenty to thirty? --Anatoli T. ^{(обсудить}/^вклад) 21:41, 14 January 2015 (UTC)

Of the languages I search, I find only our English and Latin content good enough to be my primary source. Spanish and German are close. Italian has a lot of entries but most of them are so vague, it is easier to try to make sense of monolingual dictionaries. I’m sure Finnish is on the same level as English and Latin because there are a lot of entries and all the ones I run across are good. — Ungoliant ^(falai) 22:12, 14 January 2015 (UTC)

Don't you wish Danish coverage could be as good as the languages Ungoliant listed? NativeCat ^{drop by and say Hi!} 23:29, 14 January 2015 (UTC)

I would do Danish, but Norwegian is taking up all of my available time (in between burying our dead cat today). Donnanz (talk) 23:44, 14 January 2015 (UTC)

Apart from the above, I would say the coverage of Russian, Chinese, Japanese, French, Portuguese, Serbo-Croatian, Dutch are also quite good, including definitions, probably Polish, Hungarian. Number of Korean lemmas is close 20,000. --Anatoli T. ^{(обсудить}/^вклад) 00:26, 15 January 2015 (UTC)

• @Ungoliant MMDCCLXIV: Latin isn't as good as you think. Our entries are great, although they tend not to be as good as other online resources, but Latin is lacking in so many translation tables that one can't use en.wikt for English to Latin. That's a long-term project for me to attack when I've more time (which may never happen). —Μετάknowledge^{discuss/deeds} 23:35, 14 January 2015 (UTC)

  English often disappoints me. We have good breadth of coverage but quality of definitions for many common words is poor or, at best, uneven. DCDuring TALK 23:39, 14 January 2015 (UTC)

  My main contact with Latin is through etymology, and since I began editing there have been only a handful of cases where a Latin etymon is entryless. But yeah, the English-to-Latin translation coverage is horrible. If you’re interested, I can generate one of these for Latin. — Ungoliant ^(falai) 23:46, 14 January 2015 (UTC)

I don't think any Wiktionary has a good enough coverage of Danish as it should. I'm currently working a lot at the Danish Wiktionary, and even it doesn't have nearly as many Danish words (or {insert language here} words) as it could, which is a little disappointing. Really this problem goes on for all Wiktionaries, and English Wiktionary is not the only Wiktionary that has this problem. However, I feel we'll never be complete. That's the point of a wiki, it is never complete, there is always new content to add/edit. NativeCat ^{drop by and say Hi!} 00:33, 15 January 2015 (UTC)

"Never complete" is a strong statement. What we need, is the dictionary to be useful and used. If a Wiktionary covers a large number of words by frequency in a good manner, then all fancy, archaic, interesting otherwise words can be added later. Many editors focus on entries they like, not the words that are really necessary by dictionary standards. Russian entries, for example (and a couple of other Slavic languages), have inflections for most words, which are missing in most published dictionaries, which is a clear advantage, since inflections are not straightforward and can't be easily construed from grammar references. Chinese, apart from Mandarin, now has thousands of transliterations/pronunciations for other topolects, which is also hard to get (separate topolects, including Mandarin may still not be able to compete with other dictionaries but our entries have not only pinyin but also IPA and zhuyin, other dictionaries don't have all three). Electronic dictionaries available mostly provide only Mandarin and Cantonese, with Cantonese having less coverage. There's, of course, room for a lot of improvements. Finnish is probably the best online dictionary available. --Anatoli T. ^{(обсудить}/^вклад) 00:50, 15 January 2015 (UTC)

Spanish adjective forms

Over the past two weeks, Wonderfool has created no less than thirteen sockpuppets (one of which has been used four times), acting as bots to create missing Spanish participle and adjective forms. As he shows no sign of giving up, and the blocks only serve to slow him down, it seems to me that we should try to take some sort of alternative action. I don't really want to get involved in edit warring, or, frankly, any discussion outside of my field of expertise, so I'm merely going to make a couple suggestions as to solving or mitigating this problem:

1. Create an abuse filter. I'd do this, but (a) I don't know if there's policy involved, (b) I don't know if I'm allowed, and (c) I don't know how to write one (I could figure it out, but someone would still have to check over it.)
2. Request and create a bot to add these entries, thereby removing the need for Wonderfool to do so. I don't especially want to take that much time to do this, but if this is deemed a good solution and nobody else will, I can. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 19:19, 14 January 2015 (UTC)

Special:AbuseFilter/22 already exists for that purpose. Remember not to make the filter too broad. — Keφr 19:28, 14 January 2015 (UTC)

Ah! I'd missed it, thanks. I'd assumed such a filter would be closer to the end of the list, though... has he done this before? ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 20:09, 14 January 2015 (UTC)

Okay, I've made some modifications. Please do check over the filter if you can. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 15:19, 15 January 2015 (UTC)

• Hi. It would be super if someone ran a bot to create these forms, which should free me up some time for more useful stuff around here. I did find a handful of errors in this category and this one, which I've corrected, so ideally the bot should be run be someone with Spanish knowledge. User:Adrian F4/Bot code has some forms which can be used by a bot account, which all have been checked. Regards - WF. --Adrian F4 (talk) 21:12, 15 January 2015 (UTC)

  Until then, I'll take charge of adding any forms -WF. --Adrian F4 (talk) 21:14, 15 January 2015 (UTC)

  BTW, the filter you made doesn't really work. I can still create the pages. Perhaps you could make it stricter, or just forget about it. -WF --Adrian F4 (talk) 21:18, 15 January 2015 (UTC)

  Another suggestion is giving me a flood flag, like DCDuring kindly did, and I can add loads of good entries without bothering the RC patrol. Up to you, really. -WF --Adrian F4 (talk) 21:19, 15 January 2015 (UTC)

  Yes, but it's a terrible suggestion. Renard Migrant (talk) 22:58, 15 January 2015 (UTC)

New code for Norman

Today I discovered that there is a new ISO 639-3 code, valid for only a few days, nrf, for Jèrriais roa-jer plus Guernésiais roa-grn. Would it be straightforward and uncontroversial for us to get a bot to merge those two ad hoc codes into the new, official one? The bot should also make liberal use of {{label}}s so that we can still categorize J and G separately, much as we categorize the B, C, M, and S lects of Serbo-Croatian.

Also, a few months ago, Chuck Entz brought up here the issue of Sercquiais, the langue d'oïl variety spoken on Sark in the Channel Islands. Ungoliant and I supported the idea of merging all the Channel Islands varieties into Norman roa-nor, but no one else commented and the discussion fizzled out.

So I'm raising the issue again, but this time for the new code nrf. Can we decide to apply nrf to all Norman varieties, rather than just for Jèrriais+Guernésiais, as ISO has it? This would have not only the obvious advantage of collecting several dialects of what really is a single language into a single code, but would also allow us to use "Norman" as the canonical name for nrf instead having to come up with something like "Jèrriais–Guernésiais" or "Channel Islands Norman" or something.

Is this something that needs to be voted on?

Pinging NativeCat, Chuck Entz, Ungoliant, and Embryomystic. —Aɴɢʀ (talk) 22:08, 15 January 2015 (UTC)

Support merging Jèrriais, Guernésiais and Norman, but it probably should have a vote. — Ungoliant ^(falai) 22:39, 15 January 2015 (UTC)

I'll admit to being a little attached to the idea of the different Norman varieties as separate things, but considering that they already have a shared Wikipedia (under nrm), it certainly does make sense. And between Jèrriais and Guernésiais, especially, there is a lot of duplication that could be trimmed down. I've not been able to find as much info on Continental Norman or Sercquiais, but I'm fairly certain there's duplication with the former, perhaps not so much the latter (with its quite distinct orthography). embryomystic (talk) 22:43, 15 January 2015 (UTC)

If we merge them, then we need to change the mapping to nrm in Module:wikimedia languages. —CodeCat 22:49, 15 January 2015 (UTC)

Better yet, Wikimedia needs to move its nrm projects to nrf. —Aɴɢʀ (talk) 23:05, 15 January 2015 (UTC)

I support this. It makes it less confusing. If it's just a mere dialect, not even considered a language in itself, we should not have it as a header, IMO. We should do like we do with Serbo-Croatian and add words pertaining only to a specific dialect using Template:context. NativeCat ^{drop by and say Hi!} 23:08, 15 January 2015 (UTC)

Support merging the varieties of Norman into nrf. The fact that there's one unified Norman Wikipedia is telling. - -sche (discuss) 08:40, 17 January 2015 (UTC)

As someone with Norman ancestry, I support merging the varieties of Norman; they are dialects and not separate languages. The dialects often have spelling differences. But we don't consider American English a separate language just because it uses color rather than colour. Taxman1913 (talk) 17:10, 21 January 2015 (UTC)

• So can we consider this to have consensus, or do I have to start a vote before it will actually happen? —Aɴɢʀ (talk) 22:11, 24 January 2015 (UTC)

  Since there's only support, no opposition, and since the dialects in question were never recognized by even the generous/lax folks at the ISO as languages, anyway (we had to give them exceptional codes), I say you could just start merging them. FWIW, my standard is to set up votes only for cases that discussion (either on-wiki or in the real world) has shown to be contentious either linguistically or politically (like Moldovan→Romanian). And this doesn't appear to be contentious. Let me know if you need help merging them. - -sche (discuss) 22:57, 24 January 2015 (UTC)

  As a first step, I've added nrf as "Norman" (with all the dialects as alt names). - -sche (discuss) 23:11, 24 January 2015 (UTC)

  Short of going through all of Category:Guernésiais language, Category:Jèrriais language, and Category:Norman language manually, I don't know what to do to merge them. I'm not a bot operator, nor would I even know how to begin programming a bot. —Aɴɢʀ (talk) 12:49, 25 January 2015 (UTC)

I'm not sure a bot would be a good idea, since we have multiple language sections that need to be merged, with the potential for differences in the content between them. I'm sure a great deal of the content is going to substantially overlap, but where there are differences, we need to think about how to represent them. We also need to preserve the dialectical information (have we added "Guernésiais" and "Jèrriais" to Module:labels/data so we can use context labels?). Chuck Entz (talk) 16:11, 25 January 2015 (UTC)

Module:labels/data already includes "Guernsey" and "Jersey", so we can use those. It will generate the names "Guernsey Norman" and "Jersey Norman", which are probably easier to understand than Guernésiais and Jèrriais anyway. Without a bot, it will take for freakin' ever, since Jèrriais alone has over 9000 entries. —Aɴɢʀ (talk) 18:41, 25 January 2015 (UTC)

Proposal for bot edits to French

Two proposals:

1. Remove gender templates from adjective forms which have gender indication in the definition. For example, {{masculine plural of|word|lang=fr}}.
2. Change m-p and f-p to m and f inside {{head|fr|plural}}. So {{head|fr|plural|g=m-p}} becomes {{head|fr|plural|g=m}}. Rationale: 'plural of' is already in the definition.

Objections? Renard Migrant (talk) 18:17, 16 January 2015 (UTC)

I don't object. However, the proposal would be easier to assess if you gave us two example diffs or if you gave links to two entries that would be affected, so that we can check the current situation. Not as means of objection but rather from curiosity: is this a fun activity for you? Or do you think there will be tangible benefits for the user of the dictionary, more tangible than your adding new entries? --Dan Polansky (talk) 09:40, 17 January 2015 (UTC)

Regroupements for a noun example. Renard Migrant (talk) 17:43, 17 January 2015 (UTC)

Blanches adjective example. Renard Migrant (talk) 17:47, 17 January 2015 (UTC)

Support. I would also support removing the gender from non-lemma form headwords altogether, as it reduces duplication and makes maintenance easier. Of course, genders aren't modified that often, but they still can be, and nobody is going to remember to change all the form-of entries as well, especially if there are many of them. —CodeCat 18:01, 17 January 2015 (UTC)

I would strongly oppose that for inflected forms that have inherent gender, like maisons (French) which is inherently feminine. I think it's misleading to remove the gender, because it might look like French plurals don't have gender, furthermore it's use-unfriendly because the user has to click on the singular to get the gender. I feel like you're proposing to replace a good system with an inferior system, so I oppose it. Renard Migrant (talk) 19:12, 17 January 2015 (UTC)

Codes the ISO deleted or added in 2014

In 2014, the ISO deleted some codes and added others. Here are the changes they made and my thoughts on whether we should follow suit. If you have any comments of your own, please comment.

Deleted codes: ymt (→mtm), thx (→oyb), xsj (→suj), aue (→ktz), ime, sgo, dzd, bmy, bxx, byy, kbf, ynh.

1. Merged ymt "Mator-Taygi-Karagas" into mtm "Mator". I suggest we follow suit. Mator proper, Taygi and Karagas were three similar dialects, all extinct since the 1800s, of the Mator language; even if we wanted to distinguish them (which I don't recommend), it wouldn't help to have one code for the language and one code for all-3-dialects-at-once.
2. Merged thx "The" into oyb "Oy". I'd follow suit. Oy (Oi), The, Sapuan, Cheng/Jeng and Sok form a dialect cluster of which Oi is the main variety. According to Paul Sidwell, The and Oi are spoken in villages located within sight of each other and they are 97% similar lexically.
3. Merged xsj "Subi" into suj "Shubi". The former was just a duplicate of the latter; we should delete it, too.
4. Merged aue into ktz. Both are varieties of ǃKung (khi-kun). In the past, we decided to merge several varieties of ǃKung and only treat the macrolanguage as a language. Let's follow the ISO in merging aue into ktz. (And then maybe discuss whether or not to merge ktz into khi-kun...)
5. Deleted ime "Imeraguen" as spurious. We should follow suit. To quote the code-change form, the code was "presumably based on a blanket statement by Gerteiny (1967:56) [...] that it is "Reportedly a variety of Hassaniyya structured on an Azer (Soninke) base". However, it is unclear what Gerteiny's statement is based on, and no data or other evidence for the Azer-Hassaaniyya mix is adduced. Other ethnographers with field experience, notably Fortier (2004), say they speak Hassaaniyya with some special vocabulary related to fishing. The specific forms of the fishing vocabulary are cited, and they clearly derive from a Berber (not Azer) source. [...] There is insufficient evidence to posit a past or present separate language for the Imeraguen."
6. Deleted sgo "Songa" as spurious. We should follow suit, since there is, as the ISO code-change request form notes, "no sign of a language with that name in" the area where it was claimed to exist, though there are a number of things someone could have mistaken for it. (The afore-mentioned form suggests the code was meant to represent a KiSanga variety of lea, while Ethnologue theorizes someone mistook the name of the byi "Buyu"-speaking Songa tribe as a language name.)
7. Deleted dzd "Daza". I'd follow suit. Despite Ethnologue's comment that it is "distinct from Dazaga dzg", the ISO found no evidence that any lect besides dzg existed with such a name.
8. Deleted bmy "Bemba", saying "it is difficult to interpret this entry as anything other than a duplication of the Bembe [bmb] entry [...]. The population of Bemba [bmy] is given as 296 000 which is a plausible figure if it refers to Bembe [bmb], and there is absolutely no sign of a different Bemba language of such a large population in Kivu province [...] The alternative name Kinyabemba for Bemba [bmy] suggests that the information on Bemba [bmy] came from transient Bembe in Rwanda and therefore the equality with Bembe [bmb] may have been overlooked."
9. Deleted bxx "Borna" as spurious. We don't need to do anything, since we never got around to granting it a code in the first place, on account of the facts that its primary and alternate names all presented language conflicts with other lects which were already encoded, and it was noted to be obviously spurious.
10. Deleted byy "Buya"; it was just an error (potentially a mere typo) for byi "Buyu". Obviously we should follow the ISO in deleting it.
11. Deleted kbf "Kakauhua" as spurious. It was ostensibly an unattested extinct language.
12. Deleted ynh "Yangho" as spurious. It was a phantom language mentioned by a missionary but not found by linguists.

Added codes: jje, rts, pgd, cbq, syx, nxo, txj, gku, iks and two discussed elsewhere.

1. jje for Jeju (they call it "Jejueo"), which we have until now coded as qfa-kor-jjm. It should be straightforward for us to make the switch.
2. rts for Yurats. Wikipedia says "Yurats was probably either a transitional variety connecting the Nenets and Enets languages of the Samoyedic family, or an archaic dialect of Enets", so I would hold off on adding it until a need for it is documented.
3. pgd for Gandhari ("Gāndhārī"), one of the northwestern prakrits.
4. nrf, which has its own discussion section, which see: WT:Beer parlour/2015/January#New_code_for_Norman.
5. cbq for "Cuba" / "Tsucuba" (neither name occurs in the literature I can find; "Tsutsuba" occurs, though not in English). The code was added based on a first-hand account documenting "clear lexical differences between" Cuba and its neighbour, Lopa (lop), and documenting that "the noun class prefixes are also very different".
6. syx for Shamay (Osamayi, Samayi, Samay, Shamayi), a Bantu language.
7. nxo for Ndambomo, another Bantu language.
8. txj for Tarjumo, a liturgical language of Nigeria which descended from but is confirmed by several sources to be unintelligible with Kanembu (kbl). It is written in the Arabic and (in Western scholarly works) Latin scripts. (One source claims that it is also called Turjiman or Turjiman Kanembu, both meaning "interpreter", because it was the language of many works that explained Quranic verses; however, I can't find actual uses of either of those names.)
9. kjv, which has its own discussion section, which see: WT:Beer_parlour/2015/January#Literary_Kajkavian_Serbo-Croatian.
10. gku for "ǂUngkue"/"ǂKunkwe", which, together with ǁKuǁe, formed the now-extinct language Seroa (kqu). (When it added gku, the ISO restricted kqu to ǁKuǁe.) I'm not sure whether we should add it or not.
11. iks for Inuit Sign Language.

- -sche (discuss) 08:59, 17 January 2015 (UTC)

There are some I wanted to request the deletion of, but I can't find a 'contact us' page. Renard Migrant (talk) 17:39, 17 January 2015 (UTC)

According to the SIL page on page on submitting change requests, you can fill out this form and e-mail it to the address listed here. The form will become part of the public record. - -sche (discuss) 22:04, 17 January 2015 (UTC)

Data point: the entire known corpus of Yurats amounts to < 150 isolated words. I feel that extinct languages that have only been attested at this level of detail would, in general, be better recorded as single pages the Appendix namespace than as entries in the Main namespace. Do we have any general policy on the CFIness of extinct and poorly recorded languages? --Tropylium (talk) 01:52, 18 January 2015 (UTC)

The most relevant point would be WT:LDL. But we do include quite a few languages that are very poorly attested, such as Mycenaean Greek, Proto-Norse, Crimean Gothic and Oscan. —CodeCat 01:56, 18 January 2015 (UTC)

In general, the criteria for inclusion actually make it easier to have extinct and poorly-documented languages in the main namespace than to have well-attested living languages there, by requiring more citations of the latter, heh. That's because (we wanted to make it easier to include extinct and poorly-documented languages, and) this site is geared towards covering attested natural languages in the main namespace. I think that makes some sense, too — have all such languages in one namespace. (We include a couple languages of which only a single word is attested.) But for languages with few words, we could certainly also have an Appendix: or Index: for it.
In this particular case, however, because it's not at all clear that Yurats it actually a distinct language, I recommended we not give it a code, but we could make an appendix of it. - -sche (discuss) 03:01, 18 January 2015 (UTC)

Update: I've deleted all of the codes the ISO deleted except aue, and I've added all of the codes they added except cbq, gku and the code noted above as being excluded (rts); aue, cbq and gku I simply haven't gotten around to dealing with yet (aue and gku since they're slightly messier than I anticipated, and cbq because I'd like to figure out which of its names is most common, but none of them seem to be attested at all). - -sche (discuss) 22:12, 30 January 2015 (UTC)

Literary Kajkavian Serbo-Croatian

A few days ago, the ISO added a code for literary Kajkavian Serbo-Croatian: kjv, defined as denoting specifically the 16th to 19th century literary language and not modern Kajkavian dialects. What should we do? We could decline to follow suit, and continue to handle Kajkavian as sh; or we could add kjv to Module:languages and start having ==Kajkavian== entries; or we could add it to Module:etymology language so it could be cited in etymologies but wouldn't get its own L2. Or we could repurpose the code to refer to modern Kajkavian (the same way we repurposed ltc from "Late Middle Chinese" to "Middle Chinese") and then do one of the last two things. Or...
Pinging Serbo-Croatian speakers User:Ivan Štambuk, User:Biblbroks, User:Dijan.
- -sche (discuss) 09:41, 17 January 2015 (UTC)

There is no difference between "Kajkavian literary language" and modern Kajkavian dialects (spoken and written), apart from orthographic and typographic conventions used. By the 16th century the development of all SC dialects was pretty much over (there are few changes in case desinence usages, but that's it). At any case, this should be discussed if and when the number of Kajkavian entries and their categorization and periodization becomes an issue, and not before. --Ivan Štambuk (talk) 10:03, 17 January 2015 (UTC)

The goal of ISO 639 is apparently not one language, one code, but just to have codes for whatever is useful, even if a language gets represent twice, three times or whatever. Renard Migrant (talk) 19:15, 17 January 2015 (UTC)

I read your comment as "don't add kjv (at this time)". That's fine by me. (As for categorization, we do have Category:Kajkavian Serbo-Croatian.) To ask a more specific question, is Kajkavian mentioned often enough in etymologies that it would be useful to have an etymology code for it, the way we have frc for "Cajun French" (Category:English terms derived from Cajun French)? - -sche (discuss) 21:54, 17 January 2015 (UTC)

For my point, very good example. For ISO 639, the fact that Cajun French has a code does not imply that it's not the same as French. It has a code because such a code could prove useful. Renard Migrant (talk) 21:57, 17 January 2015 (UTC)

Kajkavian, Chakavian and Shtokavian are the three main varieties making up Serbo-Croatian, where Shtokavian is the one used as the base for writing. Linguistics topics refer to the other two varieties often enough, but I don't know if there are any specific phonological isoglosses that would separate them from the larger whole. In fact, the distinction is based on a vocabulary isogloss, the word for "what": kaj, ča, što. —CodeCat 22:02, 17 January 2015 (UTC)

@-sche: Your powers of observation serve you well - the point was indeed that we shouldn't put the cart before the horse and make lengthy discussions and arrangements for activity that might as well never materialize. Regarding the the Kajkavian borrowings into the the literary form - we're dealing with perhaps at best a couple of hundred terms in the entire language, mostly in regional usage and I would be hard-pressed to find attestations for them outside Kajkavian works. Kajkavianisms and other words from subliterary dialects were mercilessly weeded out from the standard language (along with LWs from Turkish etc.) by Vukovians during the standardization in the 19th century. (It's a Balkans thing - keep your language as pure as your society through occasional cullings of the unfit). If anything, there is a need for deriving the other way around - literary SC borrowings into Kajkavian (which would be kind of nonsensical to name - Kajkavian Serbo-Croatian terms derived from Serbo-Croatian). Most of the kaj words were added by User:Fejstkajkafski who is a dialectologist (I think), but they don't seem to edit anymore. When it becomes a necessity we can create codes for Kajkavian as a whole and/or its subdialects (which are many, they differ a lot, are not always mutually intelligible and are today for the most part written in scholarly transcription by linguists who study it since native speakers always utilize the standard language in writing). --Ivan Štambuk (talk) 22:27, 17 January 2015 (UTC)

This reminds me of the language ISO is pleased to call "Hiberno-Scottish Gaelic", a term they apparently made up, which has been given the code ghc. Basically, it's a cover term for Early Modern Irish and Early Modern Scottish Gaelic and is hardly more different from the modern varieties (abstracting away from the Irish spelling reform) than Early Modern English is from modern English. We do not recognize ghc as a separate language, but simply treat 13th- to 17th-century Irish as ga and 13th- to 18th-century Scottish Gaelic as gd. I'm inclined to believe Ivan when he says the differences between "Literary Kajkavian" and modern dialectal Kajkavian are more orthographic than linguistic, and to oppose recognition of the kjv code, at least for the time being. —Aɴɢʀ (talk) 10:33, 19 January 2015 (UTC)

OK, I have updated WT:LANGTREAT to note that we treat kjv under sh’s code and header. - -sche (discuss) 20:44, 21 January 2015 (UTC)

Redundant Austronesian entities

I took a look at our Austronesian categorization system, and there seem to be some messy things going on. Some initial observations:

• Western Malayo-Polynesian (pqw). Generally agreed to be a negatively defined areal group, cf. wikipedia:Western Malayo-Polynesian languages. Does not even have any children listed, but someone regardless seems to have gone on a spree to copypaste etymologies for durian in a zillion different languages, with a "Proto-WMP" intermediate form cited, instantly well-populating Category:Terms derived from Proto-Western Malayo-Polynesian. I propose this mess be merged with Malayo-Polynesian (poz).
• Central Malayo-Polynesian (plf). Likewise an areal group, cf. wikipedia:Central-Eastern Malayo-Polynesian languages. Includes several languages classified under this group, however. I propose merging this with the slighly better-defined Central-Eastern Malayo-Polynesian (poz-cet), although we currently seem to not have this as a categorization level; only the protolang.
• Proto-Eastern Polynesian (poz-pep-pro) and Proto-Nuclear Polynesian (poz-pnp-pro) exist for some reason, although the corresponding general categories (East, Nuclear) do not. The only reconstructed entries that currently seem to exist in these are *aka 'root', *wai 'water' and Eastern *maŋo 'shark', the first two of which both exist as Proto-Polynesian items as well. I propose merging these languages with Proto-Polynesian (poz-pol-pro). Perhaps however as dialects, given that these groups are considered to form valid clades.

--Tropylium (talk) 02:38, 18 January 2015 (UTC)

If you're surprised that Proto-Eastern Polynesian exists and has a code, just know that until recently it existed twice and had two codes. (See this RFM.)
Anyway, pinging User:Amir Hamzah 2008 and User:Metaknowledge, who are the main editors of our Polynesian proto-language entries.
- -sche (discuss) 03:19, 18 January 2015 (UTC)

I agree with you on deleting Western Malayo-Polynesian and merging Central Malayo-Polynesian to Central-Eastern. - -sche (discuss) 03:19, 18 January 2015 (UTC)

The non-Polynesian mess is due to the easy availability of Blust's Austronesian Comparative Dictionary online. I refer to it a lot because it has a great deal of useful data, but you have to be aware of the author's biases, and you also have to be aware that he regularizes the orthography of the reflexes to make it easier to compare between languages. That source doesn't provide proto-forms for Polynesian, but they're available at the Combined Hawaiian Dictionary and Pollex Online. There are actually enough differences between PPN and PEP for the latter to be worth maintaining, even if we don't have much of anything in it as yet. I haven't really looked at the differences between Tongic and Nuclear Polynesian enough (or at least recently enough to remember anything), so I can't say much about it. If anything, I would add Proto-Central Eastern Polynesian, rather than taking away PEP. Chuck Entz (talk) 05:59, 18 January 2015 (UTC)

"There are actually enough differences between PPN and PEP for the latter to be worth maintaining". Yes, I'm sure there are differences — but we don't usually create separate proto-stages just to highlight the phonetic evolution of a word (for example, early PIE *h₂érh₃trom 'plough' is considered sufficient, and we do not create a "Late PIE" entry *árə₃trom, or a "Mid PIE" entry *h₂árh₃ətrom). If you wanted to be systematic about this type of a thing, it leads to tons of unmaintainable duplication, where approximately every Proto-Pol. entry has a corresponding Proto-EP entry — and then easily a dozen other nodes downward towards Proto-Austronesian. Yet it seems to me that the differences are not quite enough to make the two proto-stages more than two dialects of the same languages, so we ought to be able to cover this type of situation well enough by noting things like "evolution s > h is Eastern Polynesian", or labeling Eastern lexical innovations as dialectal Proto-Pol.

Core point being that etymological appendices do not exist for the purpose of documenting various reconstructible proto-language stages; they exist for demonstrating the relationships of attested languages. When chronology is known in great detail, this seems to necessarily lead to having to treat closely successive stages on single pages. --Tropylium (talk) 20:17, 18 January 2015 (UTC)

These are not different stages of PIE but different notation systems - underlying (aka (morpho)phonological or etymological) vs. surface (aka phonetic, i.e. the one you get with comparative method). The established notation is however inconsistent, though some recent works tend to use slashes and square brackets to distinguish the two. E.g. our entry *bʰréh₂tēr is technically nonsensical, it should be either */bʰréh₂ters/ or *[bʰráh₂tēr] (or *[brā́ter] if you one doesn't subscribe to the theory that the Balto-Slavic acute is of laryngeal origin..). No idea about Polynesian stages though, but if the separate clades are generally accepted than intermediary steps are OK to have, especially if there are restrictions in lexicon to various subbranches (e.g. West-Germanic only word should be reconstructed as Proto-West-Germanic and not Proto-Germanic). --Ivan Štambuk (talk) 00:21, 25 January 2015 (UTC)

What? Of course those are different stages. Laryngeal theory is precisely about (among other similar developments) the idea that (some cases of) Late PIE *[a-] historically developed from an earlier *[Ha-], which historically developed from earlier *[h₂a-], which historically developed from Early PIE *[h₂e-]. One can claim that the sound changes continued to hold an existence as phonological processes (i.e. that *[Ha-] would have still been phonologically */h₂e-/), but don't let this distract you. All phonological processes have a diachronic origin.

…Anyway, I'm switching to your West Germanic example, as something less prone to getting the discussion stuck in details. Of course we should reconstruct any exclusively West Germanic words as West Germanic and not as Proto-Germanic proper. The question is how should we notate this. In principle, we could create a separate Proto-West Germanic appendix. But I continue to think that this is a poor approach, since an entire separate appendix for a only marginally different proto-language stage entails massive duplication of work. We do not create pages like "Appendix:Proto-West-Germanic/balluz" to go with Appendix:Proto-Germanic/balluz, even though a word's existence in West Germanic languages and Proto-Germanic implies its existence in Proto-West Germanic. What we can do instead, what we already do, and what I am proposing should be done in similar cases as well (like here, Proto-Polynesian versus Proto-East Polynesian etc.), is that we can set up Proto-West Germanic as a dialect of Proto-Germanic, and file words only attested in West Germanic under Category:Regional Proto-Germanic.

Note again also the absense of entries just to mark phonetic development! People working on this area have been content to have just e.g. Appendix:Proto-Germanic/ēnu, not creating duplicate entries like "Appendix:Proto-Germanic/ānu" to demonstrate the Northwest Germanic development *ē > *ā. It's even entirely possible to be explicit about a thing like this regardless, the easiest way being to include a line such as

* Northwest Germanic: *ānu

in the list of descendants, and then indent all NW Germanic entries to depth-2. --Tropylium (talk) 21:52, 25 January 2015 (UTC)

Nope, they're synchronic really: */bʰréh₂ters/ → *[bʰráh₂tēr], not */bʰréh₂ters/ > *[bʰráh₂tēr]. It's the same thing just in two different notations. *a is just an allophone of *e colored by an adjacent *h₂. We just use the former notation to distinguish such *a from the "real" *a that is sometimes postulated. Nobody knows the ancestral form in Early and Mid PIE because there is nothing to compare them against, and or theories based on internal reconstruction are too speculative because the evidence is very scarce.

It's phonemically */bʰréh₂tēr/. Resolving -ēr into -ers is not phonemic, because there could in theory be another word for which the underlying phonemes are also */bʰréh₂tēr/, but which can't be morphologically resolved into -ers. The proper analysis is that the phonemic distinction between -ēr and -ers is neutralised. —CodeCat 23:45, 25 January 2015 (UTC)

If the distinction is neutralized it in principle doesn't matter which form you use, except in this case *ers# → *ēr# by Szemerényi, so we know that the "real" underlying phonological form for this specific word is */bʰréh₂ters/. It doesn't matter if there are other reconstructions with the same form at the surface level. Their existence or absence does not invalidate the underlying form of other words. --Ivan Štambuk (talk) 00:27, 26 January 2015 (UTC)

The differences between PWGmc and PGMc seems substantial though (shouldn't it be PWGmc *ballu really?), and when we have protolangs separated by centuries we can't really speak of dialects anymore. I suspect that there are many such "small" protolangs that should better be fitted into some larger grouping, but we don't really have space constraints and if there is a body of scholarship supporting them (not just the reconstructions, but the entire protolanguage, with inflections etc.) I see no point in forbidding them. They're the problem of those who add them. Not everything has to be perfectly structured, it's all a perpetual work in progress here. --Ivan Štambuk (talk) 23:15, 25 January 2015 (UTC)

Space constraints are a red herring. My concerns are internal consistency, and editor attention constraints on what can be reliably kept up to date. Or, more specifically: as our treatment of things like Northwest Germanic demonstrates, "having an intermediate proto-stage" and "having a separate appendix / separate language status for an intermediate proto-stage" are two different things. No one wants to stop recognizing Proto-Eastern Polynesian; I am only proposing covering it in the same appendix entries as corresponding Proto-Polynesian forms.

(This might also be the first time I hear "it is a work in progress" used as an argument against improving organization.)

Also I guess a more general discussion on preferrable ways to organize etymology appendices might be worthwhile. I think I'm going to start that below. --Tropylium (talk) 00:11, 26 January 2015 (UTC)

Actually, Proto-West-Germanic never existed in all likelihood, given the current thoughts in the field. West Germanic doesn't appear to be a clade. While it certainly has some innovations that are shared among the group, it seems that West Germanic was still in a continuum with North Germanic when those changes occurred. This is similar to the West and South Slavic language groups, for which a single proto-language stage is not reconstructable either. —CodeCat 22:07, 25 January 2015 (UTC)

The Ringe & Taylor 2014 book has two chapters on it. Situation with Slavic division is different - we're know for certain that they are geographical groupings and as speech communities they never existed because the earlier form of the language is already attested in terms of Old Church Slavonic, which already demonstrates dialectal variation.--Ivan Štambuk (talk) 23:15, 25 January 2015 (UTC)

• I'm certainly responsible for some of the mess; I haven't time to respond in full, but suffice it to say that entries in PNP or PEP are not that we may show every stage in the phonetic development of a proto-word but rather that we may have entries for words that by editorial laziness or simple lack of cognates we cannot place in PPn in good faith and yet are demonstrably valid up to a certain level in the Polynesian tree. —Μετάknowledge^{discuss/deeds} 21:17, 22 January 2015 (UTC)

• Returning to the subject at hand... Western Malayo-Polynesian and Central Malayo-Polynesian are negatively-defined areal rather than valid genetic groups, so proto-languages for them seem nonsensical. Is there any logical opposition to deleting the first one and merging the second to Central-Eastern? - -sche (discuss) 00:27, 26 January 2015 (UTC)
  • No argument from me. They're just the kind of artifacts you would expect from Blust's dependance on lexicostatistical cladistic analysis, so it's not surprising that he would provide lots of them- but that doesn't make them valid. Chuck Entz (talk) 01:45, 26 January 2015 (UTC)

Header levels

To the best of my knowledge, there is no 'official' list of header levels, WT:ELE mentions the headers, but not the levels. A particular one to start with, should Descendants be L3 when the descendants are not from a particular part of speech? I think the most official list we have is User:AutoFormat/Headers. Renard Migrant (talk) 11:55, 18 January 2015 (UTC)

What about DTLHS's list? —CodeCat 20:14, 20 January 2015 (UTC)

What makes something "official"? Shall we have a vote? My list also doesn't take into account the relative positions of headers. DTLHS (talk) 20:20, 20 January 2015 (UTC)

I'm not advocating officialness just asking the question. Renard Migrant (talk) 20:41, 20 January 2015 (UTC)

Proto-Arawak

Hi, could someone do me a favor and create a language module for Proto-Arawak?

m["arw-pro"] = {
names = {"Proto-Arawak", "Proto-Arawakan", "Proto-Maipuran"},
type = "regular",
scripts = {"Latn"},
family = "awd",
}

Thanks for your help! --Victar (talk) 17:49, 20 January 2015 (UTC)

I've created it, but as awd-pro, since the usual naming scheme for proto-languages is to add "-pro" to the family code. I also used "Proto-Arawakan" as the canonical name because "Arawakan" was the name of the family, but given that Proto-Arawakan and Proto-Arawak have about as common over the last few decades, and Arawak has always been more common than Arawakan, feel free to start an WT:RFM or ask me to if you think the family and protolanguage should be renamed. - -sche (discuss) 17:46, 21 January 2015 (UTC)

Thanks a ton -sche! --Victar (talk) 19:05, 21 January 2015 (UTC)

Taíno vs Taino

Hi, the proper way to spell Taíno in English is Taíno, not Taino (sans i-acute). Can we change its canonical name to reflect this? --Victar (talk) 18:34, 20 January 2015 (UTC)

It's not one person's decision what the 'proper' way to spelling a word is. Sure, we could change it if we collectively wanted to, but do we? Renard Migrant (talk) 20:42, 20 January 2015 (UTC)

Appendix:Taíno/kasike‎ may need moving and correcting. Renard Migrant (talk) 20:46, 20 January 2015 (UTC)

By proper, I mean the spelling used in all the vast majority of academic papers and on Wikipedia. So if an admin could change it before start adding more entries, I would appreciate it. --Victar (talk) 20:53, 20 January 2015 (UTC)

All academic papers? All of them, in the history of the English language, that's what you're saying isn't it? Renard Migrant (talk) 22:21, 20 January 2015 (UTC)

Yes, all in the history of man. ;-) But seriously, if it's the form used in academic papers and on Wikipedia, it should be the form used on Wiktionary. Do you have the capability to change it? --Victar (talk) 22:27, 20 January 2015 (UTC)

Wrong! I found one that uses Taino. Hard lines. Renard Migrant (talk) 22:33, 20 January 2015 (UTC)

OK, sure, but I can tell you that it is not the common spelling in academic papers, and not he spelling used on Wikipedia. --Victar (talk) 22:38, 20 January 2015 (UTC)

Check google ngrams. DTLHS (talk) 22:43, 20 January 2015 (UTC)

I don't know if that's the best method, since OCR software frequently misses /í/ and people often casually omit diacritical marks. Taino is also a Japanese surname. But if anything, the spelling should match the Wikipedia articles, w:Taíno, w:Taíno language. --Victar (talk) 22:58, 20 January 2015 (UTC)

Why? We are not Wikipedia’s vassals. — Ungoliant ^(falai) 23:04, 20 January 2015 (UTC)

That's true, but consistency is a good thing and if Wikipedia thought Taíno the preferred spelling, that says something in and of itself.. --Victar (talk) 23:13, 20 January 2015 (UTC)

I'll propose the renaming on Wikipedia then. Renard Migrant (talk) 12:30, 21 January 2015 (UTC)

... and the Ethnologue and Wiktionary thought Taino the preferred spelling. If we do change the name, it should be because the Wiktionary community found it better to use the new name, not because we have to kneel before Wikipedia. — Ungoliant ^(falai) 23:25, 20 January 2015 (UTC)

I don't know about Ethnologue and whether they make full use of diacritical marks, but Wiktionary had no Taíno entries until I just added three, so that's why it's being brought up now. Also, are we not the Wiktionary community having a discussion on this right at this very moment? --Victar (talk) 23:40, 20 January 2015 (UTC)

• While I grant that we intend (generally) to be a descriptive dictionary in how we create and edit our term entries (and as such, we should probably have entries for both spellings), in my subjective experience, we also seem to hew somewhat to academic norms when it comes to the terminology used for categories and other infrastructure, such as language names and codes. Avoiding ambiguity and all that. ‑‑ Eiríkr Útlendi │ Tala við mig 23:16, 20 January 2015 (UTC)

For reference, Languages of the Pre-Columbian Antilles is probably the main published work on Taíno terminology and will be often cited for entries. It too uses the spelling Taíno. --Victar (talk) 23:25, 20 January 2015 (UTC)

As is documented in wt:Languages, "Whenever possible, common English names of languages are used, and diacritics are avoided." Judging my ngrams and by the comments about academic literature, this may be a case similar to that of Maori (Māori). I'd stick with the diacritic-less name since it's common (well-attested) and easier to type. - -sche (discuss) 22:00, 22 January 2015 (UTC)

Well, like I mentioned above, ngrams isn't a good tool in this case so you have to really look at the current published academic literature to make an assessment. My impression is that Taíno is the far preferred spelling in papers on the topic. --Victar (talk) 23:20, 22 January 2015 (UTC)

I see you're point though with the example of Māori being the more "correct" spelling, but for the sake of ease in typing, Maori is preferable. I think the major difference though is Taíno is a reconstructed language, with a finite amount terms taken from academic papers, where as Māori is a living language with tens of thousands of terms. --Victar (talk) 02:46, 23 January 2015 (UTC)

JAnDbot

Discussion moved from WT:RFDO#JAnDbot.

Please, unblock user:JAnDbot. I am working on maintenaning interwiki links on (language) categories and I need to use it in enwikt too (now I must use my own account). Please note, that bot is blocked for more than 6 years and the reason is now obsolete. JAn Dudík (talk) 11:00, 21 January 2015 (UTC)

It's an illegal bot because it hasn't passed a bot vote. That's the reason. Renard Migrant (talk) 12:35, 21 January 2015 (UTC)

Bot did not passed a vote for bot flag, but there is no reason for leave it blockd for many years. I dont want to work on interwiki on articles, but on categories only. And I want to use my bot's account, beacuse of SUL (in the other case I could make new account, but it is problematic). There are many categories which have no interwiki even if exist in some other languages. JAn Dudík (talk) 14:05, 21 January 2015 (UTC)

No, the requirement for a vote has nothing to do with the bot flag, but has everything to do with whom we trust enough to allow the capability to perform high-volume, unsupervised edits. It doesn't matter what account you use- if you operate a bot without permission, you're subject to being blocked. Chuck Entz (talk) 14:28, 21 January 2015 (UTC)

The user is now operating his own account as a bot... - -sche (discuss) 22:24, 26 January 2015 (UTC)

I've blocked the user for a day, and think the bot account should stay blocked because of this. —CodeCat 22:27, 26 January 2015 (UTC)

Please, can you say me, how can I work on category interwiki? When I want to use bot you say 'no, it should stay blocked. When I use my own account, tou say he is using his account as a bot, block him. English wictionary is the biggest one, so even If I edit here only in must-edit-cases (incorrect interwiki, no interwiki, no other language with this interwiki, missing 5 or more links), there are many edits. There is 170 other wictionaries, where I work, and there are no problems with these edits. But english Wictionary is so selfcentric and selfish, that there say Problems with interwiki on other Wictionaries is not our problem. We dont want to have correct interwiki here. And some interwiki conflict? WTF? JAn Dudík (talk) 07:33, 28 January 2015 (UTC)

I would suggest starting a new bot vote and saying: These are the kind of interwiki links I want to add. Hopefully there will be more participation in the bot vote, and people will be able to look at your recent contributions and see if they're correct or not. I'm sorry we're such sticklers about bot policy... - -sche (discuss) 19:59, 28 January 2015 (UTC)

I've drafted Wiktionary:Votes/bt-2015-02/User:JAnDbot for bot status for you. - -sche (discuss) 22:33, 7 February 2015 (UTC)

• I have a quick question - how does your bot work? It's straight forward enough to see how an interwiki bot knows to link Tamil with no:Tamil and tr:Tamil, but how does it know to link Category:Tamil language with no:Kategori:Tamilsk and tr:Kategori:Tamilce. Do you use the Wiktionary translations to link categories, or is there something else you use (language codes?) Smurrayinchester (talk) 18:49, 8 February 2015 (UTC)

@: Articles are made autonomously, but work on categories means lot of manually work (if no interwiki is present, I must manually find and type correct category on some other project). JAn Dudík (talk) 21:17, 18 February 2015 (UTC)

The vote has ended, passes. JAnDbot is unblocked and flagged as a bot. —Stephen ^(Talk) 03:17, 26 February 2015 (UTC)

Voting to always supersede policy

I think we should seriously consider a vote that says voting always supersedes 'policy'. By this, individual pages (no matter what the namespace) and groups of pages will not be subject to any written policies if there is a vote on that individual entry. The prime example is Wiktionary:Votes/2014-11/Entries which do not meet CFI to be deleted even if there is a consensus to keep, voting comes before policy for deletion matters. The only reason I can think of not to vote on this issue is that it's equivalent to a vote on whether water's wet. It's already what we do. Renard Migrant (talk) 12:43, 21 January 2015 (UTC)

Consensuality and attestation

I have just and rather belatedly realised that the work I had last done yesterday evening had been undone on two grounds: lack of consensus and multiple quoting.

If consensus is not required for undoing but is required for doing plain work, and if it is not a general principle of Wikipedia, though it is elsewhere in dictionaries, that quotation is primarily for attestation while examples are primarily examples, then I'm giving Wikipedia away, not without very great chagrin.—ReidAA (talk) 21:31, 24 January 2015 (UTC)

Mari templates

Our templates for Mari language seem to be inconsistent with ISO 639-3 and even between themselves.

chm

External links section in the entry Mari links this to Mari the macro language, which is consistent with ISO, but {{etyl|chm}} links to Wikipedia article on Eastern Mari.

mhr

External links section in the entry Mari links this to Eastern Mari, which is consistent with ISO although the preferred name for the language seems to be Meadow Mari, but {{etyl|mhr}} returns an error message.

mrj

External links section in the entry Mari links this to Western Mari, which is consistent with ISO although the preferred name for the language seems to be Hill Mari, and {{etyl|mrj}} links to Wikipedia article on Mari language.

Could somebody familiar with the workings of our language templates fix this? I will write Wiktionary entries for Meadow Mari and Hill Mari. --Hekaheka (talk) 11:09, 25 January 2015 (UTC)

We had a discussion in WT:RFM a while ago. @-sche Could you help dig it up, please?

I don't find anything on Mari in WT:RFM, not in the active discussion nor in the archives. --Hekaheka (talk) 12:59, 25 January 2015 (UTC)

I'm sure it was in RFM. --Anatoli T. ^{(обсудить}/^вклад) 13:58, 25 January 2015 (UTC)

"chm" is the code for Mari and Eastern Mari, which have been merged to Eastern Mari, since Eastern Mari (Meadow Mari) is the standard language of Mari people. "mrj" is reserved for Western Mari (Hill Mari), which is considered a dialect and "mhr" is not used @Wiktionary. If "Mari" is used on its own, Eastern Mari is implied, there is no other (macro) Mari language. We decided to only use terms "Eastern Mari" (="Mari") and "Western Mari" here, not "Meadow Mari", "Hill Mari" or just "Mari". --Anatoli T. ^{(обсудить}/^вклад) 11:33, 25 January 2015 (UTC)

"chm" is the code for Mari and Eastern Mari, which have been merged to Eastern Mari -- I can see that we have done it. I just pointed out that we have it differently from ISO 639-3. Has this merger been done somewhere else outside the Wiktionary space or is this our own invention? Anyway, I maintain that I'm right in saying that our current practice is confusing. Perhaps a clarifying usage note under the entry "Mari" would do the trick. My Russian is elementary, but I understand from ru-Wikipedia that Hill Mari (Горномарийский язык) has some sort of official status in Mari El. Anatoli - can you elaborate a little on this point? --Hekaheka (talk) 12:59, 25 January 2015 (UTC)

Eastern Mari, Russian and Western Mari are three official languages in Mari El but Eastern Mari has 10 times more speakers than Western, is used wider and is often just called Mari (in whatever language). Despite separate code, there's no separate Mari language. Let me know if you want to know more. --Anatoli T. ^{(обсудить}/^вклад) 13:58, 25 January 2015 (UTC)

There have been two discussions of Mari, wt:Beer_parlour/2013/September#Merging_Mari_and_Buryat_varieties and wt:Language treatment/Discussions#Merging_Buryat_dialects.3B_also.2C_merging_Mari_dialects, which led to the status recorded on wt:LANGTREAT: "Both the macrolanguage and its subdivision mrj are treated as languages, but the macrolanguage code is used in place of the code (chm) which the ISO gave the standard variety of the language." I thought the use of the macrolanguage's code (chm) for Eastern Mari (properly mhr) was because most sources used chm for Eastern Mari text (à la what Anatoli is saying about how it "is often just called Mari"); if that's not the case, we could always retire chm and add mhr. But note that using a macrolanguage's code for the standard variety is something we've done before, e.g. we use lv for standard Latvian even though technically lv includes Latgalian, and standard Latvian alone would be lvs; likewise we use et and not ekk for Estonian. (And the decision to recognize either the dialects (under whatever codes) or the macrolanguage, but not both, is, well, to avoid redundancy.) - -sche (discuss) 17:15, 25 January 2015 (UTC)

All right, I think I now understand the way of thinking followed here. In the end the exact definitions of languages are often a matter of convention, so let's follow the one that has been reached. I have written a usage note to the entry for Mari and also otherwise edited it. As the terms "Hill Mari" and "Meadow Mari" are widely used, I think we should somehow explain how they relate to the whole. I made an attempt for that end, too. Please check what you think of the entry as it is now. --Hekaheka (talk) 19:06, 25 January 2015 (UTC)

Surely Wiktionary:About Eastern Mari and Wiktionary:About Western Mari (both currently nonexistent!) should be the primary places to record policy on the representation of Mari on WT? Although I observe that about pages are neglected for a lot of smaller languages. --Tropylium (talk) 22:03, 25 January 2015 (UTC)

Feel free to set up such pages. FWIW, though, WT:LANGTREAT is the central place to record mergers of dialects, splits of languages, and information like that Western Mari is mrj while Eastern Mari is chm. The about pages often concern themselves with orthography, transliteration, entry formatting, etc. - -sche (discuss) 22:06, 25 January 2015 (UTC)

General principles of protolang appendices

It seems that WT:PROTO and WT:ETYM do not actually go much into this topic at all. A new section in the former, or a separate think tank at Wiktionary:Proto-languages might be in order.

So, some guidelines I would like to have stated explicitly — or, in case it turns out other editors disagree, discussed explicitly:

1. Appendix pages are not dictionary entries, and are hence not automatically subject to WT:ELE.
   • This means e.g. that we are not obliged to create different entries for different, but related proto-words. In principle an etymological proto-root appendix may be only a single stem, and the different morphological variants indicated by the different descendants discussed in prose instead.
   • Obviously many etymology appendices regardless are effectively about a single reconstructed word. I agree with the current statement at WT:PROTO that in this case,

     (…) the layout of the entries generally conforms to WT:ELE, although some compromises may be made for the sake of usability.

2. Documenting a proto-language is not the main motivation for having etymology appendices in the first place: they instead exist primarily to highlight the etymological relationships of attested words.
   • All proto-language roots should have at least one existing Wiktionary entry listed as a descendant.
   • Synonymous proto-forms differing only in e.g. declension class should be preferrably discussed on a single page, not split across several "appendix entries".
   • Likewise, even if a subclade with a slightly different proto-language can be established, it should be by default treated together with its parent root. If no parent root of a particular item is known, it should be filed under "Regional Proto-Foo".
     • Which does not mean a blanket ban on establishing closely-spaced protolang levels, of course, if reasons for creating some were to regardless exist.
3. Roots in intermediate proto-languages which have a well-known parent — such as Proto-Germanic, Proto-Indo-Iranian, Proto-Finnic, Proto-Oceanic — should be provided with either an inherited or loan etymology, marked as words of unknown origin, or tagged with a request for etymology.
   • Not all bottom-level proto-languages have been well-reconstructed, though. E.g. no consensus exists on what Proto-Afro-Asiatic looked like, and so entries in e.g. Proto-Berber or Proto-Semitic would probably not benefit from mass-tagging with requests for etymology just because a reference to cognates elsewhere in AA has not been found.
   • We currently also have a single entry in Category:Proto-Indo-European terms with unknown etymologies for some reason.
   • Similar to WT:WDL, establishing a list of well-reconstructed proto-languages might be useful to have for reference. Not for enforcing new standards on the languages on it — but as a warning sign to anyone setting out to create appendix pages dealing with proto-languages that don't have any "standard" reconstruction.

--Tropylium (talk) 01:50, 26 January 2015 (UTC)

As for Category:Proto-Indo-European terms with unknown etymologies, I just reverted the edit responsible. The etymology said "Origin unknown", and someone couldn't resist the temptation to replace that with {{unk.}}, which is kind of silly for a top-level proto-language. Chuck Entz (talk) 02:23, 26 January 2015 (UTC)

Is documenting all Unicode characters within the scope of Wiktionary?

We have a lot of information about a wide variety of Unicode characters. In many cases, there is a Translingual section just to give a "definition" for some obscure symbol. I have my doubts about whether this falls within the scope of Wiktionary. We're a database on languages, words and their meanings, but arbitrary symbols aren't necessarily used in any of those. So I think we should reconsider this, and set some specific criteria on what symbols to include. Either that or we should explicitly state that Wiktionary is a Unicode database. —CodeCat 20:15, 26 January 2015 (UTC)

I do not think any special criteria are necessary for symbols: attestation in natural-language texts should suffice. I would go further, actually, and establish a principle of describing characters rather than code points. For one, I do not think that fullwidth forms of Latin letters deserve separate entries. — Keφr 20:34, 26 January 2015 (UTC)

At least we managed to reach agreement to use regular Latin letters rather than fullwidth letters in the names of entries like CD and CD机.
I could get behind redirecting the individual fullwidth letters to their regular-width counterparts.
- -sche (discuss) 21:37, 26 January 2015 (UTC)

I'm inclined to be more restrictive than that. Symbols should only be allowed if they can be used as entities representing words, not as entities representing themselves. That is, they are used as representing something other than the symbol. That would mean "❑" or "漢" would not be allowed for English, because I imagine the only occasion where you could find them in running text is as mentions, in the sense that the symbol stands for itself as an entity. It's not used as a word to stand for something else. —CodeCat 22:09, 26 January 2015 (UTC)

Our usual attestation criterion requires "conveying meaning", so I took this as implied. — Keφr 22:39, 26 January 2015 (UTC)

Yes, but the outcome of a recent vote is that consensus can make CFI mean whatever we want it to mean. —CodeCat 22:42, 26 January 2015 (UTC)

Some people may find translingual sections useful but I don't. If character information is to be kept - stroke orders, input methods, references, such as links to Unihan database, they are better kept outside Chinese sections, so this is a pro-argument.

Chinese (single) characters often convey basic meanings, without any connection to a part of speech. That's why translingual sections for Chinese characters lacked PoS. The actual PoS is determined by the usage in a phrase but there could be multiple interpretations and no consistency. There is no inherent PoS in a Chinese word. Published Chinese dictionaries sometimes use PoS info, sometimes don't and they have a lot of mismatches. We should allow ===Definitions=== header, at least for single characters. Definitions from Translingual should be moved to appropriate language sections - Chinese, Japanese, Korean, Vietnamese where appropriate. --Anatoli T. ^{(обсудить}/^вклад) 23:11, 26 January 2015 (UTC)

This discussion is more about characters like U+1F6A6 VERTICAL TRAFFIC LIGHT than 漢字. — Keφr 23:28, 26 January 2015 (UTC)

Well, Chinese characters were mentioned and the presence of a definition line, that's why I commented. The definitions in the Translingual sections may be useful until the moment they are moved (checking required) into language sections. --Anatoli T. ^{(обсудить}/^вклад) 23:43, 26 January 2015 (UTC)

This is a completely different issue. This is about whether arbitrary Unicode characters warrant entries. 漢字 were only used as an example of characters that may be mentioned, instead of used for their meaning. — Keφr 10:08, 27 January 2015 (UTC)

Yes. Chinese characters convey meaning in the languages where they are used, and don't merely stand for themselves. So they can be included for those languages. —CodeCat 23:34, 26 January 2015 (UTC)

Let's keep the meaningless Unicode to appendices, please. We have them for whoever's looking, but we need not have entries for them. bd2412 T 00:39, 27 January 2015 (UTC)

I propose to make entries of characters whose only "definition" is their Unicode character name speedy-deletable. Does this warrant a full vote? — Keφr 10:08, 27 January 2015 (UTC)

Yes, because I'd oppose. I think entries of characters whose only definition is their Unicode character name should be hard redirects to whatever Appendix lists them, rather than being redlinks. —Aɴɢʀ (talk) 13:03, 27 January 2015 (UTC)

In the spirit of compromise, I propose to speedily redirect entries of characters whose only "definition" is their Unicode character to whatever Appendix lists them. bd2412 T 13:52, 27 January 2015 (UTC)

Yes, I think that's the best solution. If we delete, someone may be tempted to recreate the entries. Chuck Entz (talk) 15:10, 27 January 2015 (UTC)

On the other hand, redirecting might discourage creation of entries with legitimate definitions. For one, [[ℸ]] currently has no attested definition, but it does not mean that it never will. [[⨇]] currently does not exist, but I have seen the symbol used as a logical AND operator over arbitrary sets: I have a really hard time finding good attestation for it, though. Redlinks make it explicit that there is no entry for a given character; redirecting gives the impression that further content is not needed, even superfluous — even though a determined editor can create an entry over the redirect. Secondly, readers may take the Unicode character name as a definition — at which it often does poorly, as I noted in the RFD to be archived at Talk:⦰. And lastly, I just dislike cross-namespace redirections. — Keφr 21:53, 29 January 2015 (UTC)

You, Aɴɢʀ and bd2412 specified "characters whose only definition is their Unicode character name", though? This seems to primarily imply "decorative" Unicode blocks like "miscellaneous symbols", "box drawings", "arrows". If there is reason to suspect that there is a more specific definition in existence — as is the case for all orthographic and mathematical characters, for starters — no redirect should be created, IMO. --Tropylium (talk) 22:14, 29 January 2015 (UTC)

Nothing in this proposal, as written, implies it. There are plenty of mathematical and other technical characters with exactly this kind of non-definition (say, [[⧧]] or [[⧝]]), but this is not yet an indication that this is the only possible one. — Keφr 13:33, 30 January 2015 (UTC)

Sorry, poor L2 use of "implies" here, I suppose. If an alleged mathematical symbol fails to be attested in any use, I agree that this proposal would allow redirecting that into the appendix namespace as well. --Tropylium (talk) 04:38, 31 January 2015 (UTC)

Support DCDuring TALK 15:52, 27 January 2015 (UTC)

Support Equinox ◑ 19:29, 28 January 2015 (UTC)

Support (but see above). --Tropylium (talk) 22:14, 29 January 2015 (UTC)

• Documenting all Unicode characters could be within scope of Wiktionary if we so decide, and I tend to support this for the sake of user convenience. Having the information in the appendix while having a link to the appendix from the mainspace seems to do the information service for the user, although I am not sure what the benefit is over having this information in the mainspace. As I have pointed out, we include letter entries as letters despite the fact that the letters have no meaning as letters; the objection was that letters at least form meaning-carrying larger objects, which I accept as an interesting one.

  As for "consensus can make CFI mean whatever we want it to mean": no, that is not the outcome of the recent vote. The CFI means what it means and nothing else, and the vote did not suggest otherwise. The opposers in the vote did not suggest that we should be lying about what the CFI says, merely that we should not consider CFI 100% binding. --Dan Polansky (talk) 19:45, 30 January 2015 (UTC)

Any final thoughts from CodeCat and Kephir on redirecting to appendices? bd2412 T 16:52, 6 February 2015 (UTC)

I would support that too. —CodeCat 17:14, 6 February 2015 (UTC)

Above I have given reasons for not using Unicode character names as definitions; since redirecting would essentially give these names the appearance of being definitions, I oppose. — Keφr 22:27, 8 February 2015 (UTC)

Since the support seems to overwhelm the opposition, I have gone ahead and redirected ⦰ to Appendix:Unicode/Miscellaneous Mathematical Symbols-B. It would be nice, however, if it was possible to make the redirect point directly to the row containing this entry, if anyone can tell me how to do that. Cheers! bd2412 T 21:25, 16 February 2015 (UTC)

@Kephir That would require some changes to Module:character list, so that it adds an anchor to each table row. —CodeCat 22:03, 16 February 2015 (UTC)

Goodbye (with regrets and thanks)

As an editor who has made over ten thousand individual edits over the last year or two, I was dismayed when another editor began undoing all my edits of the past few days without any warning, much less with running his reasons past me.

When this had just started I put an entry into the beer parlour justifying my edits in general terms. This has received no comments from the community.

I am aghast at this sort of behaviour by another editor, and dismayed that it doesn't seem to worry the community. My dismay is because I saw a future for Wikipedia in which the online environment, having enormous data capacity and the ability to link within and beyond Wikipedia, had the potential to eventually greatly outdo my favourite reference book, The Oxford English Dictionary.

However, this intracommunity brutality has quashed my expectations and has made me realise that I have become addicted to trying to improve the Wiktionary. The only way to cure an addiction is to give it up completely. So, and sadly, this is goodbye with thanks to the brute.—ReidAA (talk) 00:38, 27 January 2015 (UTC)

The reason it has no comments from the community is likely that this user is well known for this kind of harassment and point-making, so it's kind of the same old for most of us. Furthermore, stepping up against it would just cause a repeat of the kind of drama that the user is known for causing. So I think people kept quiet to avoid trouble. I know that was my reason for not responding, in any case. —CodeCat 01:07, 27 January 2015 (UTC)

I seem to recall that ReidAA was previously cautioned about using the same fairly lengthy passage over and over again, when the quote at issue appeared to be pressing a political viewpoint (even if the intention was not to press that viewpoint). It is quite frankly very rarely the case that the same quote will be the best showcase for each word used in it. I tend to think that the reasons for the contested reversions were pretty clear. bd2412 T 01:15, 27 January 2015 (UTC)

I don't see what the point of all the reversions was. It looks to me as if someone was just ticked off at him for reasons beyond understanding. Instead of having mediocre-to-average citations for uncited senses we have none. How is that an improvement? DCDuring TALK 01:54, 27 January 2015 (UTC)

It seems to me that the wrong party left. DCDuring TALK 01:56, 27 January 2015 (UTC)

Re: "the wrong party left". Agree here. --Anatoli T. ^{(обсудить}/^вклад) 04:27, 27 January 2015 (UTC)

I, too, agree. And Reid isn't the first highly-productive user Dan has driven away from the project; remember Speednat? Unfortunately, the community seems to be very reluctant to ban people, even when they are clearly hurting the project. - -sche (discuss) 19:16, 27 January 2015 (UTC)

Let the reader please check User talk:Speednat. The editor seemed to be copying definitions word-for-word from a copyrighted source; shortly after I asked him about that (User_talk:Speednat/2012#Webster.27s_Third_definitions), he stopped adding definitions and went to enter attesting quotations which I suspected were from a copyrighted dictionary as well which lead to User_talk:Speednat#Source_of_quotations; shortly after that, they left. To see that kind of editing from that editor as "highly-productive" is an error. The reader may have a look at User talk:Speednat, and check whether the complaints that I have raised on that page were justified. --Dan Polansky (talk) 20:14, 27 January 2015 (UTC)

I may have jumped the gun with my assessment, but that was the experience I recall. bd2412 T 03:52, 27 January 2015 (UTC)

I draw your attention to User talk:ReidAA items 11 et seq. I think one person single-handedly managed to drive a promising, detail-oriented contributor out of the project. I wish I had been able to pay attention to this over the last 2-3 weeks. DCDuring TALK 04:22, 27 January 2015 (UTC)

While most of the back-and-forth there does involve a single editor, there are other conversations where this editor seems to have driven others to exasperation. Of course, we probably all do that from time to time. bd2412 T 15:11, 27 January 2015 (UTC)

I think the project is worse to have lost him as a contributor, especially as he was beefing up our English language content, especially the more common words, which could well use the attention. Many of his typographic concerns are legitimate and have been neglected and could have been addressed by attention from our technical contributors together with his willingness to attend to the details. DCDuring TALK 15:47, 27 January 2015 (UTC)

I've been asked to draw attention to the recent contributions of Reid and of Dan, where the latter are just more reverts of the former. If we can come to consensus about which version of the entries in question (put away, etc) is best, perhaps we can entice Reid to return. - -sche (discuss) 19:16, 27 January 2015 (UTC)

I request that, above all, user ReidAA is prevented from (a) removing spaces from after #, and (b) performing non-consensual switching of context and label templates. If, by contrast, user ReidAA is not prevented from switching context to lb, then I do not see how I can be prevented from switching lb to context, given the best evidence about consensus or its lack available at Wiktionary:Votes/2014-08/Templates context and label. As for (c) having a single quotation used more than 10 times, I oppose that and I want to see community consensus for this before this continues; this could even by copyright violation, since fair use rationale gets weaker with this: we need quotations for attestation, but we do not need to reuse a single quotation e.g. 20 times. As for (d) put away, I emphasize that the grouped senses were not grouped by hyponymy but by being from baseball. User ReidAA has been making these kinds of willy-nilly groupings as he saw fit without any regard to lexicographical propriety, which I surmise is inferior and should not be continued. I ask the reader to check OneLook dictionaries how they do things, and check whether any of the dictionaries is making sense groupings in ReidAA arbitrary vein. --Dan Polansky (talk) 19:33, 27 January 2015 (UTC)

Later: There would be no copyright violation with even high-multiplicity repetition of quotations from out-of-copyright works. But I find such highly repeated use highly inferior nonetheless, worse than nothing. I surmise that even attesting quotations should be good examples of use, which highly repeated use makes unlikely. --Dan Polansky (talk) 19:41, 27 January 2015 (UTC)

Edits that don't change the appearance of the page, such as whether there's a space after # and whether {{context}} or {{lb}} is used, should never be edit-warred over, or worried about in any way. If someone switches between one way and the other, let them. It has no bearing on the dictionary at all. As for copyvio, what does BD2412 say? It seems to me to be less of a copyvio to use the same quote over and over than to use different sentences from the same work to exemplify different words, because that way we're using less of the total work. And copyrighted or not, I see no problem at all in using the same sentence to illustrate multiple words. I've done it myself to illustrate words of Irish. —Aɴɢʀ (talk) 19:46, 27 January 2015 (UTC)

Re: "If someone switches between one way and the other, let them." I vehemently disagree. I especially disagree that editors should be allowed to remove spaces from after #, since it massively hinders usability of the wikitext from my standpoint. Furthermore, generally, this non-consensual switching generally leads to back-and-forth even without an appearance of an edit war: one editor switches to no space after #, and, say, after a month or more, another editor switches back to their preferred form. We've seen this with "<" vs. "from" in etymologies. This non-consensual back-and-forth is unprofessional and counterproductive; it suggests immaturity on the part of the editor pushing their preferred style. It is one of the reasons why Wikipedia has rules about U.K. vs. U.S. spelling; they do not say "if someone switches U.K. spelling to U.S. spelling, let them". Moreover, I point the uninvolved reader to User talk:ReidAA to see for how long a time I have shown restraint as for label vs. context, that I talked about this multiple time to the user to no avail. I will emphasize one more time: non-consensual switching leads to back-and-forth and should be avoided. --Dan Polansky (talk) 20:00, 27 January 2015 (UTC)

The difference between US and UK spelling is visible on the displayed page, as is the difference between "<" and "from". The difference between presence and absence of a space after # is not. And there is no way the absence of a space after # "massively hinders usability of the wikitext"; it's a purely aesthetic preference on your part. "Non-consensual switching" only leads to back-and-forth if a second editor actually goes to the trouble of switching an edit back to how it was before, which in the cases under discussion is superfluous and frankly silly, since edits of this kind have no effect on the actual content. Of the diffs provided in this thread, I have not yet seen one worth getting upset about, much less one worth reverting. —Aɴɢʀ (talk) 20:12, 27 January 2015 (UTC)

You are entitled to your indifference about wikitext, and I am entitled to my care about wikitext and its legibility. Obviously, user ReidAA cares as much as I do, albeit in the other direction; if they did not care, they could have left the spaces alone. You cannot allow ReidAA to care, and disallow the very same type of care (albeit in the other direction) to me; that is unfair and unjustifiable. --Dan Polansky (talk) 20:18, 27 January 2015 (UTC)

If he were the one reverting your edits my reaction would be the same. —Aɴɢʀ (talk) 20:25, 27 January 2015 (UTC)

I think you're just using consensuality as a pretext for forcing your own preferences. Essentially, what you're implying is that any difference from the norm whatsoever needs to be discussed at length, and people are not allowed to edit as they wish. This, in turn, means that if anyone makes a change you don't like, you can just revert it and claim they have to ask everyone nicely for permission first. But the truth is, the only permission anyone ever needs is yours, as you seem to have declared yourself the arbiter of right and wrong on Wiktionary. And if anyone speaks up about it, the result is a stream of wikilawyering, tu quoqueing and other forms of communal shaming from your part. And I don't see any restraint at all, you basically just messaged a user and told them they should be blocked for not doing things they way you want. That's not just completely unproductive, that falls squarely in the "intimidating behaviour/harassment" ban reason. I would have blocked you for it already if I could be sure your sympathisers would not just unblock you again. —CodeCat 20:18, 27 January 2015 (UTC)

If anyone should feel entirely free to start removing spaces from after # regardless of previous overwhelming practice, why should not another person be entirely free to be adding spaces after # as they see fit? This is plain as a day. You seem to want to disallow me to be adding spaces after # as I see fit, and to be converting lb to context as I see fit, while allowing the conversion in the other direction to another person; how more unfair can this get? --Dan Polansky (talk) 20:28, 27 January 2015 (UTC)

You can't compare those two things. The space after # is more or less universal across Wiktionary, and so is support for continuing to keep it that way. However, you well know that the issue for context and label is much less clearly established and no consensus exists. So if you want to be fair, you shouldn't just be reverting this one user for that, but also the many other regular contributors that use {{cx}}, {{lb}} or {{label}}. The fact that you don't revert all editors, but do revert this one user, seems to me like you're just picking an easy target to bully, who you know won't have much support in the community or will to stand up against you. And now they left. You're a real hero, Dan, congrats. —CodeCat 20:32, 27 January 2015 (UTC)

I have absolutely no wish to perform any non-consensual switching in any new entry created by an editor with any of the {{label}}, {{context}}, {{lb}}, and {{cx}}, and I see nothing unfair about it. I see it as perfectly legitimate to use any of the four templates in newly created entries.

As for "The space after # is more or less universal", so was the use of the likes of {{slang}} without {{context}} before an illegitimate run of your bot. --Dan Polansky (talk) 20:45, 27 January 2015 (UTC)

With respect to the copyright question raised above, there is little to no impact on a fair use rationale from using the same quote multiple times. However, there is no need for us to expose ourselves unnecessarily to any risk of copyright infringement. I would import the practice that Wikipedia uses with images: if it is in the public domain, use it however you want; if it is not, then use it only if there is no comparable public domain alternative. Of course, new words (or words coined since 1923, in any case) will primarily be found in works that are under copyright, as will examples showing that old words remains in current use. I admit, when I add cites to an entry, I always try to find both the oldest use I can find, and the most recent. This also speaks to the quality of sentences as showcases for the words defined. We have a much better fair use argument for reproducing a sentence that does an excellent job of showcasing the word (where that word is the "star" of the sentence) than we do for a lengthy passage with the word incidentally buried in it, particularly if it can easily be shown that better example sentences are available. bd2412 T 20:32, 27 January 2015 (UTC)

Are function words ever the star of a passage (at least to anyone besides a grammarian) ? — This unsigned comment was added by DCDuring (talk • contribs) at 21:31, 27 January 2015 (UTC).

Probably only in very rare and odd circumstances, but I would guess those words would be the easiest to find in public domain sources. Don't forget that the public domain covers not only pre-1923 stuff, but all U.S. government-produced documents (including the limitless world of federal court opinions), and anything that any private author has deliberately released into the public domain. bd2412 T 21:37, 27 January 2015 (UTC)

Um, guys --

The very next thread here, though by another user, mimics the lead-in to this thread so closely that I surmise it is probably the same editor as ReidAA. Notably, this new editor also signs off as WF, making me wonder if we aren't barking up the wrong tree by censuring anyone for ReidAA's seeming departure... ‑‑ Eiríkr Útlendi │ Tala við mig 21:28, 27 January 2015 (UTC)

It could just as easily be WF mocking ReidAA; or another person entirely mocking them both. bd2412 T 21:35, 27 January 2015 (UTC)

Yes, Wonderfool is mimicking Reid's post for comedic purposes; he previously mimicked Semper's post in which he announced he would stop patrolling Recent changes. - -sche (discuss) 21:40, 27 January 2015 (UTC)

• Does anyone here have access to tools for seeing the IPs from which ReidAA and Regret and Reward have connected? Otherwise we're left with conjecture. ‑‑ Eiríkr Útlendi │ Tala við mig 00:20, 28 January 2015 (UTC)

  D'oh. DCDuring TALK 00:34, 28 January 2015 (UTC)

• Does it matter? FWIW I think it's most likely to be WF making fun of this thread and that ReidAA has nothing to do with it, but really it just doesn't matter either way. —Aɴɢʀ (talk) 00:38, 28 January 2015 (UTC)

• If the two are the same, then this current thread about why ReidAA left is basically entirely mooted. ‑‑ Eiríkr Útlendi │ Tala við mig 00:46, 28 January 2015 (UTC)

• Only a checkuser can do that, and they wouldn't do it just for a fishing expedition like this. As for whether the two are the same: the only thing they share is their manner of using quotes. In editing style, level of competence and basic approach, they're quite different. Chuck Entz (talk) 03:08, 28 January 2015 (UTC)

Hello (with relief and disapproval)

As an editor who has made over one hundred thousand individual edits over the last ten years or so, I was amused when another editor began deleting all my edits of the past few days without any warning, much less with running his reasons past me.

When this had just started I was creating lots of Spanish plurals, by the way.

I am not surprised by this sort of behaviour by another editor, and relieved that it doesn't seem to worry the community. My amusement is because I saw Wiktionary as a ting in the past, having enormous data capacity and the ability to link within and beyond Wikipedia.

I too am addicted to improving Wiktionary, and figure the only way to take my addiction to the next level is by staying put. So hello to all users old and new, it will be a pleasure working side by side with you all. WF --Regret and reward (talk) 16:07, 27 January 2015 (UTC)

Abuse filters for TV show content?

I don't know why — perhaps it's a rogue bot — but we very often get people creating TV episode lists on Wiktionary. This has been going on for months. I suggest new abuse filters that prevent entries with a title starting "List of", or entries that contain {{Infobox television. Equinox ◑ 00:26, 29 January 2015 (UTC)

I would say this is a relatively rare occurrence. If this were a bot, we would get that much more regularly. I think these are just idiots who cannot tell us from Wikipedia and click buttons at random. I rewrote Special:AbuseFilter/12 to catch this and some more. — Keφr 13:06, 29 January 2015 (UTC)

User:Type56op9

I have blocked this user (User:Type56op9) with no direct evidence. It is up to you guys to judge if this block is correct. --kc_kennylau (talk) 14:46, 29 January 2015 (UTC)

• I have always assumed he was WF. SemperBlotto (talk) 14:48, 29 January 2015 (UTC)
  • @SemperBlotto He has admitted it in his talk page, case closed. By the way, what does WF mean? --kc_kennylau (talk) 14:51, 29 January 2015 (UTC)
    • Oh, I talked too early... --kc_kennylau (talk) 14:54, 29 January 2015 (UTC)
      • I can’t help but like WF. He was always very sharp, but also immature for his age. Most people continue to mature in judgment, understanding, and outlook through the age of 30, and I assume WF is slowly growing up as well. I would not support him for admin status, but I think he’s an asset and his occasional bouts of wigging out didn’t result in any damage that was hard to find and fix. —Stephen ^(Talk) 21:45, 29 January 2015 (UTC)
        • I've come around to the position that indeffing WF was wrong and that he should be unblocked. Purplebackpack89 22:59, 29 January 2015 (UTC)

User:Type56op9 is unquestionably Wonderfool (aka WF), though only Wonderfool knows who Wonderfool himself is. As Stephen said, he's mostly an experienced, hard-working asset to Wiktionary, but is prone to irresponsibility and occasional bad behavior. He also tends to focus more than he should on quantity rather than quality. Many years ago, he did a bunch of massively disruptive stuff and got permanently blocked for it. For some years after that, he would constantly reappear under new accounts until someone (usually SemperBlotto) figured out it was him and blocked those accounts, too (he's gone through literally hundreds). Finally he got tired of doing that all the time, and proposed a compromise, which resulted in an agreement: as long as he stuck to one account at a time and behaved himself, he wouldn't get blocked. I would liken this to parole or work release: he's free to do what he wants, but we watch him carefully, and any admin is free to block him at any time for even minor misbehavior. This has worked for the most part, but he does still end up getting blocked from time to time, and he should never be taken at his word, and never be completely trusted. As Stephen said, he's gotten more responsible over the years and acts up less. The main problem lately: Given his track record with bots (we still occasionally find bad entries from years ago- especially from his Polish period), it's hard to trust him with one, and his fixation on high edit-volume makes him want to do things with bots and accelerated entry-creation all the time. Chuck Entz (talk) 02:26, 30 January 2015 (UTC)

User:Chuck Entz's pretty smart, you know. The best new user in the last few years, IMHO, thanks to his diplomacy, good humour, sharp observations and conscientious editing. He's hit the nail on the head with me, too. As for the Polish, yeah, I admit punching way above my weight in that one - I'd been learning Polish, but certainly not enough for quality entries here. --Type56op9 (talk) 07:49, 30 January 2015 (UTC)

BTW, it's nice to be talked about every now and then, too. Thanks! Perhaps we all should talk about each other more (nicely, if we can) and this place will be more harmonious. Of course, the opposite could certainly be true too... --Type56op9 (talk) 07:52, 30 January 2015 (UTC)

The only thing worse than being talked about is not being talked about. —Aɴɢʀ (talk) 09:38, 30 January 2015 (UTC)

Yes, but keeping your head below the parapet for a while, not causing controversy, and not being talked about can also be nice. Even though I'm not admin and have no desire to be, I would say no bots for this user. Donnanz (talk) 10:09, 30 January 2015 (UTC)

Cosmetics vs. makeup

Category:Fashion contains both Category:Cosmetics and Category:Makeup. What, if any, is the difference between them? —Aɴɢʀ (talk) 17:40, 30 January 2015 (UTC)

• There's a very fine line of distinction, I think. Can they be merged under one category - e.g. "Cosmetics and make-up"? Donnanz (talk) 17:57, 30 January 2015 (UTC)

There's not many entries in either category. It wouldn't take long to merge them. Donnanz (talk) 18:01, 30 January 2015 (UTC)

Distinguishing loans vs. inheritance

Etymologies on Wiktionary currently use the expression "(derives) from" in no less than three mutually exclusive ways.

1. A word has been inherited from an ancestor language, and its older form has been recorded/reconstructed as (*)foo.
2. A word is a loan from another language, where it appears as föö.
3. A word is a derivative of another word in the same language, and can be analyzed as fo+o.

The third is generally categorized separately (Category:Words by suffix by language, etc.), but the first two are currently completely mixed together under Category:Terms derived from other languages.

There seems to be widespread consensus that a word's etymology should be traced as far as possible along any one of these directions: e.g. name needs to be not only in Category:English terms derived from Middle English, but also Category:English terms derived from Old English, Category:English terms derived from Proto-Germanic, and ultimately Category:English terms derived from Proto-Indo-European. Similarly, alkali needs to be at least in Category:English terms derived from French, and Category:English terms derived from Arabic. (Whether we need to note other intermediate languages of transmission is less clear, but this is not the current point.)

Things get messy once we start mixing these together. Obviously we will not categorize terms according to their morphological structure in their source language, whether it is an ancestor or a loangiver. But plenty of words like bazaar are currently filed under "English derivations from PIE", despite deriving via a long Wanderwort chain and not via inheritance. (This particular word is pretty bad case, in that there isn't even a PIE cognate to the word itself, only to the roots from which it was assembled in Old Persian.)

To get some order around here, I believe introducing some finer-grained variants of {{etyl}} would help, perhaps by some flags. This would allow us to distinguish the various ways in which a word can derive from a different language. I could imagine eventually distinguishing all sorts of things, e.g.

• {{etyl|la|fr|via=desc}}: French words inherited from Latin
• {{etyl|la|fr|via=loan}}: French words loaned from (e.g. Modern) Latin
• {{etyl|la|fr|via=loan-desc}}: French words loaned from a descendant of a Latin word in another Romance language
• {{etyl|la|fr|via=desc-loan}}: French words inherited from Old French but loaned in there from (e.g. Medieval) Latin
• {{etyl|la|fr|via=und}}: French words deriving in some undetermined fashion from Latin
• {{etyl|la|fr|via=deriv-desc}}: French words derived from another French word, which has been inherited from Latin
• {{etyl|la|fr|via=deriv-loan}}: French words derived from another French word, which has been loaned from Latin
• {{etyl|la|fr|via=loan-deriv}}: French words loaned from a Latin word, which is a derivative of another Latin word
• {{etyl|la|fr|via=desc-deriv}}: French words inherited from a Latin word, which is a derivative of another Latin word
  • These last four could be useful if the intermediate derivative is archaic or unattested.

But just a loaned/inherited/neither distinction should be a good start. --Tropylium (talk) 20:22, 31 January 2015 (UTC)

We already have {{borrowing}}, so I propose we create another template to match it, for inherited terms. —CodeCat 20:50, 31 January 2015 (UTC)

Strong support. This would be a huge improvement on the meaningfulness of our etymological categories. — Ungoliant ^(falai) 20:58, 31 January 2015 (UTC)

I recognize that it will entail a lot of work, but I'll go along with making a distinction between loanwords and inherited words. However, I caution that the finer the level and greater the number of the distinctions beyond that which one tries to make, the less likely it is people will make the distinctions correctly. The last four categories you list, and especially the last two, are of questionable utility, in my opinion.
Re {{borrowing}}: it is useful, but not flexible — it can't handle steps beyond the immediate one, hence umlaut uses {{borrowing}} to cover the immediate step of "English borrowed this from German", but still has {{etyl|goh|en}}. That's where something like the via= parameters proposed above would be useful. - -sche (discuss) 01:13, 1 February 2015 (UTC)

This doesn't apply to a potential inheritance template though. Terms are almost always inherited from one language at most, and that language is always fixed. The source language would only need to be specified in the rare case of pidgins, which have multiple parents. —CodeCat 01:37, 1 February 2015 (UTC)

I think this is a good idea if kept limited to just a few types. It would indeed be useful to distinguish uninterrupted inheritance from everything else: Arabic كحول, for example, is a borrowing from English alcohol, but traces back eventually to Arabic again, so it would be nice to indicate that it's not ultimately of Old French origin. Direct borrowings are also good, as are derivations via an undetermined route. On the other hand, there are just too many different combinations of borrowing and inheritance to try to distinguish between the middle cases: various Germanic and Romance languages as well as Medieval Latin remained in contact for a long time, and terms were borrowed back and forth in various directions, with stages of inheritance in between. For instance, Old Norse terms were borrowed directly into Old English, indirectly into Norman French and then into Middle English, and inherited by modern Scandinavian languages, then borrowed into modern English, so you could have borrow-inherit-inherit vs. borrow-borrow-inherit vs. inherit-inherit-borrow. Latin, on the other hand, could be borrowed just about anywhere at any point in history from the Classical period on, and then inherited and/or borrowed repeatedly- so it could get really, really complicated. Lastly, "borrowing" within a language could be interpreted in so many ways, especially with complications such as analogical levelling and the like, that such categorization would be pretty arbitrary, at best. Chuck Entz (talk) 04:09, 1 February 2015 (UTC)

True enough, I don't expect to see most of those "combo" derivation types in widespread use anytime soon. Just noting some possibilities. An explicit "derived from, but neither by direct inheritance nor by a chain of loaning" flag (to put things in a "Category:X terms indirectly derived from Y") would probably be sufficient for almost all residual needs.

And yes, this would be a lot of work, but parts of this could surely be automated — e.g. if we have an English term whose etymology section starts "from Middle English", that can be automatically marked as "inherited", and if it starts "from X" where X is not an ancestor of English, that can be automatically marked as "loaned". It's the etymologies with longer chains that require more care, in general. But at least the plain "X terms derived from Y" categories would then work as laundry lists of etymologies that still need fine-tuning (or research).

(By the way, I would support not recognizing ultimate Wanderwort origins with a different meaning as plain "loaned". E.g. alcohol may well derive its shape from an Arabic word, but the important step where the meaning 'alcohol' was attached may have come later.) --Tropylium (talk) 05:02, 1 February 2015 (UTC)

• I oppose this huge overcomplication, adding a barrier to entry of etymologies to Wiktionary. The existence and use of {{borrowing}} is bad enough. I don't see OneLook dictionaries caring to distinguish borrowing vs. inheritance, and I don't see why they should. I saw people replace "from" with {{borrowing}}, and I do not now where they source this information from. And again, I do not see why our users would care. --Dan Polansky (talk) 10:13, 1 February 2015 (UTC)

The idea is not to mandate descent type parameters, but to just have the possibility. This would moreover not be about changing the etymologies themselves as listed in a given article, as much as subdivide the currently often overloaded "X terms derived from Y" categories. E.g. users who are interested in finding what words in Modern English are inherited from Proto-Germanic are unable to find this information anywhere, since Category:English terms derived from Proto-Germanic is also brimming with words loaned at some point from Norse, from Dutch, from Frankish via French, etc. --Tropylium (talk) 23:24, 1 February 2015 (UTC)

The solution to your problem is using {{etyl|gem-pro|-}} instead of {{etyl|gem-pro|en}}. Even better would be cutting off the etymology chain at the point of first borrowing and continuing it on the page of the etymon. There are currently etymologies of modern Japanese borrowings from English tracing the origin up to Proto-Indo-European. This is ridiculous and leads to duplication and desynchronization. We should also remove cognates from the etymology section if the proto-page exists and the cognates are listed there. --Vahag (talk) 11:16, 2 February 2015 (UTC)

Yes, to a point. Of course we should avoid lengthy etymologies for recent borrowings, but we don't have entries for many of the terms in the etymologies, and most people adding etymologies aren't competent in all the languages referenced. Do we really want to create an added incentive to make bad entries? You probably don't work much with many of the languages in English etymologies, but I've seen more than my share of unaccented Greek reverse-transliterated from some dictionary's etymology, complete with epsilon instead of eta, nu-gamma instead of gamma-gamma, etc. Chuck Entz (talk) 14:25, 2 February 2015 (UTC)

The chain should be cut and moved to the etymon's entry only after the latter is created. In the meantime the {{etyl|xx|-}} format can be used. This is what I do for Armenian entries. --Vahag (talk) 14:55, 2 February 2015 (UTC)

This is a possible alternate approach, but let's consider it a bit further:

• Switching en masse from something like {{etyl|fr|en}} to {{etyl|fr|-}}, if a word is as much as e.g. a loan French → Dutch → English would eliminate entirely, or at least vastly diminish, several "X terms derived from Y" categories. Say, Category:English terms derived from Ancient Greek. I am skeptical on if this would be an improvement.
• Leaving all but the immediate loangiving language entirely unmentioned I would definitely oppose. Any "Wanderwört of unambiguous origin" should note the language where the word's popularity arises, e.g. à la and variations in the appropriate subcategory of Category:Terms derived from French, even for languages that got it indirectly via English or Russian or whatever.
• Least radically: I would be mostly OK with leaving at least terms that derive via some combination of descent and loaning (and possibly derivation) without language-specific categorization. On the other hand, this solution would have to be elevated to the level of policy to be even remotely workable. And we'd have interesting border cases to contemplate with chronolects like Middle English. Greco-Latinate educated vocab, for one, has often been attested in literary languages from a relatively early date, but it does not sound reasonable that therefore any Modern English (or Modern German, or Modern Hungarian…) word should aim to only note its descent from some 500-700 years back, and not loan origins beyond that? ---Tropylium (talk) 17:01, 2 February 2015 (UTC)

I don't know, it is hard to formulate a mathematical rule on how far to take the chain, but to me it is intuitively clear in most cases. For example, I would take Japanese アーク, 仙 up to English only, フラン up to French and 蜜 up to PIE. Is that reasonable? I am probably using the notion of what you called a "Wanderwört of unambiguous origin". --Vahag (talk) 17:35, 3 February 2015 (UTC)

Re: We should also remove cognates from the etymology section if the proto-page exists and the cognates are listed there. This is not directly related to the discussion above, but sounds reasonable… to an extent. If a language group has any generally accepted "key" languages, cross-linking cognates in those seems like a good idea. (Isn't e.g. mentioning any possible Sanskrit and Latin cognates of Ancient Greek words already policy?) --Tropylium (talk) 17:01, 2 February 2015 (UTC)

AFAIK there is no written policy. I remember long time ago that me, User:Ivan Štambuk and few others decided to include one cognate from each independent branch of PIE. But now that we have many beautiful proto-pages, cross-linking to even "key" cognates is redundant. Only perverts like you and me are interested in such deep etymologies and they can make one extra click and go to the proto-page. --Vahag (talk) 17:35, 3 February 2015 (UTC)

The idea of dropping non-immediate steps from etymology sections, and making users click through the chain of etyma one link at a time, has been proposed before, and is interesting. I'm of two minds about it. On the one hand, Vahag is correct that duplicating the full chain of an English loanword's descent back to PIE in all the languages that borrowed it constitutes a hideous amount of content duplication and leads to those chains becoming out of sync when people correct errors in only some entries and not others. On the other hand, Chuck is correct that (a) people may feel comfortable noting that e.g. a certain English/French/etc word derives via Persian from Ancient Greek, but may not feel that they know enough Persian to create the Persian entry, which would mean the Ancient Greek connection would have to go unrecorded, or (b) people may feel erroneously confident that they can add the Persian and Ancient Greek entries, and make errors in adding them. Also, do I think it's useful to allow users to see which words from e.g. Aleut made it into German, even if they had to pass through English to get there. The solution is something sort-of like Module:etymtree, IMO. (But apparently not Module:etymtree itself, since it seems designed only for use in appendices.) Users should be able to say "this Japanese word was borrowed from English foo, and then have a module call all the ancestry of English foo up from the same central repository that the English entry for foo, and the Dutch and descendant-by-borrowing foo, etc, etc, call upon. - -sche (discuss) 17:29, 2 February 2015 (UTC)

Etymtree is usable in the main namespace as well. — Ungoliant ^(falai) 17:36, 2 February 2015 (UTC)

-ise/-ize words

Hi, is there any reason why some -ise/-ize pairs (e.g. pedestrianise/pedestrianize or satirise/satirize) have full duplicated (if not always synced) definitions, while others (e.g. specialise/specialize or polarise/polarize) merely have "alternative spelling of"-type entries for one spelling? 109.153.236.183 23:50, 31 January 2015 (UTC)

This is a point of much contention. Some (like myself) argue for having the content in a single place, and a stub-like "alternative form of X" entry at other places; this leads to arguments about people's personal favourite spellings being marginalised (ized!). There is also the fact that in some cases (I think honor/honour is one), certain obsolete senses can only be attested under one of the spellings. The pair that usually comes up in these arguments is color/colour. Equinox ◑ 23:52, 31 January 2015 (UTC)

I think it would be better to have only one entry, e.g. "pedestrianise or pedestrianize" or "honor or honour", or however it was decided to present it, to which both lookups redirected. Then the precedence would at least only be in the order of presentation. Cases of only one spelling being attested could be fairly easily handled with, for example, a "honor only" or "honour only" note against an individual definition. 109.153.236.183 00:05, 1 February 2015 (UTC)

I believe someone's working on a way to do this. (Who was it? Stand up!) There are probably some technical hurdles to overcome. Equinox ◑ 00:17, 1 February 2015 (UTC)

Maybe you're thinking of the Grease pit thread "Revisiting the issue of English UK/US spellings and entry synchroni(s|z)ation". I am interested in this subject, of course. Donnanz (talk) 00:26, 1 February 2015 (UTC)

Yeah it's here: [1]. Thanks. Equinox ◑ 00:35, 1 February 2015 (UTC)

Any idea anyone what the technical issues are? I know that Wikipedia can somehow get italic words into article titles (see e.g. here). If the same can be done in Wiktionary, then the heading could be, for example:

honor or honour

Then I suppose some fiddling would be needed with the "en-noun" and "en-verb" templates (or dual-spelling alternatives), but this doesn't seem an insurmountable hurdle. Templates are all under control of Wiktionarians and can be made to do whatever we want, right? As a minimum solution, the "en-noun" or "en-verb" line could just be repeated for the alternative spelling. 109.153.236.183 15:02, 1 February 2015 (UTC)

It's a Lua module: w:Module:Italic title. I don't know much about Lua, but I think it should be easy enough to install on Wikitionary (although it would need some rewriting to make it only italicise "or"). Smurrayinchester (talk) 10:01, 3 February 2015 (UTC)

• tangent regarding spelling-specific senses:

  Can anyone point to a concrete example of a word with a certain meaning attested only in one national standard and not in another? I'm just curious, because I can't. I can point to examples where different non-US/UK-specific spellings have different senses. I've spent the past couple of days very comprehensively citing the Norse-gaumr-derived word gaum/gorm/goam. I'm still looking for more citations, but it's looking a lot like:

  • the sense "understand; comprehend; consider" is attested 3+ times only in the spelling gaum,
  • the sense "gawk; to stare or gape" is attested 3+ times only in the spelling gorm, and
  • the sense "see, to recognize, to take notice of" is attested 3+ times only in the spelling goam,

  even though all three senses are found in dictionaries, and in use <3 times, in both of the first two spellings, and they constitute, on a spoken level, the same word.

  Point being that US/UK differences are not the only (and perhaps not even the main) cause of words' meanings being split across different entries.

  (I agree with 109.153 that spelling-specific senses "could be fairly easily handled with, for example, a "honor only" or "honour only" note against an individual definition".)

  - -sche (discuss) 19:53, 4 February 2015 (UTC)

  I'm not entirely sure what you're looking for. Something along the lines of a sense of favorite that favourite does not have? If so, there's a meaning of program that programme doesn't have, namely a computer program (which is spelled program in en-GB as well as en-US). —Aɴɢʀ (talk) 21:29, 4 February 2015 (UTC)

  My opinion: entries should be centralised and all senses should be attested on one of the spellings. Some reservations, e.g. is [[disk]] also used for intervertebral disc, as disc? --Anatoli T. ^{(обсудить}/^вклад) 22:06, 4 February 2015 (UTC)

• @Aɴɢʀ: When the subject of centralizing -ize/-ise and -or/-our and other US/UK pairs of spellings' content onto single pages comes up, people sometimes say "what if a given sense is only attested in one spelling?" My points are (1) senses being split by spelling isn't just a US/UK thing, (2) splits can be handled in the way the IP suggests, and (3) senses being split specifically by US/UK spelling seems to be not only too vanishingly rare to represent an obstacle, but perhaps entirely nonexistent. An example of a US/UK spelling split would be if "realize" were used to mean "say foobar" in America, always with the -ize spelling, while it was never used to mean that in Britain or with the -ise spelling. (In such a case, if we centralized page would have to contain a note that the "say foobar" sense way US-only.) In the case of program/programme, I note that google books:"computer programme" is actually very well attested. Even if it weren't, the fact that both national varieties of English used the spelling program would be a different kind of issue. (Specifically, it would mean there were two things we needed entries for: one thing spelt "programme" in the UK and "program" in the US, and another thing spelt "program" in both countries. In the "realize" example, in contrast, we'd be talking about a thing that was attested only in one country and in one spelling, and that wasn't used in the other country in any spelling.) that turned into a ramble, I apologize/apologise... - -sche (discuss) 22:30, 4 February 2015 (UTC)

  So something more along the lines of British arsed (as in I can't be arsed to do it), which isn't *assed in American English, because that word isn't used with that meaning in American English? I know that's not a perfect example because arse and ass differ in more than just spelling, but if spelling were the only difference between them, would that be what you're looking for? —Aɴɢʀ (talk) 23:17, 4 February 2015 (UTC)

  A {{qualifier|only "arse" spelling}} could be used then or something similar. --Anatoli T. ^{(обсудить}/^вклад) 23:20, 4 February 2015 (UTC)

  Ooh, yes, if it's true that "assed" isn't used that way, then that's the sort of thing I was talkng about. :] And as Anatoli says, it's straightforward to handle with labels (and/or usage notes). - -sche (discuss) 03:01, 5 February 2015 (UTC)