Wiktionary:Beer parlour/2014/August

Definition from Wiktionary, the free dictionary
Jump to: navigation, search
discussion rooms: Tea roomEtym. scr.Info deskBeer parlourGrease pit ← July 2014 · August 2014 · September 2014 →

Contents

Normalised spellings and CFI[edit]

For some languages, we commonly respell the words into a common form. This is done for Old Norse, Old High German, Middle Dutch, and other old languages. As it is now, CFI does not actually allow for this practice, but I think it should be allowed. So we should probably codify this practice as an exemption. Something along the lines of "for languages for which a normalised spelling is adopted, the normalised spelling itself does not need to be attested, as long there are unnormalised spellings of the same word that do meet CFI". —CodeCat 23:09, 3 August 2014 (UTC)

Support. That's definitely the case with Old Church Slavonic or Old Russian. Most quotations of these in modern Russian use modern Cyrillic letters, instead of old letters, which makes the terms in old spellings difficult to attest. --Anatoli T. (обсудить/вклад) 00:19, 4 August 2014 (UTC)
Support, and for each language that normalizes spellings, we would have to detail the normalization rules on its "WT:About X" page. We can consider an non-normalized spelling to attest its normalized spelling obtained by following the normalization rules that we have listed for the language. --WikiTiki89 13:38, 4 August 2014 (UTC)
I forgot to mention that quotations should always be added in the original non-normalized spellings whenever possible, and known non-normalized spellings should be listed in alternative forms sections. --WikiTiki89 13:53, 4 August 2014 (UTC)
  • Words should be added as they are spelled in attestation. This "normalized spelling" idiocy is another attempt to impose artificial uniformity where there is none, namely in attestations of all languages before the 19th century where there were usually no enforced rules of spelling. It will make Wiktionary completely useless as a resource because we would never know whether the added word was attested as such, or is a a guessed transcription according to a scheme devised by some wiki nickname. Any kind of "normalized" spellings should be used strictly as redirects to real spellings. -Ivan Štambuk (talk) 13:46, 4 August 2014 (UTC)
    It does not impose uniformity, it merely makes it easier to find the entry where the definition is located. Also, all dictionaries do this, which nullifies your usual argument of breaking accepted conventions. --WikiTiki89 13:56, 4 August 2014 (UTC)
    If the point were in finding entries, then the normalized spellings themselves would be redirects, not the main entries. Instead, what is suggested is that all of the main entries be somehow normalized, regardless of how they are attested, containing all of the definitions, citations and so on for all of the spellings that they resolve to under some lossy scheme, and actually attested entries be soft redirects, and paradoxically listed as "alternative spellings" under the normalized entry (how can real attestations be alternative spellings to something made up?!)
    When it comes to ancient languages, all of the paper dictionaries have space constraints that require usage of a standardized spelling scheme to help look up entries. A single word could have a dozen different spellings. However, online dictionaries do not suffer from such limitations. We can have everything - the original script in Unicode and not Latin transliteration, citations, as well as a list of widely used scholarly transcriptions, normalization schemes, reconstructed pronunciations or whatever - but the latter not as full-blown entries, because they are not real words but reconstructions. --Ivan Štambuk (talk) 14:19, 4 August 2014 (UTC)
    Online dictionaries have screen real estate constraints as well. I would not like to picture what an inflection table would look like if it includes all attested variant spellings. --WikiTiki89 14:29, 4 August 2014 (UTC)
    For ancient languages inflection tables are not that important. People don't use them to learn to speak those languages (except maybe Latin and Sanskrit, but that's insignificant). For them, much more important points are accuracy and reliability. Inflection tables which only contain attested forms in their original spelling are much more important than inflection tables containing reconstructed forms that were possibly never attested in those spellings. It's a difference between Wiktionary as a serious reference work, and Wiktionary as a conlang community. --Ivan Štambuk (talk) 14:49, 4 August 2014 (UTC)
There's also the problem of unattested lemma forms. We have an entry for πρίαμαι (príamai), for example, but according to Liddell & Scott that particular form is not attested. That doesn't happen too often in Ancient Greek, which has an enormous corpus, but it happens very frequently in languages like Gothic and Old Irish. I've been creating entries for unattested lemmas in both of those languages, but I've been wondering if that's really such a good idea. Maybe we should put them in the Appendix namespace alongside other reconstructed forms. —Aɴɢʀ (talk) 14:54, 4 August 2014 (UTC)
If the lemma form can be easily determined, then I think it is the best place to define the term. We can note on the page that the lemma form is unattested and I guess it makes sense to be able to mark or remove the unattested inflected forms as well. Inflection tables are still very useful, especially when all or most of the forms are in fact attested. --WikiTiki89 15:24, 4 August 2014 (UTC)
Support. For Old Norse and Old High German, I expect that normalized spellings actually meet CFI, firstly because Norse texts are so regularly printed in normalized form, and secondly because print and online dictionaries (the former of which are sufficient verification, per CFI, of extinct or poorly-documented languages) invariably use normalized spellings. Another set of languages that already benefit from normalized spelling and would benefit further from having the practice codified are the indigenous languages of North America, which different dictionaries and text-collections have often used slightly different orthographies to represent. For instance, in many languages, some sources have represented long vowels with macrons (ā), or circumflexes (â), other sources have used trailing mid dots (·), and still others have used doubling (aa). Pace Ivan, I think it'd be hilariously nonsensical for e.g. one third of the inflected forms of a term or one third of a set of compounds that share a common element to use ā, while another third used aa, most of the rest used â and a few entries used , all because someone preferred to blindly copy and paste the idiosyncrasies of the different dictionaries the forms were attested in rather than think critically about them for a moment. - -sche (discuss) 17:51, 4 August 2014 (UTC)
To be fair, I don't think that Old Norse texts that are printed in normalized are allowed to attest the normalized spelling. The actual attestation should be of the original spelling(s) from when the language was still in use. --WikiTiki89 18:07, 4 August 2014 (UTC)
Paradoxically, those spellings are much harder to attest. It's much like the scripts of Gothic: it was written in Gothic script originally, but everyone "normalises" it into a transliterated form nowadays. —CodeCat 18:12, 4 August 2014 (UTC)
Yes, that is why if we allow entries at normalized spellings in CFI, we must remember that normalized texts attest the term, but not the spelling. --WikiTiki89 18:18, 4 August 2014 (UTC)
Those "normalized Gothic" texts are not attestations of Gothic language. That is not how the Gothic was written. Those are scholarly transcriptions made for scholarly purposes. They are equivalent to e.g. respelling any language in phonemic transcriptions. Nobody writes Gothic today. It's a dead language with small and fixed corpus. Those kind of transcriptions are not attestations of Gothic. --Ivan Štambuk (talk) 09:31, 5 August 2014 (UTC)
I disagree that only original editions of works can be cited; I don't see such a restriction in WT:CFI. Two editions are not independent of each other for the purposes of citing a single word/spelling (e.g. one can't cite both the American and British versions of Harry Potter and have them count as two citations of castle), but nothing I see prohibits citing different editions to confirm the existence of different words or spellings — e.g. citing an American edition of Harry Potter as a use of the word favor, even if JK Rowling's original used favour. The American edition is durably archived and verifiably uses favor several times (making clear that it isn't e.g. a typo). Likewise, the normalized editions of Norse texts are durably archived. (In most cases, they're far better archived and far more accessible — as copies exist in hundreds of libraries — than the original manuscripts, which are periodically destroyed by fires and in historical cases may even have been destroyed before any un-normalized editions of them were printed. But that's mostly superfluous to my point.) Consider also how many translations of the Bible have been cited to verify various words around here — CFI's prohibition against citing two "verbatim or near-verbatim quotations or translations of a single original source" only stops us from citing two editions of the Bible as citations of the same (spelling of a) word, it doesn't stop us from using two editions of the Bible as citations of two different words. - -sche (discuss) 20:17, 4 August 2014 (UTC)
What I meant was a reproduction of an Old Norse text printed well after Old Norse died out cannot count as a citeation of Old Norse. However, we can assume that an unnormalized reproduction reflects the original spelling and use it to attest spellings, and we can assume that a normalized reproduction does not necessarily reflect the spelling but still reflects the form of the word and we can use it to attest the term and its form, but not its spelling. --WikiTiki89 20:25, 4 August 2014 (UTC)
@-sche: I think it'd be hilariously nonsensical for e.g. one third of the inflected forms of a term or one third of a set of compounds that share a common element to use ā, while another third used aa, most of the rest used â and a few entries used a· - Indeed it would be nonsensical from the perspective of someone who imagines that Ancient Greek, Gothic, Sanskrit, Old Church Slavonic, Akkadian, Hittite, Old High German, Middle Persian, Old French and others were written by a single and unified speech community, who spoke a single language in a single point in time, as opposed to being spoken an written across many centuries (often millenia) by a diverse communities who never knew each other, who wrote in ill-fitting lossy scripts under the influence of traditional orthography not necessarily reflecting actually spoken sounds, and languages of documents X and Y who are today treated as parts of a single ancient language X would be in any other occasions treated as two completely separate languages, were they attested today.
I'm receptive to the idea of having both 1) reconstructed, template-generated inflection fitting some "idealized" model of a language as well as 2) listing only actually attested forms (I believe Old Irish conjugation and Old Persian declensions currently does that). But, simply ignoring all of the variation in order to fit them into some kind of imaginary order is a disservice to any serious potential users of Wiktionary. The only ones who would benefit from that would be non-serious users who could then claim that they "learned" some ancient language as presented by Wiktionary, even though such language never existed in the form it is being presented. It would be similar to many of our protolanguage inflection templates who present some kind of ridiculous Stammbaum-like picture of parent language dissolution reflecting a POV of a single linguist, which never existed as such.
Regarding the barely documented indigenous languages - they are a separate category. They are usually a living thing, and if one scholar uses â and another ā to represent what is indisputably the same sound, it makes sense to standardize on the most common notation and use others as redirects. But if some ancient language uses three different symbols for the [a:] sound, we cannot standardize it on anything because we don't have a clue whether those symbols meant the same thing (even though some, but not all, think they did). We can't make that kind of value judgments. If the original documents are still being published in facsimile editions, it means that no normalization is possible. There could be exceptions - e.g. Gothic with a tiny corpus and a small number of authors (one, is it? Ignoring Crimean Gothic). But for the majority it's not practical at all.
Anyway, this should all be discussed on an individual language basis. --Ivan Štambuk (talk) 10:41, 5 August 2014 (UTC)

In modern languages too, such as French, there may be normalized (recommended) spellings, and it's sometimes very difficult or impossible to find attestations for these normalized spellings. When there if an official recommendation, I think that they should be includable. For old languages, the issue is more difficult. Of course, they should be included when attested (even when the olf spelling cannot be found), but not considered as the main entry (the other entry should be as complete as the normalized one). If a (normalized or old) spelling is included even when it seems to be unattested, the fact that no attestation has been found should be made very clear in the entry. Lmaltier (talk) 18:20, 4 August 2014 (UTC)

Unattested lemma forms and CFI[edit]

Kind of a spinoff based on what Angr brought up. Currently, the common practice is to reconstruct the lemma form if it is not attested, and place the entry there. If several lemmas are possible, we generally include them all and choose one at random. This practice is primarily done with old languages, but it's easily conceivable that it could happen to modern languages as well. For example, if all we have for a particular English lemma is two attestations of fonges, one of fonging and one of fonged, then I doubt we would put the main entry at one of those entries. We'd put it at fonge, even though it's not attested. CFI doesn't say anything about this practice, but as it's so widespread both on Wiktionary and outside it, I think we should clarify and codify it. —CodeCat 18:18, 4 August 2014 (UTC)

I don't think English verbs are the best example of the phenomenon. I have worked on and observed cases where we have -ing forms and -ed forms as distinct entries, but do not have the presumed verb lemma form. In the absence of the lemma, the -ing form is often shown as noun and/or adjective and the -ed form as adjective. This seems to actually be a fairly common evolution, with the base and -s forms coming well after the -ing and -ed forms, if indeed they ever materialize in use. I cannot recall specific cases, but, if it is important, instances could probably be found. The best way would be by extracting the cases from the dump. DCDuring TALK 18:37, 4 August 2014 (UTC)
I did include the 3rd person singular present as one of the attested forms in my example, and the past form could include the past tense as well. —CodeCat 18:50, 4 August 2014 (UTC)
So we're both operating without real cases. I'll rejoin the discussion when someone, possibly me, has a real case. DCDuring TALK 20:04, 4 August 2014 (UTC)
Here is a real case: I cannot find the French verb arsenicaliser in its lemma form, but I can find it in a conjugated form : "Le praticien qui a le plus arsenicalisé le monde et dont l’expérience a le plus d’extension, de richesse et de certitude, M. Boudin, préfère actuellement l’acide arsénieux et se tient exclusivement à lui dans tous les cas : (…)" (Annales de la Société de médecine de Lyon, 1851) (it's undisputably a verb in this sentence) or "Nickel minéralisé par le fer & le cobolt sulphurés & arsenicalisés ;" (François Rozier, ‎Jean André Mongez, ‎Jean-Claude de La Métherie, Journal de physique, de chimie, d’histoire naturelle et des arts, 1777). Lmaltier (talk) 20:28, 4 August 2014 (UTC)
Here are some more real cases: Passargisch, ostweserisch, and several of the other adjectives in Category:German terms with rare senses. - -sche (discuss) 22:23, 4 August 2014 (UTC)
There are plenty of real cases from extinct languages; I already brought up πρίαμαι (príamai), which itself is not attested, but other forms of it are (see [1]). Old Irish examples include ad·gnin, ailid, and claidid. —Aɴɢʀ (talk) 22:32, 4 August 2014 (UTC)

Need for entries or a field in the 'create a new entry page' that may just be cross references for all spellings.[edit]

There is a requirement for entries that may just be cross references, for all spellings that use characters that are not in the usual Romanised character set.

Just going to the <create a new entry> page is very frustrating. The cross reference might be an extra field on the <create a new entry> page, with the title something like <have you viewed ...>.

Repeatedly I am on a page but cannot search for it to get back to it or to get to similar pages in the index to get back to it.

<kephalḗ> is the Romanisation of <κεφαλή>, but you cannot search for <κεφαλή> using <kephal ...>. It is often very difficult to imagine what you need to do to get back to a page that you have accessed when using the etymology, especially.

I expect to be at Wikimania on Wednesday ...

Genevieve Hibbs

Tocharian question[edit]

@Ivan Štambuk: and @Word dewd544: in particular since they seem to be our most prolific Tocharian editors: I see from kuse that the vowel letters that are normally represented as subscripts are represented by full letters in the entry name, but the headword line shows the subscript (in this case, kuse). Is this the best way to do this? "Kuse" and "kuse" correspond to two different spellings in the original script, don't they? If and when Unicode finally provides the Tocharian alphabet, we will presumably want to move our entries to forms written in the native script (hopefully retaining the Latin-alphabet entries as "Romanizations of..."), and if we want to do that by bot, it would be good to have entries under unambiguous names. Shouldn't the Tocharian B section of [[kuse]] be moved to [[kᵤse]] instead? The only problem I foresee is that sometimes it's "ä" that's subscript, and Unicode doesn't have a character for subscript "ä". For those cases maybe we could cheat and use "ₔ" instead. What do y'all (and anyone else interested) think? —Aɴɢʀ (talk) 14:09, 5 August 2014 (UTC)

Yes it should be moved, I wasn't even aware that subscript u sign <ᵤ> existed in Unicode until now.. --Ivan Štambuk (talk) 15:02, 5 August 2014 (UTC)
And are you OK with using "ₔ" for "ä"? Are there even any entries that currently call for that? —Aɴɢʀ (talk) 15:17, 5 August 2014 (UTC)
I'm fine with that. These issue should best be discussed on the about-page for Tocharian. We only have a few hundred Tocharian entries, and they need to be rechecked and referenced at any case. Unicode support doesn't seem to be coming anytime soon. If you feel like doing that, just knock yourself out... --Ivan Štambuk (talk) 16:21, 5 August 2014 (UTC)
We don't have an about-page for Tocharian. I don't have the resources to recheck and reference the Tocharian entries, but if I happen to see any subscripts in headword lines, I'm happy to move the info to a new entry name. —Aɴɢʀ (talk) 16:33, 5 August 2014 (UTC)
Yeah, I agree that it should be moved as well. I think the reason I didn't use those characters originally was because I didn't know the others actually existed for use on here, and also just made them based on the format that was used for the few existing entries, from what I remember. But this way is better. Word dewd544 (talk) 13:52, 13 August 2014 (UTC)
Cool. Is [[kᵤse]] the only one? —Aɴɢʀ (talk) 14:31, 13 August 2014 (UTC)
OK, I've moved the Tocharian B section to [[kᵤse]] and, I believe, fixed all the links that were pointing to it. I looked through the lemma categories of both Toch. languages and couldn't find any others with subscript vowels, but I may have overlooked something. —Aɴɢʀ (talk) 10:48, 7 August 2014 (UTC)

Representing Old Irish "tense" sonorants[edit]

Anyone interested in Celtic languages or IPA transliteration (or both, or anyone who just wants to put their oar in) is invited to join the discussion I've just started at Appendix talk:Old Irish pronunciation#Representing the tense sonorants. —Aɴɢʀ (talk) 01:25, 9 August 2014 (UTC)

Lists of dictionary headwords[edit]

Are they subject to copyright? I am interested in creating appendices containing lists of headwords of some notable dictionaries that are still under copyright, as well as some additional information not contained in them. --Ivan Štambuk (talk) 20:31, 9 August 2014 (UTC)

Compiling your own list is not copyrighted as far as I know. —CodeCat 20:32, 9 August 2014 (UTC)
Copyright is a matter subject to interpretation anyway... So let's not give a shit about it --Fsojic (talk) 20:44, 9 August 2014 (UTC)
But I don't want my own list. I want an enhanced lists of words or reconstructions exclusively from certain works so that the experience of browsing them could be simulated by clicking. Additionally, references which refer to them could back-link to such lists. It could also be good for verification and inspection of coverage. I prefer lists and tabular presentation over categories.. I recall a discussion a while back about Brian's hotlist which was kept, so I suppose it's not a big deal. But such lists would be exposed outside userspace, and that seems a bit more problematic, so I'm asking if it could be prohibited for some reason. --Ivan Štambuk (talk) 21:08, 9 August 2014 (UTC)
Intellectual property lawyer hat on. The namespace that a list appears in is irrelevant to copyright law. As far as I recall, Brian's hotlist is a compilation of headwords from several different dictionaries, and therefore can not identifiably impinge on the copyright of any one of them. I think that it would be problematic, at least, to list the headwords of a specified edition of a specified, in-copyright, printed dictionary. Such a list of words defined reflects the editorial judgment of the dictionary's authors, and is therefore likely to be covered by copyright. Doing so for one that was out of copyright would be fine. A possible workaround would be to make one set of lists of words defined in out-of-copyright versions of specified dictionaries, and a separate list containing a combination of words not defined in the out-of-copyright versions (which will basically be words that are new since their publication) but which are defined in unspecified "major dictionaries". bd2412 T 22:01, 9 August 2014 (UTC)
Interesting. Thank you very much for this explanation. --Dan Polansky (talk) 09:20, 10 August 2014 (UTC)

Can we automate Hiragana and Katakana transliteration?[edit]

I see we don't currently have automatic transliteration of Hiragana and Katakana. Is there a technical reason why we can't, or is it just that no one's gotten around to it yet? —Aɴɢʀ (talk) 11:02, 10 August 2014 (UTC)

It's possible, but it would give incorrect results when they are mixed with Kanji. So the module would have to check for the presence of Kanji characters and return nothing if found. —CodeCat 12:09, 10 August 2014 (UTC)
Mixed terms should rely on kana, e.g. 勉強する and 電子メール should use kana spellings べんきょうする (benkyō suru) and でんしメール (denshi mēru), if it only transliterated the hiragana/katakana part する and メール, it would be a mess. --Anatoli T. (обсудить/вклад) 01:00, 11 August 2014 (UTC)
Would that be a lot of work? Obviously we shouldn't do it if it means listing every single one of the 6.3 kilosagans of possible Kanji characters, but if it can be done with less than 100 characters of code, why not? —Aɴɢʀ (talk) 12:19, 10 August 2014 (UTC)
Module:ja already does it (in Japanese headwords). It's not implemented in link templates, as the transliteration may be incorrect. Wyang (talk) 23:53, 10 August 2014 (UTC)
The automatic transliteration is used in Japanese entries and usexes ({{ja-usex}}) and some other templates. It's only not used in translations. For this to happen, the translations would need to follow the same format as entries, using spaces in multipart words or phrases with particles, capitalisation (forced with symbol ^ or automatic on proper nouns). Besides, many kanji translations don't have hiragana, which is needed for transliterations to happen. --Anatoli T. (обсудить/вклад) 00:02, 11 August 2014 (UTC)
The other challenge is that the Japanese transliteration is somewhat context-driven, as I said, proper nouns (which excludes language names, demonyms, month names, weekdays) are capitalised, verb with final おう are transliterated as "-ou", rather than "ō", there are cases when morphemes need to be separated ("." is used in entries), particles は and へ are "wa" and "e", rather than "ha" and "he". --Anatoli T. (обсудить/вклад) 00:56, 11 August 2014 (UTC)

Block policy clarification[edit]

The current blocking policy page WT:BLOCK seems misleading. I propose to reduce the page content to the following wikitext:

:''See also '''[[Help:Interacting with humans]]'''''
{{policy-CFIELE}}
{{shortcut|WT:BLOCK}}

# The block tool should only be used to prevent edits that will, directly or indirectly,
hinder or harm the progress of the English Wiktionary.
# It should not be used unless less drastic means of stopping these edits are, by the assessment
of the blocking administrator, highly unlikely to succeed.

===See also===
* [[Wiktionary:Range blocks]] - when and how to block a range of IP addresses
* [[Wiktionary:Vandalism in progress]] ([[WT:VIP]]) for currently occurring or very recent vandalism
* [[Wiktionary:Vandalism]] (or [[WT:VANDAL]]) for vandalism of Wiktionary in general

As per Wiktionary:Votes/pl-2010-01/New blocking policy, the above text is the only binding part of the page.

Note that I placed "policy-CFIELE" there, so that the criteria for further modification of this page should be identical to those of CFI and ELE.

What do you think? --Dan Polansky (talk) 12:06, 10 August 2014 (UTC)

Special:Abusefilter is supposed to filter enough to allow to talk with the staying editors if they're wrong, and so encourage them toward perfection. How many valuable professionals could post their personal site in reference by ignorance, we can't treat all of them as some incorrigible spammers, it would contravene to WT:Be bold.
Moreover, I'm still considering that if the current WT:BLOCK had been applied with my known Wikimedia bot (3 millions editions and 21 flags), the blocker wouldn't have to refuse to assume any hurried arbitrary decision. I saw too much waste because of friendly fire by the past.
That's why letting a message was a sine qua non condition before forbidding indefinitely the open wiki. JackPotte (talk) 13:10, 10 August 2014 (UTC)
Apart from that we could also recruit more patrollers, for example by giving this status to them automatically after 500 editions, like on the French Wikipedia. JackPotte (talk) 16:24, 11 August 2014 (UTC)

Template:pedia[edit]

Template:pedia was redirected from one page to another earlier today, resulting in a number of pages being broken. Instead of Template:pedia redirecting to Template:projectlink/Wikipedia, I request that Template:projectlink/Wikipedia redirect to Template:pedia instead. Since Template:pedia is linked to from [ https://en.wiktionary.org/w/index.php?title=Special:WhatLinksHere/Template:pedia&namespace=0&limit=5000 thousands of pages], it seems the more likely target. Purplebackpack89 23:56, 11 August 2014 (UTC)

Support it being at Template:pedia
  1. Purplebackpack89 23:56, 11 August 2014 (UTC)
Support it being at Template:projectlink/Wikipedia
Discussion

Did you even look at the page history? Template:pedia has been a redirect since 2007. —CodeCat 23:58, 11 August 2014 (UTC)

And when you moved it earlier today, it wasn't working on the pages I looked at. It shouldn't have been moved by you earlier today, and you shouldn't have deleted the page you did. You also shouldn't have edit-warred, and you should have provided better edit summaries. Purplebackpack89 00:02, 12 August 2014 (UTC)
That's because you reverted my move while I was still in the process of updating all the redirects to point to the new location. Your revert actually broke the template altogether because it ended up pointing to a deleted page. You should have taken more care before making changes when you didn't know what you were doing. You should also have taken more care to get the facts clear before posting erroneous and misinformed "polls" like you did, which do nothing but embarass you and waste the time of other editors who have better things to do than deal with you. —CodeCat 00:05, 12 August 2014 (UTC)
You coulda saved yourself the work of not updating all redirects by not making the move in the first place. Nothing will convince me that it was a good idea to make that move. Heck, things would have worked just fine if you'd let Template:pedia have the full text it did in my last edit. Nobody will ever use Template:projectlink/Wikipedia, because just adding Template:pedia is so much easier. Why don't we just have Template:projectlink/Wikipedia redirect to Template:pedia? Everything would be so much simpler that way Purplebackpack89 00:24, 12 August 2014 (UTC)
Template:projectlink/Wikipedia is not meant to be used directly in entries anyway. Rather it's meant to be used through {{projectlink}}, which supports many other projects. {{pedia}} is just a remnant from before it was converted to {{projectlink}} back in 2007. All the projectlink pages are named beginning with PL:, including Template:PL:pedia, which Template:pedia was originally a redirect to. All I did was move Template:PL:pedia to Template:projectlink/Wikipedia. I am intending to move all the other PL: templates too, as they are properly subtemplates of Template:projectlink and are only meant to be used in conjunction with it. Having them as subpages makes that relationship more clear. I really don't understand why you are making such drama out of it. —CodeCat 00:34, 12 August 2014 (UTC)
Because you broke pages, and they wouldn't have been broken if you hadn't messed around with the template. You probably shouldn't have deleted Template:PL:pedia either. Purplebackpack89 00:42, 12 August 2014 (UTC)
It has no transclusions, so why would we keep it? It's useless. —CodeCat 00:47, 12 August 2014 (UTC)
  • By the way, it was COI for CodeCat to protect a page she was edit-warring on. For the life of me, I don't understand why CodeCat is still an administrator. She edit-wars frequently, she rarely explains what she's doing, and she protects things she's engaged in edit wars on. Purplebackpack89 00:28, 12 August 2014 (UTC)
    • You're just looking for reasons to get your right when I've already countered your other arguments. You're pretty much pulling the idea that I was edit warring out of your hat in an attempt to put me in a bad light while excusing yourself. If someone breaks things or makes other bad edits repeatedly, there is nothing wrong with edit warring. It's just un-breaking the wiki. Imagine if we had to start a discussion whenever someone kept re-inserting "poop" into an entry. It would be rediculous! —CodeCat 00:34, 12 August 2014 (UTC)
      • For starters, the last edit I made to Template:pedia wasn't a bad edit. I want you to look closely at it before calling it a bad edit. Secondly, I was acting in good faith trying to restore a template that was showing up as broken on a page. Somebody who inserts "poop" into a page is vandalizing. In one of those cases, it is acceptable to edit-war. In the other, it isn't. If you don't understand which is which, and you think it's OK to edit-war to revert good-faith edits without even an edit summary explaining why you did what you did, then you have no business being an admin. Purplebackpack89 00:42, 12 August 2014 (UTC)
        • To be fair, I think very little of what you do on Wiktionary is truly good faith. You mostly get on people's nerves and are obstructive almost on principle, and people have said so many times in the past. You've even driven away other valued and productive editors with your behaviour. So if I shouldn't be an admin, then I suggest you shouldn't be on Wiktionary at all. —CodeCat 00:47, 12 August 2014 (UTC)
          • I resent your accusation. I tried to fix that template because it was showing up as broken, not to piss you off. I vote keep at RfD because I believe the project would be improved with more articles, not to piss Mglovesfun off. Every mainspace and RfD edit I make is in good faith and with a view to improving the Wiktionary Purplebackpack89 00:50, 12 August 2014 (UTC)
            • Of course, but so are all of my edits. :) I never said that you did it to piss me off. That's not what the other editors who have complained said either. But good faith edits are not equal to good edits, and are therefore not exempt from being reverted. Having good intentions also doesn't prevent you from getting on people's nerves. A while ago there was User:KYPark who kept inserting rather outlandish etymologies at WT:ES, and would get very philosophical about the ideas while not really contributing or making any kind of point. He got upset when we started moving them to his userspace because he didn't understand that it didn't belong there, and after his behaviour continued for about a year or so, he got blocked, I think even several times. There was no discussion about a block, but nobody really minded that he was blocked because he had annoyed and frustrated so many people that nobody was willing to stand up for keeping him. They were glad he was finally gone. The reason I am telling all this is that something similar may eventually happen to you as well. You would do well to try to be a friend of the larger Wiktionary community, because all the good faith in the world will not help you if they are fed up. —CodeCat 01:02, 12 August 2014 (UTC)
From the edit history of Template:pedia it is apparent that Purplebackpack's starting assumptions (since amended) are mistaken. Template:pedia was not moved; it has been a redirect since 2007. And it was not CodeCat's updating of the redirect target but Purplebackpack's revert of that which seems to have broken some existing uses during the update that was being made. Purplebackpack's subsequent unilateral insertion of thousands of bytes of duplicated code also created quite a mess. Purplebackpack says "I was acting in good faith". As Wikipedia observes at w:WP:CIR, "[some users] believe that good faith is all that is required to be a useful contributor. Sadly, this is not the case at all. Competence is required as well. A mess created in a sincere effort to help is still a mess." - -sche (discuss) 01:07, 12 August 2014 (UTC)
  • Explain how the code insertation created a mess. Purplebackpack89 01:10, 12 August 2014 (UTC)
  • Also, there are certain things that acting in good faith entitles you to. One of them is a clear explanation when you are reverted. CodeCat did not give one. Purplebackpack89 01:12, 12 August 2014 (UTC)
    • I was more focused on undoing the damage than on giving an explanation. Fixing thousands of entries had a higher priority to me than satisfying one user. —CodeCat 01:20, 12 August 2014 (UTC)
      • Maybe you shouldn't have broken them, then... FWIW, CIR isn't policy here or even on Wikipedia, it's merely an essay, and it's a bad idea, because it flies in the face of being BOLD, and taking chances with edits. It's also walking too fine a line, because it's impossible to understand why a particular editor did a particular edit. Finally, it requires a level of communication that is present on Wikipedia but not on Wiktionary; Wikipedia not only has fewer things that can be broken (since they don't use as many templates and lack a rigidity of article structure), it also is better at explaining to editors what's wrong. Purplebackpack89 02:57, 12 August 2014 (UTC)
      • Furthermore, @CodeCat:, your attitude that explaining your edits to other editors is of little or no import is disheartening, to say nothing of being wrong. You complain about me being hard-headed, but I've mentioned this to you at least half a dozen times, and other editors have mentioned it as well, and you've ignored them. It's very disingenuous for you to make a CIR-based argument when you have not been forthcoming about why you're right. Purplebackpack89 04:36, 12 August 2014 (UTC)
        • (edit conflict) CodeCat didn't break anything- you did. You assumed bad faith, and didn't bother to ask or investigate. I'm not going to apologize for CodeCat- sometimes I vehemently disagree with her actions, and I've done my share of griping about it. I've even reverted a few of her edits- but only when things were seriously broken and she wasn't around to fix them, and only after carefully analyzing everything to make sure I wasn't going to make things worse.
        • You see, normal people would post a complaint on her talk page or in the forums first and demand to know why she was doing it. You, on the other hand, know better than everyone else and reserve the right to unilaterally step in and take over any time it sort of looks like someone might be doing something wrong- shoot first, and ask questions later. And then, when it's demonstrated that you were mistaken, you don't admit you were wrong, you don't apologize- no, you attack the person you interfered with for not explaining things so even you could understand. After all, you never make mistakes- the only way you could ever be wrong is if someone else misleads you into being wrong. Chuck Entz (talk) 07:54, 12 August 2014 (UTC)
          • Chuck, I saw something was broken and tried to fix it. I did that in good faith, and felt I was owed an explanation for why my edits were wrong. Purplebackpack89 14:07, 12 August 2014 (UTC)
  • Now as before, I find the CodeCat pattern of discussion-free and summary-free edits to infrastructure objectionable. CodeCat hardly ever explains themsemselves, but require explanation for opposition to their edits. CodeCat lacks the maturity to understand that excessive change with little added value is bad. --Dan Polansky (talk) 07:11, 12 August 2014 (UTC)
  • PBP, for this and other edit-warring incidents you have been stripped of rollback and autopatrolled privileges. The latter increases the chance that someone without "COI" will notice any disputes with you, so you should be thankful, really. Further misbehaviour will be met by a block. Keφr 08:10, 12 August 2014 (UTC)
    This should be immediately undone, since the BPB vs. CodeCat incident had nothing to do with autopatrolling and rollback flags. Especially the autopatrolling should be returned back, since the mainspace edits of PBP are largely undisputed, and removing the flag will increase patrolling cost to the patrollers. Furthermore, the threat of a block is inappropriate, since PBP was edit warring with CodeCat on a page which CodeCat edited without consensus; a block or desysopping of CodeCat could be in order, given the long-term pattern of their editing behavior. --Dan Polansky (talk) 08:31, 12 August 2014 (UTC)
    Flags restored. We can't just punish one party in a conflict, especially considering that the reverts were (perceived as) legitimate, and that there is no pattern of (perceived) abuse of those flags.. --Ivan Štambuk (talk) 08:55, 12 August 2014 (UTC)
    The reverts were perceived as legitimate by whom? Keφr 09:07, 12 August 2014 (UTC)
    By them, obviously, otherwise they wouldn't have done them. --Ivan Štambuk (talk) 12:33, 12 August 2014 (UTC)
    I stand by what I did. PBP may not have used the rollback button here, but the repeated combative and misinformed edit-warring is evidence that he cannot be trusted with it. Patrolling burden should not be a problem; PBP has made one edit yesterday, two the previous day, three edits two days ago, and previous 13 edits were on 3rd of August, so his edits are quite infrequent. However, the few edits he makes do need attention apparently. In my opinion a block is not only appropriate, but long overdue. This is not just a single incident, and PBP refusing to learn (from past mistakes and from everything else) is a huge red flag. And nothing prevents you from starting a desysopping vote. Keφr 09:01, 12 August 2014 (UTC)
    In Wiktionary:Beer_parlour/2014/June#Purplebackpack89, in the hidden section "Rights removal", two editors supported flag removal while four editors opposed, two of which explained that since PBP has not abused the flags, they should not be removed. Again: since the editor has not abused the flags, they should not be removed. Furthermore, nowhere in this thread have you noticed that CodeCat refuses to learn. You have singled out the fairly harmless PBP, and conveniently ignored the editor who by my lights have caused actual damage in the mainspace, unlike PBP whose only damage are drama threads in Beer parlour, a fairly unimportant thing. In this very thread, the drama was sustained by CodeCat, who continued to respond to PBP posts. But again, the drama itself is fairly harmless, an attribute of an open wiki where people can actually speak up. --Dan Polansky (talk) 09:17, 12 August 2014 (UTC)
    Wasting people's time on futile discussions is not "fairly harmless". And again, if you think CodeCat's actions are so egregious, what are you waiting for? For any punishment to be effective, CodeCat needs to be desysopped first. Keφr 09:35, 12 August 2014 (UTC)
    There is no sound desysopping process. That is why Ruakh left before he would have to deal with CodeCat in this environment. The only desysopping process that we tried relied on the 2/3-supermajority consensus for desysopping. And this of course enables CodeCat to perform mass changes with unclear support, possibly even less than plain majority support, and be fairly sure they will not get desyssopped, since there probably is something like 45% or more of supporters of what they are doing; I have invented the 45% number and I do not really know the scope of support for their various changes. --Dan Polansky (talk) 09:46, 12 August 2014 (UTC)
    "Nowhere in this thread have you noticed that CodeCat refuses to learn." What she has refused to learn is that the editing process would be a helluvalot easier for everybody concerned if she used edit summaries. She has repeated refused to even consider doing so. Purplebackpack89 14:01, 12 August 2014 (UTC)
As annoyed as I am with PBP's behavior and attendant whining in this incident, I disagree with removing his flags over it. I simply don't see the relevance. This reminds me a lot of the whole Gtroy/Acdcrocks/LuciferWildcat affair: in that case, improper harassment generated enough sympathy that he was able to continue with his prolific creation of subpar and often fabricated content far longer than he would have otherwise. Chuck Entz (talk) 13:57, 12 August 2014 (UTC)
Are you suggesting we should repeat the same mistake by letting him loose? As you see, there is enough strife in this community without stubborn ignoramuses adding to it. Keφr 14:29, 12 August 2014 (UTC)
  • One thing's for sure: Kephir overstepped his bounds with his removal of rights. His "beef" against me has translated into HOUNDing and irrational admin actions, and this after I told him multiple times that interacting with me is unproductive. I am very close to considering he be forbidden from interacting with me for the good of the community. Purplebackpack89 14:07, 12 August 2014 (UTC)

Practically this whole discussion is making me facepalm...stop getting so bent out of shape over such piddly little things... To precounter any (slightly) likely accusations, I'm not taking sides here.

  • PBP: You thought CodeCat fucked things up, but in reality things only got messed up because you didn't let them finish the redirections and such that they were doing. Accept that you made a mistake and move on. I get that this was a somewhat more widely used/high profile, etc template not some obscure thing but maybe instead of, as Chuck said, assuming bad faith you should have posted to CodeCat's page to the effect of "Do you realise you broke this thing?" (since that's what you thought happened) before reverting. User: PalkiaX50 talk to meh 14:59, 12 August 2014 (UTC)
Where do you get the idea I was assuming bad faith? And you guys fail to acknowledge that CodeCat bears some responsibility for not initially communicating why she did what she did. Purplebackpack89 16:22, 12 August 2014 (UTC)

Lots of errors in Old French nouns (and in verbs too, before I fixed them)[edit]

(This is a bit of a rant. No offense intended to whoever created all the mistakes ... maybe User:Mglovesfun?)

I notice a bunch of mistakes in Old French noun declension. Of the first 4 words I checked out, 3 had incorrect declensions.

  • seror is mistakenly listed under suer; suer is the nom. sg. and seror the obl. sg. but WT has them reversed.
  • empereor is the std form, but WT claims that empereür is standard and redirects the former to the latter when it should be reversed; it also messes up the nom. sg. (should be emperere not empereres) and obl. pl. (should be empereürs not empereres).
  • ameor has the same declension as empereor (nom. amere - ameor, obl. ameor - ameors) but is listed with a totally different declension, broken in a different way from empereür.

A little more looking reveals

  • chanteor has the same declension as ameor and empereor but is listed with a messed-up declension that is different from both the messed-up declensions of ameor and empereor/empereür; what a mess. Its etymology is also broken ... it lists a mistaken cantor instead of cantātor.
  • BTW empereür's etym. is slightly messed up, listing a non-existent Latin word imperātōr with a stray long mark over the o.
  • chaceor, robeor, troveor have the mistaken declension of ameor.
  • compaignon and felon should have similar declensions; both are broken, each differently from the other.
  • nonain and *ante (should be antain) again should have similar declensions and are broken, each differently from the other.

I'm sure there are tons more. How did this get so messed up?

BTW, the Old French verb conjugations were utterly messed up, too, and full of wrong-way redirects as well, but I've put a lot of work into fixing them.

I'd suggest in the future that it would be better to have no declensions/conjugations at all than completely wrong ones.

Benwing (talk) 09:17, 12 August 2014 (UTC)

  • Well, what happens with editors in the more obscure languages is naturally that less people know anything about them, so it is tough to find out if they are correct or not. The same happened with plenty of other editors. There was a guy called User:Razorflame who editted in tonnes of languages, but users better than him kept pointing out his mistakes in these languages, which made him move to other languages - before long he was editting in Kannada, and became the self-proclaimed Kannada expert (and since nobody else knew anything about the language, he was allowed to edit to his heart's content, doubtlessly filling this project with crappy Kannada entries). The same thing has happened myriad times, for example User:Wonderfool with Asturian - he claims to be married to an Asturian woman who knows the language, and since nobody else edits in that English, he becomes the "local expert". My suggestion (and hope) is to fix as many of the Old French entries as you can. It's highly probable that Mglovesfun has made plenty of mistakes, so we appreciate any new editors in less widespread languages like. Wonderfool too would appreciate other Asturians to correct his work. --Type56op9 (talk) 14:30, 12 August 2014 (UTC)
    Maybe you can even get your Asturian wife to check all your edits. --WikiTiki89 15:57, 12 August 2014 (UTC)

CodeCat pushing original research[edit]

User:CodeCat is again pushing large-scale original research (OR) into etymology sections of mainspace articles, as well as articles for protolanguage reconstructions in the appendix namespace, but this time removing cited scholarship (which he added in several instances) and replacing it with his own fabrications (en example). In the past he objected to tagging his made-up theories with the template {{original research}} which he unilaterally deleted out of process having removed all of the instances of articles being tagged with it (en example). Neither of these were discussed anywhere and CodeCat never uses edit summaries. His behavior is detrimental to the both credibility of Wiktionary as well as discouraging for any editors involved in those areas who see their work undone in such dictatorial manner. --Ivan Štambuk (talk) 12:31, 12 August 2014 (UTC)

We've been over this before. Wiktionary does not have a policy or prohibition against original research, and just because you say it's unwanted doesn't mean it is. —CodeCat 12:33, 12 August 2014 (UTC)
No we haven't been over this. Many have objected to this practice. And what you're doing here is something entirely different - bending different (legitimate and scholarly-supported) theories into something original and thus useless, but seemingly supported by references. And you do it repeatedly, without discussion, and when it's reverted you revert back to the disputed version containing your original research, claiming that the disputed version should be discussed first before reverting. --Ivan Štambuk (talk) 12:38, 12 August 2014 (UTC)
Yes, because I disagree with your moves. You also disagree with mine. So we're at a standoff. That's why I called for a discussion, to form a real consensus on WT:AINE-BSL rather than to just edit war over it. This discussion is not going to get us anywhere as long as it's just the two of us. —CodeCat 12:41, 12 August 2014 (UTC)
I object to CodeCat placing their unsourced original theories where sourced theories exist. --Dan Polansky (talk) 12:54, 12 August 2014 (UTC)
As for Wiktionary:About Proto-Balto-Slavic, each sentence present there that is not based on consensus should be tagged "[disputed]" or the like, to make it clear the page does not represent consensus. --Dan Polansky (talk) 12:56, 12 August 2014 (UTC)
Every single part of the Proto-Balto-Slavic reconstructions I created can be sourced. What is not sourced is the exact written form of the words. Instead, I converted them to use a common notation, just like we do for other reconstructed languages. I don't understand what is so controversial about it. —CodeCat 13:02, 12 August 2014 (UTC)
Okay. I object to CodeCat replacing (or renaming) particular forms that are sourced with particular forms that are unsourced. --Dan Polansky (talk) 13:04, 12 August 2014 (UTC)
Just to put things into context here. Are you suggesting that if a source attests, say, Indo-European *teutā or *teutéh₂, and someone creates an entry with that name, then we are not allowed to move that to *tewtéh₂ even if no source attests it in that exact written form? Because that's the equivalent of what I've been doing for Balto-Slavic. —CodeCat 13:10, 12 August 2014 (UTC)
Yes, that is correct, just that by "even if no source" you probably meant "since no source". The sourced exact written forms should prevail unless there is an overwhelming consensus to the contrary. --Dan Polansky (talk) 13:18, 12 August 2014 (UTC)
Well, I would say that we already have a consensus as WT:AINE already details how forms are to be normalised. Some of it I've written, but some parts of it were already there before (in particular the bit about laryngeals). There has not been any dispute about that practice, and it has been enforced by other editors as well, so I believe consensus can be assumed. So then I would conclude that there is, in fact, a consensus for moving those entries to the normalised form *tewtéh₂. Furthermore, there is also an established practice to normalise even attested languages, including most prominently Old Norse and Old English, but also languages with a prescribed standard orthography. So I can only assume that normalising spellings is a well-established practice for Wiktionary and if I was supposed to treat it as something controversial or disputed, I would have expected more evidence for that. —CodeCat 13:23, 12 August 2014 (UTC)
Given my past experience with you, I don't believe a single word that you say about "consensus". So please deliver objective evidence of consensus; I will not consider any consensus claims made in the absence of such objective evidence. --Dan Polansky (talk) 13:26, 12 August 2014 (UTC)
You can't prove a negative. Consensus exists through the lack of dispute. As there has not been any dispute regarding the normalising of spellings in PIE, consensus can be assumed. —CodeCat 13:29, 12 August 2014 (UTC)
Re: "Consensus exists through the lack of dispute": Absolutely not. Either an overwhelming common practice or a discussion is a prerequisite for there being a consensus; both can be demonstarted by objective evidence. Since we now know that your consensus claims are based merely on your perceived "lack of dispute" and conventiently fit your long-standing pattern of mass editing without consensus, the need for you to provide objective evidence has been corroborated. --Dan Polansky (talk) 13:33, 12 August 2014 (UTC)
So then there is actually no consensus between us on what consensus is. That's going to be difficult. —CodeCat 13:36, 12 August 2014 (UTC)
Re: "Consensus exists through the lack of dispute": That is an absolutely outrageous view of consensus. In any event, obviously we now have evidence of lack of consensus. DCDuring TALK 13:37, 12 August 2014 (UTC)
For what it may be worth: AINE and AINE-BSL before CodeCat edited them. The latter was only edited by Ivan before. Keφr 14:40, 12 August 2014 (UTC)
What should also be looked at is how many of the entries that existed at that time followed the policy on those pages. I'm fairly sure that when Ivan created the page for PBS, it was when the dispute had already started, and most of the Balto-Slavic entries that existed at the time did not in fact adhere to the practices that Ivan was detailing. So it was not an attempt to codify practices but to establish his own as canonical in contradiction to what was already present on Wiktionary. The edits I made after that corrected that, while also inserting practices I felt were more reasonable, but were not established by anyone prior to that. For the PIE page, I believe at the time most of our entries (there weren't that many yet) also didn't adhere to the spelling norms on that page. So I probably edited the page to reflect the reality, although it's long ago so I don't really remember. —CodeCat 22:08, 12 August 2014 (UTC)
User:CodeCat: Maybe you should compile a list of those instead of saying "fairly sure". Apparently some people here do not take your "fairly sure" very seriously. Which is not necessarily the problem with those people. Keφr 10:09, 13 August 2014 (UTC)
What you did is 1) changed the policy page removing the stuff you don't like and overriding it with your original research 2) mass renamed a bunch of pages, and edited the references to them in the articles 3) undid my reverts when I challenged those changes.
Your claim that "most of the Balto-Slavic entries that existed at the time did not in fact adhere to the practices that Ivan was detailing" is nothing but lies. Most of them were referenced except for the reconstructions that are a figment of your imagination and cannot be found anywhere in the literature, and which are not deleted due to your interpretation of "no policy against original research" = "I can do whatever I like".
And now again you mention spelling norms which this has nothing to do with. These different reconstructions represent completely different theories and cannot be unified under a single "normalized" spelling. I explained that countless times. --Ivan Štambuk (talk) 10:32, 13 August 2014 (UTC)
  • The so-called "normalized spellings" argument is a red herring, and an attempt to push a particular POV. Writing a instead of o, ź instead of ž, or writting a glottal stop sign ʔ or not is not merely a "normalization of spellings" - these represent completely different protolanguages, reflecting different theories by different linguists. Usage of innocent terms such as normalization is merely an attempt to trivialize implications of such edits. These differently reconstructed protolanguages in fact represent completely incompatible theories and cannot be reconciled via notational convention. It's not like w = u̯ in Proto-Indo-European, in the example given by CodeCat above. --Ivan Štambuk (talk) 14:29, 12 August 2014 (UTC)
    • As far as I know, the majority of Balto-Slavic linguists accepts the existence of a so-called "acute" register for Proto-Balto-Slavic. The use of *ś rather than *š reflects a real phonetic difference in Proto-Balto-Slavic (that of PIE *ḱ versus *s + RUKI) and this difference is maintained in Slavic as *s versus *x/*š, and I'm not aware of any dispute about this either. I'm less certain about *o versus *a, but using *a in all cases seems like the more conservative approach, at least until more sources start popping up supporting *o. So far I've only seen Kortlandt's arguments, but his theories are hardly mainstream. —CodeCat 14:36, 12 August 2014 (UTC)
      Indeed they do, but most do not treat them directly as glottal stops - usage of ʔ by Leiden school is in a completely different framework (glottalic theory of PIE, PIE *H > segmental *ʔ) than others who e.g. claim that it was just phonologically redundant feature in long syllables, or perhaps not (when preserving the *V: vs. *VH distinction). I was referring to from PIE patalovelars and not the RUKI-induced . Regarding the *a vs. *o - the only Proto-Balto-Slavic dictionary published since Trautmann's 1927 book uses *o, so it's pretty far extreme to impose *a like you've been doing. Different reconstructions = different theories, and NPOV requires us to abandon any attempts of notational "normalization" and treat all of the incompatible sources equally. --Ivan Štambuk (talk) 15:26, 12 August 2014 (UTC)
      The use of a superscript glottalization symbol is not meant to indicate that it was indeed a glottal stop of any kind. It's just an abstract symbol that stands for the acute register, whatever its nature was. As for *ž, I realise that you meant that, but it doesn't make much sense to write *ž < PIE *ǵ but *ś < PIE *ḱ. We should use the same diacritic for both of them. And the sources I've seen so far mostly indicate the PIE palatovelars with an acute accent. If we indicate them as *š and *ž instead, then what symbol shall we use for the RUKI-induced variation of *s? Concerning *a versus *o, the problem is that if we distinguish them only sometimes but not other times, that's going to confuse users who may think there is real significance to this. They might think, quite reasonably, that if one noun ends in *-os and another ends in *-as, that this reflects some real difference rather than different theories. Therefore, I opted to not distinguish them in the normalisation, so that no false impressions are given about this. We could of course normalise in the other direction, but the issue there is that the *o versus *a distinction is not reconstructable in the majority of cases. —CodeCat 15:44, 12 August 2014 (UTC)
  • I think it's high time that the original research policy on etymologies be formulated and voted on. If at any point a legitimate and referenced scholarship that I added could be overridden by some anon on the basis of a WT:ES discussion I want to know it so that I know what I don't want to waste my time on. --Ivan Štambuk (talk) 16:22, 12 August 2014 (UTC)
Why should proto languages be exempt from all attestation rules? It's true that they can't be attested in the same way as attested languages but that doesn't allow us to just use invented forms. Wiktionary is no scholarly platform for linguistics, if you want that, go to Wikiversity. This isn't the first time CodeCat's behavior over here has been criticized - soon I'll be at a point where I am forced to take action in form of a WT:VOTE. -- Liliana 16:27, 12 August 2014 (UTC)
I would love to see that. While CC's OR Finnic roots don't bother me all that much as I once told her it would be better for both her and Wikt. if she used an appropriate platform for her OR, e.g., Amazon self publishing (I might even buy it.)
And then there is the much more serious issue of meddling with actual referenced etymologies, see diff where arguably the most authoritative source on lv etymologies – Karulis – was replaced by a root from god knows where all the while the superscript [1] at the end would have a reader believe that it was, in fact, Karulis who offered that root, essentially we end up with what is called (I think) a fabricated citation on Wikipedia. Removing intermediate steps (as in diff) in a referenced etymology also evades me, there are two homographic terms with completely different meanings why would trimming down a ref'd etymology be a good idea?
Re: OR Uralic roots I was disappointed by CC's ignoring of important corollary information from authoritative sources when crafting her OR root appendices, for example etymoloogiasõnaraamat was explicit in the fact that Finnic word for "shoulder; help" is ultimately an Iranian borrowing which for some reason she saw as not worthy of inclusion in her appendices for this root (e.g., Appendix:Proto-Finnic/api) and I'm lost as to why... Neitrāls vārds (talk) 18:14, 14 August 2014 (UTC)
*pūteiti is supposed be the infinitive form except that infinitive for Proto-Balto-Slavic verbs cannot be reconstructed because Old Prussian evidence doesn't agree with East Baltic and Slavic. But let's just ignore that conveniently in order to "fix" the inherent flaws of the tree model of language change and "normalize" entries... --Ivan Štambuk (talk) 16:30, 16 August 2014 (UTC)

What is consensus?[edit]

There is some disagreement about the exact nature of consensus in the above discussion. Specifically, the debate is over whether lack of debate implies consensus on a particular thing. My view is that it does, as we generally tend to follow the practice that an edit is ok until someone reverts it or complains about it. So my question is, in the absence of any discussion, can consensus be assumed? And if not, what should be done with the many unwritten and undiscussed rules that were never formally "consensusified"? Also, if I'm correct that consensus is needed for every edit on Wiktionary, what does that mean for the millions of edits on mainspace entries for which no discussion was made beforehand? Is requiring explicit consensus for every change workable? —CodeCat 13:47, 12 August 2014 (UTC)

  • This is a clarion call for instituting the BRD process they have at Wikipedia. An undiscussed edit is a "bold" edit. If another editor disagrees with that edit, he/she can revert it. At that point, you discuss. If no one disagrees with a bold edit, it isn't discussed. Purplebackpack89 13:57, 12 August 2014 (UTC)
  • A planned massive change cannot be claimed to be supported "by consensus" if there is no discussion and no evidence of overwhelming common practice, merely "lack of dispute", and the lack of dispute is caused by the fact that the change was not proposed in a public forum in the first place. As for the need of discussion, mass changes absolutely should not be put on par with single edits of mainspace articles. When specific claims of consensus are made without reference to a discussion or a vote where people expressed their agreement, the consensus is less certain but still possible, and can be proven by pointing out to a long-standing overwhelming common practice, sample of which can be provided by the claimant. When such a hypothesis of consensus is presented in a public forum, the rest of the editors can try to find a significant volume of refuting counterexamples to the putative common practice claim.

    The "dispute", "consensus" sequence is the opposite one: if I make an edit in a mainspace and no one oppose it but no one also becomes aware of the edit, there is no point in talking about consensus. It is only after there is at least a shred of dispute that talk about consensus and consensus forming becomes meaningful in the first place. --Dan Polansky (talk) 14:06, 12 August 2014 (UTC)

    • I think we should write WT:Consensus and have it approved by vote. —CodeCat 14:23, 12 August 2014 (UTC)
      • You should better write User:CodeCat/Consensus, so that everyone can know that, by your lights, a proposal that you did not even make is automatically supported by consensus since no one managed to dispute it. I think trying to write WT:Consensus could have some nasty repercussions, since it delves into meta-levels and infinite regress; it is this infinite regress that you are abusing here. --Dan Polansky (talk) 14:30, 12 August 2014 (UTC)
        • Wikipedia does fine with their w:WP:Consensus, and I think we need something similar. In fact, I think copying and amending it would be good. —CodeCat 14:39, 12 August 2014 (UTC)
  • Let us return to the substance. The problem with w:WP:Consensus is that it is bullcrap not so great. Consider this: "Any edit that is not disputed or reverted by another editor can be assumed to have consensus." That cannot be the case. A consensus is a state of general agreement. No one can be thought to agree with an edit of which they are not even aware of. As long as the edit is undisputed, it is not known whether it is supported by consensus, but in the absence of indication to the contrary, there is no need to revert the edit. Agreement and disagreement with an edit is only possible after the edit arrived to the attention of the person agreeing or disagreeing. --Dan Polansky (talk) 15:29, 12 August 2014 (UTC)
  • Well there was opposition to you doing OR in etymologies but that didn't stop you from continuing to do so. Do you perceive stopped being reverted as "consensus was formed" ? --Ivan Štambuk (talk) 15:40, 12 August 2014 (UTC)
  • The sense in which we have "consensus" at Wiktionary when folks fail to object to something is reminiscent of a "consensus" of users in response to arbitrary changes of user interface on must-use systems such as some of those of the federal government or Google or one's IP provider. One of the points of a wiki is to elicit contributor commitment by getting beyond such supposed consensus to one based on authentic participation. That users continue to participate in Wiktionary despite unsatisfactory changes and unsatisfactory methods of enacting changes wrought by technical contributors is a tribute to the preexisting commitment those users have because of the wiki idea and their prior efforts. It is also a tribute to the burgeoning complexity of the way in which many aspects of Wiktionary are implemented, which concentrates power in those very few who have both the time and the motivation to enact that complexity. Whether that complexity is actually necessary rather than a way of increasing the power of the motivated contributors is now moot. We now are stuck with the complexity and are help hostage to the whims of such contributors.
So our "consensus" seems to reflect the realities of power, more than anything else, though laziness and weakness of commitment to the project may also shoulder some of the blame. CodeCat is simply making that explicit. DCDuring TALK 16:59, 12 August 2014 (UTC)
Mostly the latter I think, as you can probably see by observing our "consensus-building" processes. You often either have the discussion stuck in nowhere because people are unable to agree on a minor detail, or complete lack of participation or "meh whatever" responses. And then someone uses a pretentions Latin phrase to justify reverting any changes. I am not surprised at all that CodeCat tends to bypass those venues. All this apathy can be really frustrating. Keφr 10:56, 13 August 2014 (UTC)
Insinuations may sound great, but do not belong on an open wiki and in an open discourse. The "someone" would be me. The Latin phrase would be "no consensus => status quo ante". The phrase "status quo" has been used over the years repeatedly by other editors as well, and has been an important principle that I have not introduced. As for e.g. Wiktionary:Votes/bt-2009-12/User:JackBot (to which you are untrasparently referring above, which is a poor practice), 6 editors participated; since the vote did not state the task for the bot, I would have opposed as per Ruakh and msh210 there; I don't know what you are complaining about. We do not need neverending repeated mass changes, especially those with significant oppositions; making many changes is enjoyed by immature juveniles, who avoid the real building of the real dictionary, unlike e.g. SemperBlotto or Equinox. I am sure the reusers and parsers of Wiktionary data (there are some) do not enjoy incessant changes either. --Dan Polansky (talk) 11:16, 13 August 2014 (UTC)
  • Blimey! How is it that the talk page of the article I'm editing on Wikipedia about the Israeli-Palestinian conflict is more collegial than this? I am tempted to collapse the entire last fourth of this month-subpage for generating "more heat than light". I will start blocking editors if they continue to speculate so incivilly about other editors' gender or genitalia; such speculation is not only irrelevant to our stated goal of making a dictionary but outright harmful to that goal, because it creates an exceedingly hostile environment that turns off potential contributors and tries to drive away existing contributors (it did drive Cloudcuckoolander to leave for a while, IIRC). It meets every criterion of our written blocking policy; it directly and deliberately hinders/harms the progress of Wiktionary in the way aforementioned, it has continued despite less drastic means of stopping it being attempted, it wastes everyone's time and it causes editors distress by directly insulting them and being continually impolite towards them.
    Some of you clearly feel that some of CodeCat's edits, and her tendency to implement them without discussion and even in the face of opposition, are also harmful to the project (and it was said above that they drove Ruakh to scale back participation in the project); if you would like to propose to desysop CodeCat, or block her if she continues her own behaviour, those options remain open [to all of you and to those of you who are admins, respectively], though continuing a civil discussion seems like a better course of action. (But, to be clear, CodeCat's behaviour and the misgendering above are not comparable.) - -sche (discuss) 21:46, 12 August 2014 (UTC)
    Thank you. Keφr 10:02, 13 August 2014 (UTC)
    I believe that many do not believe that we can desysop CodeCat without even more damage to the project, as the skimpy documentation for our system would make it difficult to maintain. I think this is probably a wrong belief, as shown by our operating successfully without CodeCat's presence after the last conflict, but who wants to test it for an extended period? I think our ever-increasing dependence on complex modules and relentless "tidying" of templates that would compete with such dependence has the effect of increasing CodeCat's power over the project. DCDuring TALK 22:07, 12 August 2014 (UTC)
    I think DCDuring said we use too many templates and they're too complex. If that's what he said, I agree. Purplebackpack89 22:16, 12 August 2014 (UTC)
  • For the uninformed: -sche (sic) is one of the major enablers of CodeCat's editing without consensus, and himself guilty of repeated controversial mass editing without consensus. --Dan Polansky (talk) 22:28, 12 August 2014 (UTC)
    Right. User:-sche apparently blocked me for "speculating about CodeCat's gender", while at the same time we can see threats and ad hominem attacks against editors like User:Purplebackpack89 go unpunished. --Ivan Štambuk (talk) 14:41, 14 August 2014 (UTC)
    To clarify, I blocked you because you are the only person I've seen continue to speculate after I asked (warned) people to stop that particular long-practised irrelevant/harmful behaviour. (I decided not to institute ex post facto blocks to other users as long as they stopped, and so far they have.) I am none too happy about Kephir calling Purplebackpack a "lying illiterate troll", but I hope that sort of incivility can be discouraged by discussion. (Also, I am not the only one with a mop, other people could step up to the plate and issue warnings and blocks if ad hominem attacks continue...) - -sche (discuss) 16:34, 14 August 2014 (UTC)
    There was already misunderstanding regarding my usage of the personal pronoun he so it was necessary for me to elaborate on that. I simply declared my position on the topic I didn't brought up in the first place, so if you want to issue warnings and blocks you're barking up the wrong tree. --Ivan Štambuk (talk) 16:19, 16 August 2014 (UTC)
  • As someone who has interacted with both editors (Codecat and Polansky) in both negative and positive terms i can say i am a neutral and impartial party in this conflict. Nonetheless, although i have also encountered posturing behaviour by Polansky, i think that has been balanced out by his helpful lessons he's given me on how guidelines on wiktionary work. Pass a Method (talk) 14:46, 14 August 2014 (UTC)

Superprotect certain pages such that sysop permissions are not sufficient to edit them[edit]

Users may be interested in [2], which creates a new protection level called "superprotect" for "protecting pages such that sysop permissions are not sufficient to edit them". The new protection was developed after a series of events on en.WP, and was applied to de.WP's MediaWiki:Common.js after a series of events there. (De.WP held a RFC on Media Viewer and found that consensus was for it to be disabled by default and opt-in rather than enabled by default and opt-out, a de.WP admin implemented that consensus via w:de:MediaWiki:Common.js, and an edit war occurred between that admin and another admin + a WMF person. Events on en.WP were similar, except en.WP admins didn't edit-war.) There has been some discussion of the new protection level at en.WP (permalink to current revision), though it has generated more heat than light, and there is a RFC on Meta.
You may also be amused by this; if you don't speak German, the key bit (after the initial post by Bene* in English) was BHC's reply "does Bene* even have the right to edit MediaWiki:Common.js now?"
- -sche (discuss) 19:04, 13 August 2014 (UTC)

I'm sure that's a tool CodeCat would love. -- Liliana 19:15, 13 August 2014 (UTC)
All power to the technocrats. DCDuring TALK 19:22, 13 August 2014 (UTC)
I think the idea is to lock a page for a few days, during which nobody can edit, discuss what the right thing to do is, and then do the right thing when the protection ends. I can get behind that Purplebackpack89 19:24, 13 August 2014 (UTC)
What we really need is shadowbanning ;) Equinox 19:28, 13 August 2014 (UTC)
Yeah, then we coulda banned Mglovefun so I wouldn't have had to read all his low-level digs of me. Purplebackpack89 20:25, 13 August 2014 (UTC)
Was that comment really necessary? -- Liliana 21:14, 13 August 2014 (UTC)
Yes. It serves as evidence that PBP's favourite pastime here is not dictionary-building, but trolling. Of which I think there is abundance already, but whatever. Also, meet kettle, pot. Your remark about CodeCat above in this section was equally gratuitous. Keφr 21:19, 13 August 2014 (UTC)
Kephir, calling me a troll is inaccurate (witness how many entries I've created in the last 24 hours alone) and is evidence of your continual campaign to get me banned from the project for relatively innocuous edits. I have told you numerous times that interacting with me is unproductive. One more remark like that and I will request a one-sided interaction ban on you interacting with me. Purplebackpack89 21:28, 13 August 2014 (UTC)
A "one-sided interaction ban" sounds absurd. Googling the phrase only finds you demanding them in a few places. Equinox 21:30, 13 August 2014 (UTC)
But don't you agree with the general idea that things would be better off if Kephir stopped interacting with me? Purplebackpack89 21:35, 13 August 2014 (UTC)
Oh c'mon. Do you really need to bring your seemingly-unlimited paranoia in here? -- Liliana 21:40, 13 August 2014 (UTC)
Paranoia, Liliana? Dude tried to remove my autopatrol and rollback rights less than 48 hours ago (see above discussion where he was shouted down). He has been targeting me for months since that disruptive pump thread a month and a half ago. He freely admits to monitoring my edits. And he just accused me of primarily being a troll, which is a flat-out lie. Purplebackpack89 21:44, 13 August 2014 (UTC)
Since you mentioned that, you have edited a whopping number of six pages in main namespace in the last 24 hours; what do you want, a biscuit? As for you being a troll, your ban at SEW and subsequent lack of remorse for what caused it should be enough evidence of that.
Also, if you really thought that interactions with me are such a waste of your time, you could simply avoid having them and ignore me. Simple as that. Keφr 22:15, 13 August 2014 (UTC)
How many new entries, Kephir? Three new entries in the last 24 hours (one of which had multiple definitions).
I cannot ignore you because you keep inserting yourself into my editing, even though I've asked you to not do so many times, and am asking you do so again. You have admitted to monitoring my edits, not to better content but to find dirt on me. A perfect example of this is your bullshit removal of my rights a day and a half ago, even though it was blatantly clear that the punishment didn't fit the crime, and you did not have a community consensus to do so. So, I ask you once more: will you voluntarily stop interacting with me for the good of the community? Because it's crystal clear that you continuing to interact with me is unproductive. Purplebackpack89 22:21, 13 August 2014 (UTC)
Three entries? Whoop-de-doo. Wonderfool makes more in 15 minutes. What does it prove? "You have admitted to monitoring my edits, not to better content but to find dirt on me" — show a quotation of me saying that. Keφr 22:49, 13 August 2014 (UTC)
In the only discussion I've only had with you on your talk page. If you were merely looking at edits for errors, it wouldn't matter who created them. I take it you're not going to agree to stop monitoring my edits, nor to stop using your tools in a way that results in a wheel war? What a shame. I thought you'd be the bigger man. Purplebackpack89 22:54, 13 August 2014 (UTC)
I asked for a quotation. If you fail at reading comprehension, you should not participate in writing a dictionary. Keφr 22:58, 13 August 2014 (UTC)
Your quote "Also, stop randomly jumping accounts for no reason, it makes it harder to review your "contributions", for lack of a better word." would indicate that you have been "reviewing" my contributions. But quotation, schmotation. I ask you once more: are you going to stop doing it? Purplebackpack89 23:01, 13 August 2014 (UTC)
I noticed two different accounts using the same signature at random. Did I "admit to monitoring [your] edits, not to better content but to find dirt on [you]" as you claimed? No. Lying illiterate troll. Keφr 23:21, 13 August 2014 (UTC)
The Wiktionary community has the right to review any edits anyone makes; this is a public project. I would be happy if someone were reviewing my edits, because any mistakes I make would be fixed. The quote about PBP from his ban on the Simple English Wikipedia (which Kephir linked to above) that I think basically sums him up very well is "it's clear that he cannot collaborate in a constructive fashion". Even though his mainspace edits may be productive, he is incapable of dealing with criticism. --WikiTiki89 23:39, 13 August 2014 (UTC)
I'm supposed to let a wheel war over my permissions and being called a "lying illiterate troll" roll off my back? I'm not the problem here, Kephir is. Kephir cannot collaborate in that he continues to call me a troll even after I've asked him to stop interacting with me altogether. Purplebackpack89 23:45, 13 August 2014 (UTC)

User:Kephir[edit]

@Purplebackpack89: If you continue to target other users in the BP, I will block you. --WikiTiki89 23:43, 13 August 2014 (UTC)

PBP, if you continue to comment here, I would too. Wyang (talk) 23:46, 13 August 2014 (UTC)
You're going to block ME because another editor attacked me? That seems unfair. Purplebackpack89 23:49, 13 August 2014 (UTC)
Why don't you scroll up and look at who is starting all of the antagonism in these discussions (hint: it's you). --WikiTiki89 23:52, 13 August 2014 (UTC)
Except I wasn't talking to Kephir, and he went in there, and called me a troll. Heck, the Mglovesfun comment wasn't even altogether serious, and that was blatently obvious! I can't believe you think that it's OK for Kephir to do that, or to wheel war over my permissions. He acts abominably towards me, and it has got to stop immediately. Purplebackpack89 23:54, 13 August 2014 (UTC)
Please don't propose ridiculous "interaction bans", and I will not have to strike them out. Notice I said "discussions" in the plural. I don't think anyone here got the humor in your remark about MG. --WikiTiki89 23:59, 13 August 2014 (UTC)

This Purplebackpack89 is very similar to User:Razorflame — no substance, disruptive, irreverent, spoilt brat. And what do they have in common? They are both American children. I blame the American school system. No European or Asian youth would behave like this. --Vahag (talk) 07:00, 14 August 2014 (UTC)

I think that's not a good generalisation, Vahagn. It won't get us anywhere. There are good and bad, well-educated and spoilt people everywhere. (No comment on the topic at hand). --Anatoli T. (обсудить/вклад) 07:08, 14 August 2014 (UTC)
This is exactly what's wrong with liberals like you, Anatoli — false equivalency. "All religions are peaceful", "all cultures are good", etc... It is only America that instills in its children an undeserved sense of specialness and entitlement. You know very well that in our societies someone like Purplebackpack89 would immediately eat a couple of slaps in the face and wouldn't dare raise voice against such a valuable editor as Kephir. --Vahag (talk) 10:20, 14 August 2014 (UTC)
First of all, I'm not defending Purplebackpack89, I'm just saying that it is wrong to judge people by their origin. Using slaps in the face has little to do with being liberal, use of force is often justified. Skinheads in Russia killed hundreds of people from the Caucasus (including Armenians) and Central Asia, just for having the wrong looks and accents, for being generalised as "less civilised" or having bad behaviour or upbringing. I'm not always justifying American politics either but it's Russia that now instills in its children an undeserved sense of specialness and entitlement. Russia may treat Armenia better than other neighbours but don't be fooled, Putinism in this stage can't have real friends, allies or simply partners, it can only have vassals or enemies. Real liberal societies, which you dislike, give a chance to everyone, regardless of where they come from and a slap of face gets the one who deserves it, no matter where they come from. I work and socialise with people of different races and colours in the country, which treats everyone equally, more than US or Russia, and I don't see any problem with that. You would think the same way if you lived in a friendlier environment. I don't blame you for your views. I don't want to be involved in political debates or discuss, which nation is better, just replying to your comment. "No European or Asian youth would behave like this". This is funny really, people behave badly "in our societies", even if they are beaten. I am actually having trouble finding civilised and open conversations in runet (Russian Internet), besides it's problematic to punish someone on the Internet for bad behaviour. --Anatoli T. (обсудить/вклад) 06:02, 15 August 2014 (UTC)
You never get my trolling, Anatoli... And by the way, regarding modern Russia — I'm probably the biggest Russophobe you know. --Vahag (talk) 06:22, 15 August 2014 (UTC)
Well, в каждой шутке есть доля ... шутки. I do get your trolling but not always, I admit. Your jokes get you into trouble, so I'm not the only one who doesn't always get your trolling. :) I am a Russian Russophobe, as far as the Russian politics go, even if I was born in the Eastern Ukraine and lived in Russia most of my life. --Anatoli T. (обсудить/вклад) 06:38, 15 August 2014 (UTC)
@Vahagn Petrosyan:, you've crossed the line with your comments. For one, it's a personal attack to call somebody a "spoilt brat". It's also inaccurate to say I'm without substance, as I have created over 100 entries. Purplebackpack89 13:32, 14 August 2014 (UTC)
Purpleback reminds me a little of Wonderfool too. Socially naive, mostly well-meaning, and occasionally a genuine asshole. Definitely could do with some more mainspace edits tho. --ElisaVan (talk) 01:31, 15 August 2014 (UTC)
Just a note, and sorry for this being a few months late, but I am not this user. Just thought I'd clear that up. Razorflame 17:48, 22 October 2014 (UTC)
Good to know. Keφr 18:05, 22 October 2014 (UTC)

Personal word list/vocabulary list ?[edit]

I use Wiktionary as my learning tool for learning foreign languages.
I am using the "Watchlist" feature as my word list but I think we need a proper personal word list.
If people could create a word list and then choose to add words that they looked up to the list, it would help them remember those words better. Burkhankhaldun (talk) 07:17, 14 August 2014 (UTC)

Hit Ctrl-D. (This is not a prank. That would be Ctrl-W.) Keφr 07:34, 14 August 2014 (UTC)
You could simply edit your own user page and add the words of interest there; it would also allow grouping and formatting using the wiki markup. Equinox 07:44, 14 August 2014 (UTC)
What Equinox said. :) If you click the button at the top of the edit window that says "advanced", the icon on the far right of the menu that appears will even help you add a sortable table if you want to put in both foreign-language words and translations, and be able to sort them. Cheers! - -sche (discuss) 16:38, 14 August 2014 (UTC)
Userspace should be user for the activity related to the improvement of project, and not as a personal diary or a learning tool. That kind of functionality is outside the purpose of this project. --Ivan Štambuk (talk) 14:55, 16 August 2014 (UTC)

Not renaming template pedialite[edit]

I have only now noticed that "pedialite" is being renamed in the mainspace to "projectlink|wikipedia", like in diff. I object. Let me also repeat my sense of exasperation about the perpetrator of that renaming. This is very angering, and I feel helpless. I don't think I would quit Wiktionary about this, since Wiktionary is too great a project regardless, but the exasperation does move me in that direction. --Dan Polansky (talk) 09:20, 14 August 2014 (UTC)

Short markup is highly beneficial to editors who actually create entries. (I suppose it makes no odds to those who primarily work on templates. Ha.) Every time I have to start using {{cx|cookery|lang=en}} instead of {{cookery}}, or {{l|en|-mone}} instead of [[-mone]] (or have to dig my edit cursor through masses of such code generated by others), I feel a similar frustration to Dan's. I don't care about the final underlying representation — it could be a huge complex XML document — but the stuff I type in a text box, including others' work as presented to me for ongoing editing, should omit the complexities. Editing facilities do not seem to be keeping up with internal changes. Equinox 09:28, 14 August 2014 (UTC)
Amen. Content seems to be taking a back seat to a single individual's urge to "tidy". DCDuring TALK 12:21, 14 August 2014 (UTC)
And yet, when things are confusing, people wonder why we don't attract many new editors. My efforts to increase consistency and tidiness is aimed at reducing the substantial mental load that comes with editing Wiktionary, so that it becomes more accessible to newcomers. Wiktionary is too strongly biased towards its existing user base. —CodeCat 13:03, 14 August 2014 (UTC)
  • I object as well. The confusing thing is that Template:pedia and Template:pedialite are redirects to something, rather than being templates themselves. I seriously doubt that any editor would understand why projectlink exists at all. The reason that Wiktionary is too strongly biased to its existing user base is an over-reliance on an increasingly few templates, and a lack of redundancy in templates. Purplebackpack89 13:22, 14 August 2014 (UTC)
    • I see it as the opposite. I see having too many templates that perform similar functions as the problem. Redundancy should be eliminated, not increased. The more templates are alike, and the less different kinds with subtle differences, the easier it is for new editors to learn them. But to address the point about pedia and pedialite specifically, I was not going to delete them. I was only converting them to something equivalent. After all, I've seen other people's bots convert {{cx}} to {{context}} which is no different. It's just eliminating the shortcut, as shortcuts are intended primarily to give editors less to type. —CodeCat 14:25, 14 August 2014 (UTC)
  • This is a common tactic: first a "redundant" template is orphaned by bots, and afterwards they are listed for deletion on the obscure template discussion board which barely anyone keeps track of, and after a discussion involving one or two editors they get summarily deleted. You take a 1 month wikibreak and suddenly templates that were stable for years, had short and easily recallable names are all gone and replaced with some verbose monstrosities. --Ivan Štambuk (talk) 14:50, 14 August 2014 (UTC)
  • After a fortuitous e/c that illustrates my point.
    @Purplebackpack: You seem to misunderstand. You should know that the only important confusion is that faced by those who control the system, not any confusion on the part of actual users or contributors. That confusion is asserted to affect new contributors, though a moment's thought would clearly show the assertion to be implausible.
Most new potential contributors are likely to be contributing in one or two languages, their native language or their native language plus English. Any uniformity of system design across languages makes no difference to them. Redirecting templates are the means of getting the best out of our technical infrastructure while allowing customization for individual languages.
In the case at hand, we have redirecting templates applied to a different class of items. It would be easy enough to address the naming inconsistencies among our project-linking templates. The problem is that CodeCat/Mewbot fails to listen to or heed objections, let alone seek input in advance. I can only conclude based on the consistent pattern of behavior that CodeCat can't handle disagreement and avoids it by doing elaborate endruns and using technical malarkey to put an end to any discussion that might lead to a frustration of ambitions or whatever. I suppose that we can expect various neuroses (or worse) among contributors to become evident over time. It is only when the resulting behavior causes problems or inconvenience for others that we are entitled to object. This seems to be one of those times. DCDuring TALK 15:03, 14 August 2014 (UTC)
I see value in merging templates in the backend, but there is no reason that {{pedialite}} can't be kept as a redirect. I also see no value in mass-converting uses of this template. --WikiTiki89 15:11, 14 August 2014 (UTC)
I agree with Wikitiki on all points. Improving the backend of the template is good. The (semi-)memorable name ({{pedialite}}) should be kept as a redirect, and RFDed if someone wants it to be deleted. Renaming existing usages is not harmful, but it's not helpful or necessary, either. (Ditto renaming existing usages of {{cx}}, as Mglovesfun did.) - -sche (discuss) 18:36, 14 August 2014 (UTC)
I noticed that DP is now running an unapproved bot making undiscussed edits to many pages. I find it ironic that in trying to undo what he considers my wrongs, he commits those very same wrongs himself. Apparently the rules don't apply, and there's no need to discuss anything, if you already know you're right? This kettle disapproves of the pot. —CodeCat 20:35, 22 August 2014 (UTC)
A neat trick of yours. It shows your modus operandi in the clearest. I am merely restoring the state before your undiscussed changes. I do not need to actively seek consensus to restore status quo ante. --Dan Polansky (talk) 20:42, 22 August 2014 (UTC)
Thank you for confirming that you did not intend to discuss your edits or your bot with anyone. —CodeCat 20:44, 22 August 2014 (UTC)
Again, I am not instating a change; I am abolishing a change. This very thread shows the degree of consensus or its lack for the change that I am abolishing. --Dan Polansky (talk) 20:59, 22 August 2014 (UTC)
Can you show the vote or other kind of discussion that demonstrates that there is consensus for mass-reverting other editors without discussion and with an unapproved bot? —CodeCat 21:10, 22 August 2014 (UTC)
I recall instances unilateral reversal of changes to individual entries to the status quo ante. But I don't see any particular reason why the same should not apply to multiple bot-installed items. I see no particular reason why any single person's unvoted-on changes deserves any special protection from reversion or reversal. In the case of someone who simply institutes changes without any consensus, motivated principally by a purely personal compulsion to tidy, and then leaves to others the task of completing the change (Redlinked categories come to mind.), reversion would seem to be warranted and a failure to do so to reward bad behavior. DCDuring TALK 21:57, 22 August 2014 (UTC)
No, I cannot show you the vote showing that this wiki should be governed by consensus; that I accepted as a given when I have joined the project. Re: "Unapproved bot": this is AWB from a menial-work user controlled by me, not a bot. --Dan Polansky (talk) 22:10, 22 August 2014 (UTC)

Debotting MewBot[edit]

FYI: Wiktionary:Votes/2014-08/Debotting MewBot. --Dan Polansky (talk) 09:34, 14 August 2014 (UTC)

Inupiak[edit]

How do I go about rendering Inupiak entries when a specific font is required to view the intended characters? When I copy and paste words from the original typeface into wiki entries, they are rendered with completely wrong characters. I don't want to create entries that aren't accurate, so it's probably best if I wait for a solution before creating more work for myself by having to change them later on! The font can be found hereJakeybeanTALK 17:53, 14 August 2014 (UTC)

Use the proper Unicode characters: Ġ, ġ, Ḷ, ḷ, Ł, ł, Ł̣, ł̣, Ñ, ñ, Ŋ, ŋ. — Ungoliant (falai) 17:59, 14 August 2014 (UTC)
Brilliant, thank you. —JakeybeanTALK 18:29, 14 August 2014 (UTC)

Empowering JackBot[edit]

FYI: Wiktionary:Votes/bt-2014-08/User:JackBot for bot status. --Dan Polansky (talk) 20:33, 14 August 2014 (UTC)

Old Provençal or Old Occitan?[edit]

Wiktionary knows of a language called "Old Provençal", which nowadays is normally termed "Old Occitan". Should it be renamed? Benwing (talk) 04:38, 15 August 2014 (UTC)

SupportCodeCat 13:03, 15 August 2014 (UTC)
@CodeCat: Since you repeatedly complained of people posting no rationales, do you have any? --Dan Polansky (talk) 13:47, 15 August 2014 (UTC)
I support the rationale that Benwing gave. —CodeCat 13:49, 15 August 2014 (UTC)
@CodeCat: Thank you. How can we verify that "Old Provençal" is nowadays normally termed "Old Occitan"? --Dan Polansky (talk) 13:51, 15 August 2014 (UTC)
My book "Introduction to Old Occitan" says re. Occitan vs. Provençal:
Occitan enjoys increasing acceptance in all the languages of scholarship on the subject (despite the resistance of Provençal partisans) [with a footnote here] and will be adopted here.
The footnote says
"Only among specialists outside France has Occitan come to be the generally accepted term for the language" (Field 233).
Since we are "outside France" then we should use Old Occitan. Note that we're already using Occitan for the modern language, since Provençal is properly speaking only one of the dialects of Occitan. Benwing (talk) 16:17, 15 August 2014 (UTC)
Thank you very much. I have now also looked at Old Provençal,Old Occitan at Google Ngram Viewer and W:Old Occitan, and support. I've seen User:Renard Migrant edit Occitan, so I am pinging him, in case he has input. --Dan Polansky (talk) 16:48, 15 August 2014 (UTC)
Support as well. --WikiTiki89 17:14, 15 August 2014 (UTC)

Removing the number sign (#) from voting templates[edit]

I propose to remove the automatic number sign (#) from the voting templates {{support}}, {{oppose}}, and others. There has been discussion about this before (can someone find where?) and I recall that the objections were that people might not precede the template with a number sign, but I from looking at our votes, all of the uses I see have a number sign before the template. This will solve those annoying indentation problems with these templates, by putting all of the indentation control outside of the templates.

To clarify, this means that this will still work:

# {{support}} ~~~~

This will no longer work:

{{support}} ~~~~

But this will now work as expected:

: {{support}} ~~~~

As part of updating these templates, we can even merge the backends so that all of the voting templates will function the same way (such as {{vote delete}}).

--WikiTiki89 13:02, 15 August 2014 (UTC)

  • Support. I feel the templates should not generate the number, and should not be used like they are used in the second option above. It seems to me that the # sign is not part of the support itself, unlike the support icon and the text "support". And someone may want to write # Weak {{support}} ~~~~, and this should work seamlessly. --Dan Polansky (talk) 13:49, 15 August 2014 (UTC)
Support per nom and per Dan's point that the current format makes it hard to cast qualified votes like "weak support". A short previous discussion was here (prompted by edits to Template:vote-generic and other vote-setup templates). - -sche (discuss) 18:40, 15 August 2014 (UTC)

Additions/changes to Template:en-verb[edit]

Recently I converted this template to Lua, but without changing how its parameters work at all. But with Lua I think we can streamline it some and hopefully make it easier to use. The way I propose is as follows:

  1. For completely regular verbs, which add -s, -ing and -ed to the page name, nothing changes. You just specify no parameters at all.
  2. For slightly irregular verbs, you can specify only the first parameter (this is new):
    1. If the first parameter equals "es", then the present 3rd singular gets that ending instead of the normal -s. (example: smash {{en-verb|es}})
    2. If it equals "ies", then the final -y of the page name is replaced by -ies in the present 3rd singular, and by -ied in the past, while the present participle will end in -ying. (example: carry {{en-verb|ies}})
    3. If it equals "d", then the past ending will be that instead of the normal -ed. (example: free {{en-verb|d}})
    4. If it is a word ending in "es", then this ending is replaced with -ing and -ed to form the present participle and past. (example: recognize {{en-verb|recognizes}})
    5. If the first parameter anything else, then it is taken as the stem to form the present participle and past. (example: plot {{en-verb|plott}})
  3. If the second parameter and possibly third and fourth parameter are present, the template works as before. This is done for backwards compatibility. In particular, you can specify the present, present participle, past and optionally past participle directly using the four positional parameters.

CodeCat 13:32, 15 August 2014 (UTC)

I agree with #1 and #3, but I think many of the things in #2 can be automated. #2.1, #2.2, #2.3, and #2.4 can all be detected automatically (in case of false positives, we can use {{en-verb|s}} to provide the default behavior). For #2.5, I think we can do something like {{en-verb|tt}}, {{en-verb|dd}}, etc. --WikiTiki89 14:15, 15 August 2014 (UTC)


Why not also add another parameter in the first place, which would be a number denoting the verb's position. So for example carry out would be like
{{en-verb|1|ies}}.
Programmatically, It will first check if the first argument is a number and treat 2nd 3rd and other arguments as if they were 1st 2nd and so on (the arguments meaning don't change).
Otherwise it will behave just like it does now.--Dixtosa (talk) 14:17, 15 August 2014 (UTC)
That is a good idea, but I think it would be better of as a positional parameter: {{en-verb|p=1}}. --WikiTiki89 14:19, 15 August 2014 (UTC)
I didn't want to make the proposal too complicated technically speaking. We could still do that in a later proposal, but for now I'd rather focus on what is there first. —CodeCat 14:40, 15 August 2014 (UTC)
The technical side doesn't matter in the proposal. My suggestions simplify the interface that editors will use, by not requiring any arguments in most cases. --WikiTiki89 14:55, 15 August 2014 (UTC)
I'm more wary of making proposals that make too many changes. I've noticed in the past some people don't like changes, so I've tried to keep them to a minimum. —CodeCat 14:56, 15 August 2014 (UTC)
What you should have noticed was that people don't care about the internal changes, but about the external changes. And my suggestions will have fewer external changes. People tend to complain when you require extra parameters or change the names of templates or parameters, but not when the template is updated to do all the work for them, while supported backwards compatibility. --WikiTiki89 15:00, 15 August 2014 (UTC)
What Wikitiki said. DCDuring TALK 15:48, 15 August 2014 (UTC)
I agree with Wikitiki about automating the things in #2. Benwing (talk) 16:23, 15 August 2014 (UTC)
How will this handle the existing cases such as tie#Verb when used in accordance with the current documentation, ie, {{en-verb|t|y|ing}}? Presumably per 3 above? DCDuring TALK 18:04, 15 August 2014 (UTC)
Yes, it looks like that is covered by #3, although ideally there should be a shortcut for it. --WikiTiki89 18:17, 15 August 2014 (UTC)
Agreed: "What you should have noticed was that people don't care about the internal changes [changes to template and module internals], but about the external changes [changes to wiki markup]." --Dan Polansky (talk) 09:15, 23 August 2014 (UTC)
@CodeCat: I have blocked User:MewBot for going ahead with this without consensus. --WikiTiki89 23:53, 24 August 2014 (UTC)
I see a consensus for it here. —CodeCat 23:53, 24 August 2014 (UTC)
I see a few people agreeing with me about my proposed changes to your changes. I also see no mention of a bot run in this discussion. --WikiTiki89 23:56, 24 August 2014 (UTC)
Your proposals were all additions to mine. I've just implemented a subset, foregoing the autodetection part. And while there is no mention of a bot run, why would there be opposition to one if we already agreed on what the new parameters are? —CodeCat 00:00, 25 August 2014 (UTC)
One reason is that if we do implement the auto-detection part, then you would have to do another bot run on the same verbs. Another reason, is that you need to start following bot procedure more closely. I would have thought that while there is a vote going on to debot your bot, you would be on your best behavior. --WikiTiki89 00:08, 25 August 2014 (UTC)
I thought I was? That's why I'm confused, I really thought I was finally doing something nobody would find a reason to be upset at me about. Maybe my bot should be blocked then, it seems I'm not good enough at judging when it's ok to use it. —CodeCat 00:13, 25 August 2014 (UTC)
You shouldn't be guessing at what people would be upset about. All you need to do is ask before every bot run. It's a simple enough rule to follow. --WikiTiki89 00:17, 25 August 2014 (UTC)
People never give a straight answer though. Just look at what happened here; I thought there was assent, when you judged it differently. I don't deal well with trying to figure out just what people mean, and clearly when I try to make sense in interpreting all the different comments and opinions, people get upset because I inevitably get it wrong. —CodeCat 00:22, 25 August 2014 (UTC)
You have to ask the question before you complain about getting unclear answers. After you were done making the changes in the module, you should have come back to this thread and asked something like "Can I start the bot run now?" --WikiTiki89 00:27, 25 August 2014 (UTC)
The module/template is specially designed to make all the parameters spell one of the inflected form, just as how I designed {{de-conj-auto}} to make all the parameters spell the pagename (advertisement!). For example, criticize would have {{en-verb|criticiz|es}} for aesthetics purposes. And talking about unclear answers, would you like people to just oppose you if they don't agree with you? Isn't that a bit rude? I oppose criticize having {{en-verb|criticiz}} as parameter. --kc_kennylau (talk) 01:49, 25 August 2014 (UTC)
That is one of the reasons I want auto-detection. I think we can all agree that just plain {{en-verb}} is much nicer than either of {{en-verb|criticiz|es}} and {{en-verb|criticiz}}. --WikiTiki89 02:00, 25 August 2014 (UTC)
I think the second is better than the first, because the second doesn't use a superfluous parameter. The "es" is literally just a no-op, like if you wrote {{en-verb|criticiz|lang=en}}, since the template doesn't need or use a lang= parameter. —CodeCat 02:02, 25 August 2014 (UTC)
What Kenny is saying is that criticiz looks very wrong by itself. And speaking of superfluous parameters, the entire word criticiz is superfluous as well, since it can be deduced from the page name. --WikiTiki89 02:06, 25 August 2014 (UTC)
Not any wronger than some of the parameters you might find in other templates. For the inflection table of Finnish suomalainen for example you write {{fi-decl-nainen|suomalai|a}}. If you assume that parameters should look like natural words, then yes, the parameters look strange. But that's a wrong assumption. And while the parameter can indeed be deduced from the page name, we haven't yet determined how many cases there are for which that deduction gives the wrong results. That's one of the reasons I preferred to play it safe at first, given that this is such a widely used template and we can't afford errors. Once the proposed changes had been made, I was going to research how feasible your additions would be. But I never got that far now... —CodeCat 02:16, 25 August 2014 (UTC)
If there is a choice between using a parameter that looks natural or one that looks unnatural, you should go with the natural one. As for the reliability, of course if we haven't discussed it, we couldn't have determined anything. I wouldn't call going ahead with a bot run "playing it safe". --WikiTiki89 02:25, 25 August 2014 (UTC)
What my bot did was apply the more limited version, the proposal I originally made. That I had already done a lot of researching and experimenting with, so I already knew that it would be possible to implement it fully as proposed, before I even proposed anything. I wanted to make sure that I wouldn't be proposing something that ended up not being feasible. —CodeCat 02:41, 25 August 2014 (UTC)
It is of course feasible, but the aestheticness is lost. --kc_kennylau (talk) 03:30, 25 August 2014 (UTC)

The meaning of "word"[edit]

We have had some disagreements of late as to what constitutes a "word". Our corpus offers a few relevant definitions, including:

  • A distinct unit of language (sounds in speech or written letters) with a particular meaning, composed of one or more morphemes, and also of one or more phonemes that determine its sound pattern.
  • A distinct unit of language which is approved by some authority.
  • Any sequence of letters or characters considered as a discrete entity.
  • Different symbols, written or spoken, arranged together in a unique sequence that approximates a thought in a person's mind.

So what is a "word" so far as "all words in all languages" is concerned? bd2412 T 18:33, 15 August 2014 (UTC)

There is that old classic definition, as MWOnline puts it: "any segment of written or printed discourse ordinarily appearing between spaces or between a space and a punctuation mark". This definition has the virtue of fitting with the use of written documents for attestation, which virtue is not shared by any of the four above. Of course, it is not the only definition that might have that virtue. DCDuring TALK 20:02, 15 August 2014 (UTC)
It also has the drawback of not working for languages that aren't written with spaces, or aren't written at all. —Aɴɢʀ (talk) 20:04, 15 August 2014 (UTC)
And, of course, MMWOnline has 10 main senses, 20 individual definitions of word, of which two senses, four definitions (including that above), may be relevant to (y)our discussion. DCDuring TALK 21:52, 15 August 2014 (UTC)
Several of the recent discussions about what constitutes a word were specifically about whether romanizations were words. On that subject, I've opined that romanizations are not words but representations of written words, like the shadows of things in Plato's Cave. Cambridge’s definition is interesting to consider in this context; it says a word is "a single unit of language that has meaning and can be spoken or written" — as if in their view words themselves are concepts, like the Platonic concept of jar, and spoken and written forms are just instances, like actual jars (and then, in my analysis, romanizations are the shadows of the jars). Cambridge's definition also implies that words have to belong to languages.
Even if one doesn't share my view of romanizations, it may be difficult to write a definition of "word" that applies to all words, doesn't apply to anything other than words, and yet isn't a paragraph long. I'll think about it, but for now, to add to the above list of other references' definitions, Wikipedia defines word as "the smallest element that may be uttered in isolation with semantic or pragmatic content (with literal or practical meaning) [...or more concisely...] the smallest meaningful unit of speech that can stand by [itself]".
- -sche (discuss) 21:20, 15 August 2014 (UTC)
I would actually go further and say that written language isn't language at all but only the representation of language, just like a painting of a pipe isn't a pipe. —Aɴɢʀ (talk) 22:10, 15 August 2014 (UTC)
I would have to disagree with you there. I would say that written languages and spoken languages are different languages using different media, but heavily influenced by each other. --WikiTiki89 22:21, 15 August 2014 (UTC)
It would be fascinating if we had a dictionary of sounds, where the "reader" could speak the sound and in response be told the meaning of it, with no visual symbols being used at all. bd2412 T 23:01, 15 August 2014 (UTC)
An IPA dictionary would be the next closest thing. I'd love to work on that, but it would be hard. —CodeCat 00:50, 16 August 2014 (UTC)
  • From MW Online:
  • a - (1) : a speech sound or series of speech sounds that symbolizes and communicates a meaning usually without being divisible into smaller units capable of independent use
    (2) : the entire set of linguistic forms produced by combining a single base with various inflectional elements without change in the part of speech elements
  • b - (1) : a written or printed character or combination of characters representing a spoken word <the number of words to a line> []
    (2) : any segment of written or printed discourse ordinarily appearing between spaces or between a space and a punctuation mark
  • -- HTH DCDuring TALK 21:52, 15 August 2014 (UTC)
    I notice that all of the definitions we've discussed so far fail to cover words from sign languages that lack written forms. Many of the above-mentioned definitions (e.g. Cambridge's) do cover words from sign languages like ASL, DGS, etc, because words from those languages can be written (using SignWriting, HamNoSys, etc), but not all sign languages can be written. - -sche (discuss) 22:36, 15 August 2014 (UTC) clarified 03:33, 16 August 2014 (UTC)
    There's another question raised by that - do we need to pick one definition for "word" or should we, in an abundance of caution, include within our sweep several different formulations? Should we include every distinct unit of sounds in speech or written letters with a particular meaning, plus every distinct unit of language which is approved by some authority, plus every sequence of letters or characters considered as a discrete entity? bd2412 T 00:45, 16 August 2014 (UTC)
    That's a good question. My initial reaction when I thought of sign languages was that it would make sense to have different definitions for (1) spoken/written languages' words and (2) sign languages' words, or even (1) spoken languages' words, (2) written languages' words and (3) sign languages' words. But it seems to me that there's a basic concept behind all of those, a basic sense of "word"—if we can figure out how to formulate it—that "spoken word", "written word" and "signed word" are subsenses of. (Does it seem that way to you?) Other senses like "unit of language approved by some authority" — which I guess is the sense people use when they say "irregardless isn't a word"? — could either be additional senses on the same level as that basic sense, or subsenses of it.
    For a definition of the basic sense, what about "the smallest unit of language which has meaning and can be expressed by itself"? Would that include or omit anything it shouldn't, bearing in mind that I'd envision it having subsenses that would provide the details on the nature of written / spoken / signed words? - -sche (discuss) 03:33, 16 August 2014 (UTC)
  • An interesting issue is caused by clitics. English has a clitic in the 's ending, but some languages have a lot of them. For example, Arabic has clitic object and possessive pronouns, plus clitic prefixes like wa- "and", fa- "so", ka- "like", sa- (future tense), etc., and many Arabic dialects have a clitic negative circumfix ma- ... . In some sense, adding a clitic to a word makes a larger word rather than two joined words; that's the nature of a clitic. At the very least, the resulting entity behaves as a single phonological word, usually with a single primary stress and possibly secondary stress(es). But it's possibly not a single linguistic word (morphological word?) in that it's not the "smallest unit of language which has meaning and can be expressed by itself" -- that would be the part without any clitics added. I would say there should be a general rule that word+clitic combinations should not be entered into Wiktionary; I think there are already agreements of this sort in specific cases, e.g. the -que, -ve, -ne clitics in Latin.
  • Compound words are also a problem since they seem to also violate the "the smallest unit of language which has meaning and can be expressed by itself" criterion but often have an idiosyncratic meaning, e.g. blackboard and blackbird are not the same as black board and black bird. Sometimes in English we write such compound words with no spaces, but not consistently cf. red tape, red-eye/redeye/red eye (flight), data base/database, file name/filename, etc. In Mandarin the issue comes up even more acutely since most words are compounds of one sort or another and there are widely varying degrees of compositionality of meaning, phonological behavior as one word or several, etc. Benwing (talk) 10:03, 16 August 2014 (UTC)
  • So I guess that we are considering abandoning the notion of the lexicon in favor of our slogan. I suppose that each word in our slogan should be similarly parsed and that any limits of our user interface be ignored in pursuit of an ideal that won't be realized before Wiktionary collapses under the burden of idealism. DCDuring TALK 12:27, 16 August 2014 (UTC)
    We abandoned the notion of the lexicon (a collection of lexemes) right at the very beginning, when we decided to have individual entries (and in some cases citations and the like) for plural forms like peaches, conjugated forms like stumbled, and superlative forms like hungriest. Your fear of the "limits of our interface" is interesting, given that the entirety of Wiktionary can fit on some USB thumb drives, and that our sister project Wikipedia is showing no signs of such problems despite having over 33 million pages across all namespaces with no signs of slowing, compared to our 4.1 million. I'd say we have plenty of room to grow. bd2412 T 14:19, 16 August 2014 (UTC)
    The limits are only that we are stuck with screen output and keyboard input. Other forms of input and output are much more limited in the portion of our content that they can accommodate. DCDuring TALK 15:06, 16 August 2014 (UTC)
    Sorry, I misinterpreted your actual shortsightedness for an altogether different kind of shortsightedness. My wife asks her phone things all the time and gets answers (often from Wikipedia). Quite a few technologies exist to allow blind people to use the Internet, generally. The limitation on our interface is that Wikimedia has not yet initiated a vocal interface. bd2412 T 15:28, 16 August 2014 (UTC)
    I am very happy to leave farsightedness and idealism to those who enjoy living in fantasy worlds. Judging by the way speech recognition works as delivered by Google and Apple and its rate of progress to date, I would say that it has no relevance whatsoever for casual users of Wiktionary for the next ten years or more. Of course, hard-core user/contributors such as ourselves may find some use for it sooner, though judging by the use of speech-recognition technology in workplace situations (ie, one language, one speaker, narrow range of vocabulary) much less challenging than ours and much more equipped with technical resources, even this may not turn out to be true. DCDuring TALK 17:34, 16 August 2014 (UTC)
    You should brush up on Moore's law. Also, web accessibility. Your ten-year projection may be a bit out of line with the current state of the art. I just tried my wife's phone with a few definitions of non-English words and it did pretty well. Of course, all of this is a separate discussion from what is a "word" for our purposes. bd2412 T 21:36, 16 August 2014 (UTC)
    Show me. DCDuring TALK 01:56, 17 August 2014 (UTC)
    How would I go about doing that? In any case, I'll concede that it's a bit beyond the scope of the question of how "word" is defined for purposes of writing the dictionary. bd2412 T 03:33, 17 August 2014 (UTC)
  • The question is too general and philosophical, and thus of no practical value. What is a word should be decided on an individual language basis, primarily on the basis of the criterion of usability: how the users and third-party software would look up Wiktionary entries using the search box/API to find out word's meaning and other metadata. The purpose of "all words in all languages" motto is not to impose exclusive inclusion of words (as opposed to non-words, a we already have countless entries for non-words), but rather to extol the liberal principles of the project, namely the absence of "authorities" which decides what goes in or not, as opposed to actual attestations of language which provide real-word evidence of usage. --Ivan Štambuk (talk) 14:27, 16 August 2014 (UTC)
    • The question may well be overly general and philosophical, but if we're in the business of offering people definitions of words, we should have a handle on what words are. bd2412 T 13:42, 18 August 2014 (UTC)
      We are in the business of collecting human knowledge, at a scale that vastly exceeds the needs 99% of people. People don't care whether something is a word or not when they look it up in the dictionary - they just want the meaning/translation and other goodies. Optimization of user interface for human consumption should be orthogonal to the underlying goal of documenting all instances of written (perhaps one day even spoken) language. The potentially harmful impact of overextending the formal definition of word to include the supposedly non-opaque compounds or set phrases, or "real" words with various affixes attached with debatable level of transparency is trivial. --Ivan Štambuk (talk) 13:45, 21 August 2014 (UTC)

I have overhauled our entry [[word]], adding several verb senses, adding more citations, etc. I left the four definitions which pertained to the linguistic sense of 'word' alone until the end of the overhaul, when I changed them like this (if it's too hard to pick through that diff, see this): I didn't reword the last three of the four senses at all (though I am about to RFV at least one of them), but I split the sense which had said
  1. A distinct unit of language (sounds in speech or written letters) with a particular meaning, composed of one or more morphemes, and also of one or more phonemes that determine its sound pattern. [from 10th c.]
into one sense and two subsenses with supporting citations:
  1. The smallest unit of language which has a particular meaning and can be expressed by itself; the smallest discrete, meaningful unit of language. (Contrast morpheme.) [from 10th c.]
    1. The smallest discrete unit of spoken language which has a particular meaning, composed of one or more phonemes and one or more morphemes.
    2. The smallest discrete unit of written language which has a particular meaning, composed of one or more letters or symbols and one or more morphemes.
Further improvements are welcome and indeed encouraged. - -sche (discuss) 03:42, 18 August 2014 (UTC)
Postscript: I modified another of the 'liguistic'-sense-related definitions, like this. - -sche (discuss) 05:06, 18 August 2014 (UTC)

Multiple Categories For 1 Label?[edit]

I run into items in Special:WantedCategories from time to time that that result from labels categorizing differently than contributors expect. For instance, kibutz is a Turkish word for an Israeli institution, so it only seemed logical to put "Israel" in the {{context}} template. This added the redlinked category Category:Israeli Turkish, because the system assumes Israel is where the term is spoken, rather than where the referent of the term is found. The same thing happens when someone uses classical to refer to something associated with ancient Greece and Rome: we have things set up to interpret that as referring to the classical stage of the language (this was done with Classical Chinese in mind), and thus we had the redlinked category Category:Classical English. Is there a way to have both a regional and a topical category for the same term, and to select between them? Or, lacking that, is there a way to turn off categorization for a single parameter in the {{context}} template, to have it display in the list, but not generating a bogus category? Chuck Entz (talk) 02:53, 16 August 2014 (UTC)

This has been raised and ignored before, but it might be the singer not the song. DCDuring TALK 02:56, 16 August 2014 (UTC)
I'd say the problem is more general in that we never had a proper way to distinguish regions as topics and as dialects. I think I brought this up before, when the label templates were being converted to Lua, but I think it was mostly ignored at the time like DCDuring said. —CodeCat 14:28, 16 August 2014 (UTC)
The problem isn't limited to regions. A region can be a usage context or a "topic", but so can a discipline, trade or profession. In addition, as Chuck points out, the categories are not necessarily intuitively connected to the label. Furthermore the mapping from labels to categories only has the sullen consent of users as it was basically imposed in the new regime. In the former regime there were many fewer topical categories, so much less often-inappropriate force-fitting was required and there could be a template for each major "context". These templates were apparently added based on a count of frequently used parameters of {{context}}, after some editing of user-inserted labels that were "close" to existing ones. The by-product was that there was an incentive for users to stick to the "contexts" for which templates existed because fewer keystrokes were required, but user flexibility was not otherwise discouraged. There were specific technical difficulties with template-only implementation of {{context}}, which difficulties made it an early target of Lua-based reform. However, there was a significant loss of capability in the simplistic implementation of the new approach. DCDuring TALK 14:53, 16 August 2014 (UTC)

where to put quotations when there are multiple spellings of a word?[edit]

Old French is notorious for having inconsistent spelling, e.g. standard arachier also appears as arrachier and arracher, among others. I have the latter two linked to the former as alternative forms, and the former lists the latter two as alternative forms. I have one quotation which would involves the form arrache, which is a form of both arrachier and arracher. Should I (a) put this quote under arrachier because that's the more standard spelling of the two, or (b) put it under arachier because it's under this lemma that all the variant forms are "gathered"? Is there a rule for this? Benwing (talk) 10:10, 16 August 2014 (UTC)

There's also: (c) put the quote under arrache, since it's specific to the spelling. Chuck Entz (talk) 17:34, 16 August 2014 (UTC)
Are you saying "the more standard spelling of the two" is not the entry that has been made the lemma? That should be fixed before anything else is done, and then the quotation should go in the lemma entry = entry for the most standard spelling (=, if I follow what you're saying, the or at least a spelling from which the form found in the quotation derives).
For English and German, I often see quotations of one spelling placed in the lemma entry even if it is a different spelling, especially if the quotations are famous uses (e.g. from Shakespeare), or contribute to showing that the word has been in continuous use for a long time. There are even quotations of Chaucer (Middle English) placed in English entries, which I have been told off for moving to Middle English entries. Quotations are also placed in the entries for the spellings they use, if that it necessary to verify the existence of those spellings. So, in the general case, it seems you can place the quotation wherever seems most appropriate. - -sche (discuss) 17:48, 16 August 2014 (UTC)
Sorry, let me clarify, arrachier is more standard than arracher but arachier is the most standard form. Both arracher and arrachier would produce the same 3rd-singular present indicative arrache. I fixed things so that both arrachier and arracher point to arachier. You (-sche) suggest putting it under arachier if I'm not mistaken. Chuck says put it under arrache but there's no entry at all for that. Benwing (talk) 19:10, 16 August 2014 (UTC)
I didn't say "put it under arrache", just that doing so should be included as one of the options. I would vote for either the lemma or the actual spelling, but I don't claim to be an expert on Wiktionary best practices. As -sche (who knows far more about this than I do) said, putting the quote in the actual spelling is mostly done when the quote is for the purpose of showing usage of the spelling itself, rather than of the term independent of its spelling. It might also be useful if you had a quote that was ambiguous enough so you weren't sure which lemma it belonged to.
In general, the issue of which form to put content on boils down to a tradeoff of usefulness vs. practical concerns: it would be most useful to have things like the definition at every spelling, but practically, that would be a nightmare to keep in synch, so we have everything in the lemma. Being consistent in format is just another aspect of usefulness: the more consistent things are, the less thought you have to put into finding things. Here, I don't see practical concerns unless you want to put it in multiple places, so you're better off figuring out where people would likely want or expect it to be, and putting it there. Chuck Entz (talk) 20:30, 16 August 2014 (UTC)

is there a policy statement somewhere indicating the standard order of "alternative forms", "etymology", "external links", etc. sections?[edit]

There seem to be standards for how these sections are ordered, but I'm not sure where this is documented. I'm guessing it's something like this:

At level 3:

  1. Alternative forms
  2. Etymology
  3. Verb / Noun / etc.
  4. External Links

At level 4, under the headword:

  1. Definition
  2. Pronunciation
  3. Conjugation / Declension
  4. Derived words, Related words
  5. Descendants
  6. Synonyms
  7. Anagrams

But not really sure.

Benwing (talk) 10:15, 16 August 2014 (UTC)

See WT:ELE. DCDuring TALK 10:52, 16 August 2014 (UTC)
Pronunciation is level 3, just after etymology. An exception is when there are multiple etymologies and they have different pronunciations. — Ungoliant (falai) 18:01, 16 August 2014 (UTC)
I'd say the exception is when there are multiple etymologies and they have the same pronunciation—in that case, Pronunciation precedes Etymology 1. —Aɴɢʀ (talk) 18:30, 16 August 2014 (UTC)
Yeah, you’re right. — Ungoliant (falai) 18:34, 16 August 2014 (UTC)

Catalan rhymes[edit]

Our current rhymes pages deal exclusively with central/eastern Catalan, which is the basis for the standard language in Catalonia itself. In these dialects, there is significant vowel reduction in unstressed syllables, so that many more words rhyme when they did not before. But in Valencia as well as large parts of western Catalonia, there is much less vowel reduction. The Balearic dialects take an intermediate position, reducing less than central Catalan but more than Valencian.

Because our rhymes pages use central Catalan as a base, they're no use for anyone outside of that dialect area. On the other hand, vowel reduction is entirely predictable, so information would be gained if we used unreduced vowels in rhymes, while none would be lost. So I'd like to propose that we do not reflect vowel reduction in Catalan rhymes, so that all dialects can be covered.

This does not solve all cross-dialect problems for Catalan, as Balearic dialects have an extra vowel phoneme that the other dialects lack, while Estern and Western Catalan differ in the application of é versus è. But it would be a step in the right direction at least. —CodeCat 19:51, 17 August 2014 (UTC)

If I understood it, this means basically to make rhymes in Western Catalan with regard to vowels (é/è is another problem). For example, Rhymes:Catalan/atʃus will be renamed to "Rhymes:Catalan/atʃos" indicating both pronunciations and Rhymes:Catalan/aɾə will be splited in "Rhymes:Catalan/aɾa" and "Rhymes:Catalan/aɾe" with a note redirecting to the other page for rhymes in Eastern Catalan. I think it is clear enough for a intermediate Catalan and it is useful for rhymes in Western Catalan or Valencian. --Vriullop (talk) 17:25, 28 September 2014 (UTC)
Yes, much like that. For Catalan as well as other languages, we already have notices on the rhyme pages, which say that the words on the page rhyme with the words on another rhymes page in certain dialects that merge or drop some sounds. The Central Catalan -r in infinitives is an example of this, English has rhotic and non-rhotic pages, and in Dutch there's final -n. In each case, we choose the rhyme that gives the most distinction. —CodeCat 17:33, 28 September 2014 (UTC)

Nynorsk translations and interwiki links[edit]

Hi,

As you know, the Bokmål translations can have an interwiki link to the Bokmål Wiktionary when the trnslation exists in this project. But for the Nynorsk translations, we don't have often this chance. Indeed, Nynorsk Wiktionary still exists currently but community seems not very active anymore (nn:Special:RecentChanges). Therefore, maybe we could route the « nn » code to « no » in Module:translations, like it’s done for « nb »; that is to say having an interwiki link to no.wiktionary for translations in Nynorsk, instead of having an interwiki link to nn.wiktionary.

Here is a previous discussion about the proposal to rally efforts on Bokmål Wiktionary: w:nn:Wikipedia:Samfunnshuset/Arkiv/2009#Wiktionary, et fellesnorsk prosjekt.

And you'll find here an analysis of Nynorsk translations on this project who whould have an interwiki link if there were routed to no:: User:Automatik/Analysis of Nynorsk translations.

What do you think of routing nn code to no for interwiki links in translations? — Automatik (talk) 00:23, 19 August 2014 (UTC)

Support. Before the Chinese merger we used to do this for Mandarin (cmn) terms, which linked to Chinese (zh) Wiktionary. BTW, Chinese merger is a success and helped increase contents for Chinese varieties (esp. Cantonese, Min Nan, Wu and Hakka), perhaps a good indication of what could have been done for Norwegian (Bokmål and Nynorsk) (vote didn't pass) and Arabic (no vote), even if the situation is not the same (e.g. Chinese lects have no inflection and PoS headers don't require genders or plural forms). --Anatoli T. (обсудить/вклад) 00:30, 19 August 2014 (UTC)
I think the question we need to ask ourselves first is what the purpose of interwiki links is in translations. Is the purpose to lead the user to more information about the term outside of what the English Wiktionary has available, regardless of what language it is presented in? Or is the purpose specifically to lead to the entry in that specific language? Or something else? If we just want to link to more information, then we really want lots of interwiki links, not just to one specific language. But if the idea is to match the language specifically, then we shouldn't be linking to Bokmål from Nynorsk translations. —CodeCat 13:20, 20 August 2014 (UTC)
The purpose is to link to a definition of the term in its own language. In this case, it's the same language. It may even make sense to link to all three of the Norwegian Wiktionaries (when the entries exist) for every Norwegian term. --WikiTiki89 14:09, 20 August 2014 (UTC)
Actually there are not three but two Norwegian Wiktionaries (no: and nn:). And 'no' stands for Bokmål even if in ISO it stands for Norwegian and 'nb' stands for Bokmål. — Automatik (talk) 21:08, 20 August 2014 (UTC)
Oh. In that case there is no question that, at the very least, we should we should point Bokmål translations to the no.wikt. I assumed there were three because Serbo-Croatian has four (Bosnian, Croatian, and Serbian, as well as the combined Serbo-Croatian). --WikiTiki89 21:22, 20 August 2014 (UTC)
Bokmål translations already points to no.wikt. I asked for Nynorsk translations because no.wikt is also written in Norwegian and is more active. But I can understand the reluctance of some (why do not point to nn.wikt if it's not closed and still exists?… for the reason explained before I guess, but I understand). — Automatik (talk) 22:14, 20 August 2014 (UTC)
Oh sorry, I completely understand now. In that case, I'm still going to say that all three of {{t+|no}}, {{t+|nn}} and {{t+|nb}} should point to both of no: and nn:. --WikiTiki89 00:15, 21 August 2014 (UTC)
But what should be done for the distinction between {{t}} and {{t+}} in this case? The template would not have any way to say that no: has an entry but nn: does not. —CodeCat 00:19, 21 August 2014 (UTC)
It would have to be a bit more complex. Something like {{t|nn|foo|iw=no,nn}}, {{t|nn|bar|iw=no}}. Each language would have to have a list of allowed interwikis. Not to mention User:Rukhabot would have to be updated (ping User:Ruakh). --WikiTiki89 00:35, 21 August 2014 (UTC)

Sorry I'm very late here. The no project isn't a Bokmål only project anymore. It encompasses both nn and nb. All words are (at least should be) clearly marked to denote if they are one or the other or both. So linking to the no project will yield the most up to date definitions and correct grammar etc. --Teodor (dc) 20:43, 22 September 2014 (UTC)

One could just as well argue that sv, da, nb and nn should link to whichever of them is the biggest, most active, with the best content or whatever at the time, since they are, linguistically, versions of the same language.

I could support having fallback links, like, for Nynorsk: 1. nb > 2. sv > 3. da. Similar arrangements for other languages (with the requirement that they are very closely related, to promote a linguistic rather than political and/or geographical link). --Njardarlogar (talk) 17:18, 23 September 2014 (UTC)

Cantonese translations[edit]

(Notifying Kc_kennylau, Wyang): The topic above gave me an idea that Cantonese (yue) translations could be linked to the Chinese (zh) Wiktionary. Although the Chinese Wiktionary doesn't always provide Cantonese transliterations or other info, it may be helpful to look up other things. Written Cantonese shares about 99% of terms with Mandarin. Perhaps Cantonese jyutping can be loaded with a bot here and Chinese Wiktionary, Wyang has created a framework for this. --Anatoli T. (обсудить/вклад) 00:38, 19 August 2014 (UTC)

@Atitarev: 係咪?(Do you mean this?) --kc_kennylau (talk) 00:41, 19 August 2014 (UTC)
@Kc kennylau: Thanks, Kenny. Well, yes, single character entries (zi) have jyutping but multi-characters often don't. Since Cantonese Wiktionary doesn't exist, it would probably make sense to use Chinese (for semantics, translations). -Anatoli T. (обсудить/вклад) 00:50, 19 August 2014 (UTC)
@Atitarev: I still don't understand what you mean. Can you provide me with an example? --kc_kennylau (talk) 00:56, 19 August 2014 (UTC)
@Kc kennylau: OK, Mandarin translation (at China#Translations) 中国 (zh) (Zhōngguó) links to zh:中国, even if it uses "cmn" language code but Cantonese 中国 (zung1 gwok3) links to nothing, uses {{t}}, not {{t+}}. I suggest to redirect yue to zh, the way cmn is used in translations and use {{t+}} when a Chinese entry exists in zh:wikt. --Anatoli T. (обсудить/вклад) 01:03, 19 August 2014 (UTC)
@Atitarev: Oh, I thought you were talking about {{zh-usex}}, that's why I didn't understand. --kc_kennylau (talk) 01:05, 19 August 2014 (UTC)
@Kc kennylau: Do you support this idea? Or you think zh:wikt shouldn't be used for Cantonese? --Anatoli T. (обсудить/вклад) 01:11, 19 August 2014 (UTC)
@Atitarev: Support. --kc_kennylau (talk) 14:30, 19 August 2014 (UTC)

Renaming existing uses of Template:term to Template:m[edit]

Recently I've noticed that User:Mulder1982 has been working through many entries and correcting and fixing etymologies, but he has also consistently replaced {{term}} with {{m}}. I fully support this action, but I think it would be more effective to use a bot to do this. Is it ok for me to run a bot to replace existing instances of {{term}} with {{m}}? —CodeCat 13:24, 20 August 2014 (UTC)

Yes, please do! It's hard work. :D But well, that has actually been a by-product. The main thing I've been doing these days is to remove redundant transliterations of Gothic. That could probably be done with bots too, I suppose. But yes, speaking for myself, do run the bot. Mulder1982 (talk) 13:27, 20 August 2014 (UTC)
I weakly support this and would point out that there is absolutely no urgency in the matter. --WikiTiki89 14:07, 20 August 2014 (UTC)
  • Oppose converting "{{term}}" to anything. Wiki markup is the user interface; a significant widespread conversion like this should be done via a vote, IMHO. See also Wiktionary:Beer parlour/2014/April#Convert Template:term.2Ft_and.2For_Template:term to Template:m.3F. By the way, where is the rationale? Moreover, Mulder1982 should not have been manually performing the conversion until they can demonstrate widespread support for such a conversion. --Dan Polansky (talk) 17:40, 20 August 2014 (UTC)
  • Support; at first I found it confusing that both {{term}} and {{m}} existed, since it wasn't obvious that they do the same thing. Being consistent would be less confusing for new users. Benwing (talk) 22:32, 20 August 2014 (UTC)
  • Support; I do this manually all the time. There is zero reason to use 10 characters (term|lang=) where 2 will do the same thing (m|). CodeCat, you mentioned once before that there are still many instances of {{term}} where no language is specified. I would say those should not be changed at all, rather than being changed to {{m|und}}; after all, probably over 95% of those cases are English anyway. —Aɴɢʀ (talk) 22:44, 20 August 2014 (UTC)
    • We can either change them to "und" and then later fix all instances of {{m|und|...}}, or we can leave them as they are. They're going to need to be fixed either way. —CodeCat 22:49, 20 August 2014 (UTC)
      • There are some cases where we anticipate and tolerate that linking templates are used with the language code 'und', aren't there? E.g. when Phaistos disc or Buyla inscription particles are linked to, or when forms from intermediate stages of reconstructed languages or language families are mentioned (pre-Proto-Algonquian (post-Proto-Algic), etc, and IIRC 'Middle Iranian'). Hence, I think it makes sense to leave language-less {{term}}s as they are, as they constitute a different category of things that need to be cleaned up from {{m|und}}s. (In some cases, language-less {{term}}s may actually need to be converted to {{m|und}}s, but in a lot of cases they need instead to be labelled as English or, in etymologies, Danish or Norwegian.) - -sche (discuss) 18:09, 21 August 2014 (UTC)
        • Of course. But if we have a cleanup category for uses needing a language, it doesn't imply that the category needs to be emptied out altogether. There will be some where the use of "und" is legitimate, but by far the most of them won't be, so a category would still be helpful. —CodeCat 18:12, 21 August 2014 (UTC)
          • It makes it harder to clean up when the cleanup category is filled with legitimate uses. It would be better to leave correct usages as {{m|und|...}} and unknown usages as {{term|...}}. --WikiTiki89 18:24, 21 August 2014 (UTC)

This was proposed just couple of days back at Wiktionary:Requests_for_moves,_mergers_and_splits#Template:term_into_Template:m, with three opposes. --Dan Polansky (talk) 22:58, 22 August 2014 (UTC)

I have created a vote: Wiktionary:Votes/2014-08/Migrating from Template:term to Template:m. --Dan Polansky (talk) 23:09, 22 August 2014 (UTC)

Capitalizing proper nouns in reconstructed languages[edit]

Should we capitalize proper nouns in reconstructed languages? For example, should the Proto-Slavic word for Rome be located at *Rimъ or *rimъ?

From my point of view, reconstructions represent sounds, not spellings, and sounds have no notion of capitalization. Additionally, if we choose to use capitalization, we would be forced to pick a particular language's rules for when to capitalize words. In the case of Proto-Slavic, the various Slavic languages themselves have varying rules for capitalization. --WikiTiki89 17:42, 20 August 2014 (UTC)

I think we should not capitalise. —CodeCat 17:57, 20 August 2014 (UTC)
If it's a convention to standardize them in the literature, which seems to be the case, we should do it as well.
Contrary to the popular misconception, reconstructions don't represent sounds but phonemes. Phonetic reconstruction (*abc -> *[abc]) is a different category and has little to do with comparative method. --Ivan Štambuk (talk) 18:02, 20 August 2014 (UTC)
Allophones can be reconstructed in some cases, though. Sometimes allophones become phonemes later on, but we can also reconstruct them through the effect they have on sound changes. For example, we can reconstruct two allophones of Old English /x/ because of later developments in late Middle English. For Proto-Germanic we know that an allophone [ŋʷ] existed because of the effects of the w:Boukolos rule. And for Proto-Indo-European the allophony of syllabic and consonantal semivowels is also well known. —CodeCat 18:08, 20 August 2014 (UTC)
Unless I'm going into detail about phonology/phonetics, there is no reason to differentiate sounds from phonemes here. Neither sounds nor phonemes have any notion of capitalization. But as to your first point, I'm willing to bet that most sources that capitalize reconstructed proper nouns use the language would use the rules of its own language to determine when a word should be capitalized. For example, Vasmer would use the Russian rules, while the Hrvatski etimološki rječnik would use the Croatian rules. In the cases where the rules differ, how would we decide which one to follow? --WikiTiki89 18:11, 20 August 2014 (UTC)
Probably we should find common rules and use them, like we do with reconstructed words. —Useigor (talk) 18:15, 20 August 2014 (UTC)
We already capitalize phonetic transcriptions like Pinyin so there is a precedent. The whole transcription of sounds vs. alphabetic words dichotomy is false - this is an issue of arbitrary convention and nothing else. Thousands of largely unwritten languages use scholarly transcription schemes invented to accommodate their phonology, with capitalization rules conveniently imposed. Rules for capitalization are generally the same across all Slavic languages - the few corer cases like demonyms could be decided by counting the preferred form in reflexes. --Ivan Štambuk (talk) 13:29, 21 August 2014 (UTC)
Pinyin has its own standard capitalization rules. --WikiTiki89 13:39, 21 August 2014 (UTC)
Which are completely arbitrary (though based on Latin-script languages) which just proves my point that something being not a "real" word but rather a transcription is no argument. --Ivan Štambuk (talk) 13:49, 21 August 2014 (UTC)
The rules themselves may have been chosen arbitrarily, but our rule is straightforward and logical (namely: follow the standard). In the case of reconstructions, there is no standard to follow, so the arbitrary choice is on us, and we shouldn't be making such arbitrary choices. --WikiTiki89 13:59, 21 August 2014 (UTC)
We do not follow the standard but the attested usage. That the usage overwhelmingly conforms to the standard is a different issue - but even if it didn't we'd still have to add them because they're attested as such. Conversely, the lack of de iure standard does not preclude the possibility of forming one on or own, as it has already been done countless times, on the basis of de facto standard established in the literature, or for pragmatic reasons. Furthermore, the precedent set by Pinyin where capitalization and other orthography rules are being assigned to a purely transcriptional notation demonstrates that the practice is far from unnatural, and that the dichotomy of word vs. transcription is a false one. --Ivan Štambuk (talk) 18:29, 21 August 2014 (UTC)
The "dichotomy" that I was referring to is not word vs. transcription, but spelling vs. transcription. Both spellings and transcriptions represent words (in fact, as we see in the discussion above, no one actually knows what a word is). I never said that capitalization in transcriptions is unnatural, only that we shouldn't use it for proto-languages. Pinyin was designed as a multipurpose system. One of these purposes is embedding Chinese words in foreign language text ("President Xí Jìnpíng did such-and-such."), which is why capitalization is useful, since foreign languages expect capitalization. However, we do not do this at all with proto-languages; we don't write "Shrines of *Perunъ were located either on top of mountains or hills", we only use this transcription to talk about the word itself. Attestations of the spelling "*Perunъ" are not really attesting anything, they are just simply how other researchers choose to transcribe the term. We are free to transcribe the term however we please (for example, we could transcribe it as *Perūnu) as long as it is internally consistent, but it is in our own best interest to ensure that the transcription conforms with some sort of norm, so that our readers who are accustomed to reading about Proto-Slavic will not be confused. However, I doubt much confusion would come from not capitalizing nouns. In fact it does our readers a disservice to pretend that *Perunъ the god is somehow a different word form *perunъ the weather phenomenon, when we in fact do not know whether the Slavs of the time saw them as the same thing or different. --WikiTiki89 18:55, 21 August 2014 (UTC)
The deity appellation is the original term, and the common noun *perunъ is secondary, derived from the proper noun. Far from being a disservice, it is in cases such as this when proper and common noun denote different entities but differ only in capitalization, that capitalization of proper nouns is helpful from lemmatization. Proper noun is uncountable and animate, whereas the common noun would be countable and inanimate, and both have separate and different set of reflexes in daughter languages.
Your argument was that "reconstructions represent sounds and not spellings", from which it is now evident that when you typed spelling you meant word. Why it shouldn't be used for reconstruction still remains a mystery. This "fictional reader" argument gets thrown a lot for every imaginable dispute and I don't buy it. If anything, readers would expect proto-terms to conform to capitalization rules of ordinary languages.
Regardless of the Pinyin's design goals, there is fundamentally no difference between "Shrines of *Perunъ" and ""President Xí". Both are not "real" words to the same extent, regardless of how you define the term. --Ivan Štambuk (talk) 00:08, 22 August 2014 (UTC)
It doesn't matter what the original term was (in fact I'm curious how we even know which one it was). The point is they may or may not be different words, they may or may not have had different spellings (if they even had any sort of writing), but the only thing we do know is that they were pronounced the same. "President Xí" and other Pinyin words are attestable as uses in English, while reconstruction notation is only attestable, and therefore only meant for, mentions. --WikiTiki89 03:42, 22 August 2014 (UTC)
Just on capitalisation of Chinese/Japanese/Korean romanisations to no-one in particular, since it was mentioned here. There are certain rules, used by various dictionaries and standards, not just for the use of loanwords in English but e.g. educational purposes. Place names and people's names are definitely capitalised, even the rules for word spacing and hyphenations are described. So, "Xí Jìnpíng" is a standard modern transliteration (pinyin) for 近平 but "Xí Jìn-píng" or "Xí Jìn Píng" is not. Months, days of the week are not capitalised There are some discrepancies as for demonyms and language names e.g. 中国人 (Chinese person), 中文 (Chinese language). Both "zhōngguórén" and "Zhōngguórén" (less commonly "Zhōngguó rén"), "zhōngwén" and "Zhōngwén" are used by various dictionaries. An agreement between editors is required on capitalisations in this case. --Anatoli T. (обсудить/вклад) 04:27, 22 August 2014 (UTC)
In some transcription schemes for proto-languages, capital letters are semantically different from lower-case letters. For example, Proto-Brythonic is often reconstructed with both *b and *B, where Proto-Brythonic *B is the descendent of Proto-Celtic *b, while Proto-Brythonic *b is the descendent of Proto-Celtic * in a leniting environment. So capitalizing proper names in those cases would be a bad idea. And I'm generally opposed to it for other proto-languages where that isn't an issue, too. —Aɴɢʀ (talk) 18:22, 20 August 2014 (UTC)
What about the reconstructed Latin words? --kc_kennylau (talk) 06:22, 22 August 2014 (UTC)
The difference with Latin, is that Latin orthography is well known. For most proto-languages, the orthography is 100% artificial. --WikiTiki89 12:07, 22 August 2014 (UTC)

More entries in the lemma category than in its subcategories?[edit]

I am trying to figure out where the lemmas are coming from in Category:Hungarian lemmas. It has more than 16000 entries but the subcategories contain only about 14000, even if I include the phrases, proverbs, suffixes, prefixes which are not specifically listed under the subcategories. Are there other categories? --Panda10 (talk) 13:24, 21 August 2014 (UTC)

You can see a full list of the categories that are recognised as "lemmas" at the top of Module:headword. —CodeCat 13:37, 21 August 2014 (UTC)

Letter petitioning WMF to reverse recent decisions[edit]

The Wikimedia Foundation recently created a new feature, "superprotect" status. The purpose is to prevent pages from being edited by elected administrators -- but permitting WMF staff to edit them. It has been put to use in only one case: to protect the deployment of the Media Viewer software on German Wikipedia, in defiance of a clear decision of that community to disable the feature by default, unless users decide to enable it.

If you oppose these actions, please add your name to this letter. If you know non-Wikimedians who support our vision for the free sharing of knowledge, and would like to add their names to the list, please ask them to sign an identical version of the letter on change.org.

-- JurgenNL (talk) 17:35, 21 August 2014 (UTC)

Process ideas for software development[edit]


Hello,

I am notifying you that a brainstorming session has been started on Meta to help the Wikimedia Foundation increase and better affect community participation in software development across all wiki projects. Basically, how can you be more involved in helping to create features on Wikimedia projects? We are inviting all interested users to voice their ideas on how communities can be more involved and informed in the product development process at the Wikimedia Foundation.

I and the rest of my team welcome you to participate. We hope to see you on Meta.

Kind regards, -- Rdicerb (WMF) talk 22:15, 21 August 2014 (UTC)

--This message was sent using MassMessage. Was there an error? Report it!

Haida lects[edit]

Wiktionary currently includes Southern Haida (hax), Northern Haida (hdn), and the macrolanguage they are sometimes considered to form, Haida (hai). Μετάknowledge and I discussed this on my talk page and are of the opinion that we should deprecate the macrolanguage hai and have only hax and hdn. The phonological and other differences between Northern and Southern Haida are, as linguist Michael Krauss puts it, "rather great, allowing only partial mutual intelligibility without practice, perhaps like Swedish and Danish, or German and Dutch." Translator Robert Bringhurst says "[i]t is, in fact, chiefly out of courtesy that northern and southern Haida are described as two dialects rather than two close but separate languages. By 1900, north and south had clearly known centuries of diverging cultural growth." Each language indeed has its own dialects (Kaigani and Masset Haida are the mutually intelligible dialects of Northern Haida, Skidegate and the now-extint Ninstints constitute Southern Haida). - -sche (discuss) 19:08, 22 August 2014 (UTC)

CFI and Non-Deities in Classical Mythology[edit]

Is there any reason we should keep most entries or senses referring to individual people in Greek mythology? It's true they're not covered by the "given name and surname" rule in CFI, but the names often have no meaning beyond their reference to the individuals themselves. Do we really need definitions that read like "daughter of so-and-so and so-and-so, wife of King so-and-so, and mother of so-and-so"? Chuck Entz (talk) 22:27, 22 August 2014 (UTC)

We should keep them for the lexicographical information that they carry, including etymology and pronunciation; that is for keeping the entries. As for keeping senses, that is kind of natural to me. A related poll: Wiktionary:Beer parlour/2010/December#Poll: Including individual people. --Dan Polansky (talk) 22:31, 22 August 2014 (UTC)
While these words do refer to the individuals themselves, the individuals may be so widely known that their names are assumed understood. If you consider the use-mention distinction, then most people will be introduced to a reader much the same as any other unfamiliar term is. But certain people are assumed to be known and their names are therefore only used and not introduced. This applies not just to Greek mythological names but also to names in modern times. For example Elvis can be seen in this way. Of course the knowledge of the person fades with time as popular culture shifts, but some names stay known for longer than others, just look at Bonnie and Clyde, Hitler, Napoleon, Caesar, Tutankhamun etc. These names definitely have "meaning". The meaning is tied to a cultural context, so that when that context is lost, the meaning goes with it. But in a sense, learning who Napoleon was is not so different from learning what a lute or an ironclad is. —CodeCat 22:38, 22 August 2014 (UTC)
We could include them with a definition such as "Character in Greek mythology" with a link to the Wikipedia article. --WikiTiki89 22:46, 22 August 2014 (UTC)
What a horrible idea. --Dan Polansky (talk) 22:46, 22 August 2014 (UTC)
Why is it a horrible idea? We already do this for many Biblical characters. --WikiTiki89 22:49, 22 August 2014 (UTC)
Actually scratch that. I was thinking of what we do for non-English entries. --WikiTiki89 22:53, 22 August 2014 (UTC)

Asteroid Names[edit]

The IP who's been adding these has been better known for adding mountains of junk edits to Japanese and Chinese entries, and to English entries having to do with magic and mythology, among others. These are so innocuous, and the other stuff is so awful, that we've let them slide, for the most part. It occurred to me, though, that there are some conceptual issues to be addressed.

Asteroid/minor planet names are assigned by an international body according to specific rules, usually based on the wishes of the discoverer(s), and are accompanied by a unique number reflecting the order in which the names have been assigned. Thus, the first asteroid discovered and named is w:1 Ceres, but the 1 is often left off. The IP describes it as Ceres being "short for 1 Ceres". I'm not sure, though, if the number is really part of the name, or whether it's added to the name to make the full official designation. In practice, the number seems to left off quite a bit in normal usage.

A more serious issue is what language header should be used. It seems clear to me that something like 1 Ceres, as a scientific name assigned by an international body, is interlingual. I'm not sure, though, if we have- or should have- entries for any of these. The question then becomes: is the sense at Ceres referring to 1 Ceres English, translingual, or some combination of both? Is the number like the author abbreviations in taxonomic names: necessary to be included at least once in a publication for the name to be technically complete, but mostly left off? Or is omitting it a sign that it's not translingual in that use? How is the name handled in other languages where script or morphology are different than the norm in the language for things like taxonomic names? Chuck Entz (talk) 04:04, 23 August 2014 (UTC)

Besides w:fr:(1) Cérès begins with these words: (1) Cérès (international designation (1) Ceres). We can see into w:Category:Asteroids that the parentheses are often used. JackPotte (talk) 08:31, 23 August 2014 (UTC)
Wikipedia tells us that there are 625 000 asteroids to be named. The IP has concentrated on the first hundred. Could we at least decide that anything beyond the top 100 will be deleted? All human names are likely to be used to name asteroids and hurricanes (and doggies, dolls and teddy bears), it's not possible to define every instance. Ceres certainly needs the asteroid definition, but I don't think the numbered forms belong to a dictionary. --Makaokalani (talk) 11:28, 29 August 2014 (UTC)
@Makaokalani: I think we need something other than an arbitrary criterion for any class of items, like asteroids, that we include. By our often-hostile reaction to items that are boring linguistically and our loving attention to items of the tiniest linguistic interest we betray the self-serving bias of many of our decisions and even policies. If this dictionary is to serve an audience beyond ourselves, we need to have principles and criteria that have more of a user focus, that we can include in CFI and even in a slogan to supplement or even replace "All words in all languages".
Asteroid names seem clearly intended to be suitable for use in all languages, to the greatest practical extent. The official name with the number is almost guaranteed to be attestable in scholarly literature and lists. If we were to exclude mere inclusion in a list on the possible grounds that such inclusion was a mention, not a use, I'd not be surprised if we didn't drastically(!) reduce inclusion. Such a principle would need to be applied in all realms to avoid the kind of bias I refer to above. But such a principle has the disadvantage of requiring many RfVs and much attestation effort.
A "policy" decision to include a class of names (eg, asteroid names with or without the order-of-discovery number) is subject to "correction" by RfVs. Do we know which of the forms is most likely to be in use translingually in running text? Both? Can the lettered portion of the full asteroid name be used again for an asteroid name, so only the number and the combination would be unique? Ie, could there be a 23400987 Ceres as well as 1 Ceres?
It seems silly to have L2 sections for every language that attestably hosts discourse about asteroids for either form of the asteroid names. The pronunciation rationale sometimes advanced for toponyms seems less applicable to asteroids, which will much more rarely be pronounced. The precedent we've established for taxonomic names says that we don't have any pronunciation section for translingual terms. I could imagine some kind of automatic system for generating language-specific IPA to help users pronounce any such term, which might be a reasonable approximation of what actually takes place in discourse using such terms. DCDuring TALK 12:26, 29 August 2014 (UTC)
It's incorrect to say that "1 Ceres" is short for Ceres, per se. Ceres is the planetoid's name, 1 Ceres is its systematic designation, which combines the object's name with its unique serial number (the number system came later than the naming, and the two are somewhat independent - early asteroids like Ceres were named without being numbered, and most modern asteroids are numbered without being named). A few asteroids/comets do have the same name (eg 209P/LINEAR and 118401 LINEAR, or the 9 Comets Shoemaker-Levy), which are taken from the names of their discoverers and have to be distinguished by their number. In running scientific text, asteroids and comets always seem to be given by name and number, but in tables only the number is given. From a translingual perspective, the English name takes precendence in scientific papers, but in common usage they may differ for transliteration reasons etc (what we call 67P/Churyumov–Gerasimenko, the German press call Tschurjumow-Gerassimenko or Tschuri), and some of the oldest objects have had other names grandfathered in (so the planetoid that we call Ceres, the Chinese call 穀神星 (star of the God of grain)).
Personally, I think a good rule of thumb would be to have any and all asteroids and comets where we can find three citations that name them without giving them any additional title or number (which would serve as an indication to the reader that they should check a sky catalogue, not a dictionary). "Ceres" is easily citable without the 1, and would therefore be kept - likewise, the more major recently-discovered minor planets like 99942 Apophis, 50000 Quaoar, 90377 Sedna and Comet Shoemaker-Levy 9 can be cited without the number. Wilson–Harrington on the other hand only appears in Google Books as "Comet Wilson–Harrington", "Wilson–Harrington (4015)" or "107P/Wilson–Harrington", and would therefore not be worth having, since in each case, it's clear that it represents an astronomical object. Halley's Comet would be an exception, I assume, since it uses the possessive (it's not Comet Halley) and has interesting translations in other languages. Smurrayinchester (talk) 08:55, 11 September 2014 (UTC)
So you would seem to favor making the language header for such names Translingual.
I am not sure that I understand the rationale for excluding any two-part asteroid name (ie, with the number) that is citable outside of a table. We rarely show any willingness to exclude multi-part names based on their ready intelligibility in context. It is neither a policy, nor an accepted practice, despite arguments that it should be. Certainly the first time one sees such a name one would almost certainly not be aware of any naming convention, though one might infer that the number and name collocate and behave as a unit. Is it really adequate for a user to be able to infer the definition "the 4015th astronomical object of this type to be formally recognized"?
Are we being consistent across topical domains in how we treat such proper names, both as to inclusion and the content of the definition? Consistency is not required, but somewhat similar cases might shed some light on lexical features that argue for inclusion and what makes for an adequate definition. Somewhat similar domains include multi-part taxonomic names, especially species names distinguished only by the discoverer's name; food additives (eg, E300) the class of which is almost always clear from context, vernacular names of taxa (eg, oak and oak tree), which often include a hyponym (eg, tree) that is arguably redundant. Another types with less broad inclusion has as an exemplar Route 66. DCDuring TALK 12:48, 11 September 2014 (UTC)
Smurrayinchester's suggestion sounds reasonable to me. - -sche (discuss) 00:33, 12 September 2014 (UTC)

Migrating from Template:context to Template:cx[edit]

FYI: Wiktionary:Votes/2014-08/Migrating from Template:context to Template:cx. --Dan Polansky (talk) 08:40, 23 August 2014 (UTC)

Where was this discussed? —CodeCat 12:33, 23 August 2014 (UTC)
You could start discussing it now, instead of retreating into a procedural defense invoking those you routinely ignore. It does make it look like you have no particular substantive arguments.
The proposal seems sensible. {{context}} could be retained as a redirect, as it affords an option that may not require consulting the documentation for casual users. The shorter name would reduce the size of the database. Making {{cx}} the effective template rather than a redirect would eliminate one extra redirect call at the time of page loading. This seems like basic efficiency, even though efficiency is a secondary concern. DCDuring TALK 13:40, 23 August 2014 (UTC)
It's usually Dan who resorts to procedure, so I was only returning what he does to me all the time. Petty maybe, but satisfying. —CodeCat 14:44, 23 August 2014 (UTC)
Ok, but where are the substantive arguments now? DCDuring TALK 15:20, 23 August 2014 (UTC)

FYI: Wiktionary:Votes/2014-08/Templates context and label. --Dan Polansky (talk) 17:47, 23 August 2014 (UTC)

CFI Misspelling Cleanup[edit]

FYI: Wiktionary:Votes/pl-2014-08/CFI Misspelling Cleanup. --Dan Polansky (talk) 09:32, 24 August 2014 (UTC)

Format and layout of language categories[edit]

Yesterday I made some changes to how language categories like Category:English language display. Instead of some prose, there is now a table which displays the information in a more systematic format, and it also shows information that did not appear before, like ancestors and other names. I do think that the table is a good addition, but I'm not really happy with how it looks. So I wonder if we should keep it this way. Is there something we could do to improve it? Or was the original format better? —CodeCat 13:46, 24 August 2014 (UTC)

A few things: 1. Maybe you could merge this table with the one on the right- having 2 different tables makes the page look messy. 2. If there are no ancestors of a language don't display that row. 3. Can descendants of a language be shown as well as ancestors? 4. In Category:English language, for example, in other names there is "Hawaiian Creole English". But "Hawaiian Creole English" isn't another name for English- it's its own thing. DTLHS (talk) 20:59, 24 August 2014 (UTC)
We include not just names for the language, but also names of varieties that are subsumed under the same language on Wiktionary. Category:Dutch language also lists Flemish for example. Showing descendants would be much harder to do, and would involve basically searching through all the languages to find any that have the current language as their ancestor, and then repeat. —CodeCat 21:05, 24 August 2014 (UTC)
(e/c) This is somewhat orthogonal to your question, but the list of "Other names" highlights something I've been thinking about for some time, which is that it's both confusing for anyone looking at our data (e.g. people looking at WT:LOL, and now also people looking at Category:English language), and possibly even undesirable from a technical standpoint, that we conflate in the names= parameter both "X, another name of language Foo" and "Y, name of one dialect of language Foo which is subsumed under the header Foo". "Modern English" is indeed another name for "English"; anything that is ISO-code-en (i.e. post-1500) "English" can also, in linguistic context, be called "Modern English". That's very different from "Hawai'ian Creole English", which is not another name for English — not interchangeable with "English" — but merely the name of one non-independent (non-L2-having) dialect. So I wonder if we shouldn't split a dialects= parameter off from the names= parameter. (There will be a few edge cases where a name refers to both a dialect and the language itself. These could be handled by listing the name in both places, or by giving one parameter priority and saying e.g. "never list anything in the dialects= parameter which is already in the names= parameter".) - -sche (discuss) 21:13, 24 August 2014 (UTC)
I agree. And while we're at it, we might as well split off a separate value for the canonical name? —CodeCat 21:18, 24 August 2014 (UTC)
@CodeCat: After considering it (and then getting distracted and forgetting about this thread for a week), I think that would make sense. As long as we're splitting off both a canonical= parameter (whether we decide to call it that or something else) and a dialects= parameter, should we also rename the rump "names" parameter to something like alt_names=? - -sche (discuss) 05:39, 15 September 2014 (UTC)
@CodeCat: Shall we start splitting up the names= parameter? If we not only add canonical= and dialects= parameters but also add an alt_names= parameter, and deprecate names= in favour of those three, we can sort old names= entries at our leisure and always know which languages have had their names sorted and which haven't. It seems like it might even be possible to have a bot fix all the languages which have only one name specified (just switch them from calling that one name a names= to calling it a canonical=). I'm not wedded to any of the specific parameter names I'm using in this post, btw, if you have better ideas. How do you think we should handle edge cases where a string is both the name of a dialect of language X, and an alternative name for the whole language X? List it in both places, or only one? - -sche (discuss) 16:03, 1 November 2014 (UTC)
I would suggest using canonicalName= to be clear, alongside otherNames= and includedSubdivisions=. That way we don't have to concern ourselves with the politics of "dialect". —CodeCat 16:50, 1 November 2014 (UTC)
Somewhat unsurprisingly, I don't like the change. --Dan Polansky (talk) 18:15, 25 August 2014 (UTC)
I agree that the box should be merged with the one on the right. --WikiTiki89 18:20, 25 August 2014 (UTC)
Should the content that is currently on the right be put to the left, right, top or bottom of the current left box? —CodeCat 18:28, 25 August 2014 (UTC)
I'm not sure, but the merged box needs to end up on the right. --WikiTiki89 18:34, 25 August 2014 (UTC)
After thinking for a moment longer, I decided that I think what is currently on the left should go on top of what is currently on the right. --WikiTiki89 18:35, 25 August 2014 (UTC)

Red link headline in new entries[edit]

When I create a new entry, the line under the part of speech is a red link even after saving the entry. See -obb. I am using the {{head}} template. Is this how it's supposed to work? --Panda10 (talk) 20:05, 24 August 2014 (UTC)

The fuck? I didn't think {{head}} was supposed to create links at all, just the pagename (or head= parameter if there is one) in boldface. —Aɴɢʀ (talk) 20:09, 24 August 2014 (UTC)
Panda10, no this is not normal. The problem is of course in breaking a term.
BTW, null edit makes it blue (because null edit makes the module rerun, which on the second run sees the suffix on the list of available articles).
It is interesting that noone has noticed that suffixes and prefixes link to themselves so far.--Dixtosa (talk) 20:45, 24 August 2014 (UTC)
I think this must be fairly recent behavior. I suspect somehow it's treating affixes as two-term entries, like hot dog, where the two words are automatically linked. Only for some reason it thinks [[-obb]] is something like [[]][[-obb]]. —Aɴɢʀ (talk) 20:56, 24 August 2014 (UTC)
This is probably caused by recent changes to Module:headword that User:Wikitiki89 made. —CodeCat 21:03, 24 August 2014 (UTC)
Yes, this is my fault. I will attempt to fix it shortly. In the meanwhile, I am going to assume that this is not harmful enough to revert. Angr diagnosed it correctly. Only I was smart enough to remove the [[]], but not smart enough to remove the link from the [[-obb]]. --WikiTiki89 22:29, 24 August 2014 (UTC)
Yes check.svg Done --WikiTiki89 23:06, 24 August 2014 (UTC)
But wait, I think hyphen should still be a separator. Now it is not--Dixtosa (talk) 12:14, 25 August 2014 (UTC)
If it isn't possible to add links when there are letters on both sides of the hyphen, while not adding links if the hyphen is the first or last character in the string, then better not to have automatic linking in words with hyphens at all, and add it manually where needed. —Aɴɢʀ (talk) 12:22, 25 August 2014 (UTC)
That is possible and fairly easy too. My first thought is one regex replace:
"text lol-lol2".replace (/([^\s\-]+?)(\-|\s)([^\s\-]+?)/g, "$1]]$2[[$3") and embrace it with "[[" and "]]" if at least one replace occurs--Dixtosa (talk) 13:03, 25 August 2014 (UTC)
It's possible, but it is not always wanted. Very often hyphenated words should be linked together. --WikiTiki89 14:56, 25 August 2014 (UTC)

IPA alphabet[edit]

Where did the symbols e̞, o̞ and ø̞ go? They are needed at least for Finnish pronunciations. It's a small language, but big in en-Wiktionary. --Hekaheka (talk) 04:37, 25 August 2014 (UTC)

Please check Module:IPA. Someone may add those symbols, if they are valid. --Anatoli T. (обсудить/вклад) 04:46, 25 August 2014 (UTC)
Heka may have been referring to MediaWiki:Edittools, from which the symbols were removed earlier this month and to which they were just re-added. - -sche (discuss) 04:52, 25 August 2014 (UTC)
I saw that too. I thought these IPA symbols generate module errors. It would be the case, if they are missing in the module. --Anatoli T. (обсудить/вклад) 04:55, 25 August 2014 (UTC)
-sche, thanks for re-listing them. --Hekaheka (talk) 05:05, 25 August 2014 (UTC)
It was User:Wyang, actually. --Anatoli T. (обсудить/вклад) 05:14, 25 August 2014 (UTC)
Anatoli, they shouldn't cause an issue, because they are in the module. They are [eoø] + "combining tack below" for lowered articulation. The latter symbol is in Module:IPA/data under "primary articulation". --Catsidhe (verba, facta) 05:12, 25 August 2014 (UTC)
OK, I haven't checked. --Anatoli T. (обсудить/вклад) 05:14, 25 August 2014 (UTC)
Why are they needed for Finnish? Does Finnish have three heights of mid vowels? In other words, are /e/, /e̞/, and /ɛ/ all separate phonemes? Or are [e], [e̞], and [ɛ] three distinct allophones of the same phoneme? Because there's no need to use the symbol e̞ at all, either allophonically or phonemically, unless you're already using both e and ɛ and need a third symbol that's distinct from both of them. —Aɴɢʀ (talk) 07:00, 25 August 2014 (UTC)
I can't explain this any better than it's currently done in Wikipedia: Finnish_phonology#Vowels. If I understand it correctly /e/, /e̞/, and /ɛ/ are indeed separate phonemes and Finnish happens to use /e̞/ and not /e/ as equivalent for the Latin alphabet "e". --Hekaheka (talk) 16:18, 26 August 2014 (UTC)
I don't think that it's only necessary to use the diacritic to indicate contrasts. If the phoneme is really /e̞/, why write /e/ which is less accurate? It would be a bit like writing /m/ instead of /n/ when a language has only one nasal consonant and its articulation is alveolar. Or writing /s/ when the language's only sibilant is postalveolar, or writing /p/ when the only labial plosive in the language is voiced. —CodeCat 16:41, 26 August 2014 (UTC)
The clause "if the phoneme is really /e̞/" doesn't make any sense. Phonemes aren't preassociated with IPA symbols. IPA symbols are convenient ways of representing phonemes and allophones, and there is always some wiggle room in their application. By longstanding phonetic convention, diacritics are to be avoided unless they illustrate some important distinction; and ordinary Latin-alphabet characters are to be preferred over modified ones whenever feasible. So if a language has only one front mid vowel, the convention is to use "e" to represent it. Using [e̞] by itself in fact doesn't make any sense, because [e̞] means "a sound more open than [e]", but if you don't use [e] in your transcription system of the language in question, then you haven't defined what sound [e̞] is more open than. You could say it's more open than the cardinal vowel [e], but in fact very few languages' vowels are located at exactly their cardinal value. (Neither English [i] nor German [i] is cardinal [i], though German is closer to it than English is.) So you have to define [e] in the context of your language before the symbol [e̞] is even meaningful. As for Finnish, Finnish phonology#Vowels does not say that /e/, /e̞/, and /ɛ/ are separate phonemes; it says that Finnish has a single mid front unrounded vowel phoneme which falls between cardinal [e] and cardinal [ɛ]. The authors of the Wikipedia reveal their ignorance of how the IPA works by insisting on /e̞/ to mark that vowel, when by actual IPA conventions it should be transcribed /e/. —Aɴɢʀ (talk) 17:51, 26 August 2014 (UTC)
What he^ said. --WikiTiki89 18:04, 26 August 2014 (UTC)
If understanding of IPA relies on such conventions, then it kind of bypasses the point of having a universally applyable and unambiguous transcription system. The reasoning here also seems a bit circular in a way. On one side, you say you need to define [e] before defining [e̞], but then what does it mean to say that the Finnish sound is between cardinal [e] and [ɛ]? Surely that in itself means that there are absolute reference points that the lowering symbol is relative to? So the way I see it, the question is whether the correct transcription is [e̞] or [ɛ̝]? If the sound is between them, then either symbol is appropriate. —CodeCat 18:21, 26 August 2014 (UTC)
The "cardinal" ones are defined (although not sure how they are defined), but as Angr already said, very few languages have vowels that coincide with the "cardinal" values, so by CodeCat's logic, all languages should use a whole pile of diacritics on every vowel and every consonant. --WikiTiki89 18:28, 26 August 2014 (UTC)
See Cardinal vowels on how they're defined. Cardinal [i], [u], and [ɑ] are defined as the most high front vowel possible, the most high back rounded vowel possible, and the most low back vowel possible, and all the others are defined as being a certain acoustic distance between those. The IPA makes no claim of being able to represent every conceivable nuance in articulation (or acoustics) in every single spoken language in an unambiguous and universally applicable way. German Haus and English house sound quite different from each other, but both are—correctly—transliterated [haʊs]. —Aɴɢʀ (talk) 18:44, 26 August 2014 (UTC)
I realise that IPA is only an approximation. I see it as a kind of set: every symbol used on its own represents a certain set of possible articulated speech sounds. Diacritics narrow that set further down. But in the case being discussed here, it's not quite clear whether the sound belongs to the [e] set or the [ɛ] set, as it's equally distant to the cardinal value of both (as I understand it). So in this case using a diacritic seems warranted to clarify that the sound in question is an edge case. —CodeCat 18:54, 26 August 2014 (UTC)
Modern Hebrew and Standard Spanish are a couple examples of other langues for which a sound right between [e] and [ɛ] is transcribed as /e/. --WikiTiki89 19:36, 26 August 2014 (UTC)
If the mid vowels of Finnish are midway between canonical [e], etc and [ɛ], etc, and our Finnish-speaking editors want to use the transcriptions [e̞], [ø̞], [o̞], then I think it would be unhelpful for people who don't speak Finnish to try to ban that notation from narrow transcriptions and mandate that the narrow transcriptions be less accurate than the Finnish-speakers want them to be. (The degree of specificity used in narrow transcriptions can and does vary from language to language, so the argument that we'd have to use many diacritics for other languages can be dismissed.)
In broad transcriptions I am inclined to accept that the vowels can be written /e/, /ø/, /o/ or /ɛ/, /œ/, /ɔ/ (de.WP uses the latter).
I would attach weight to how Finnish references transcribe the vowels. Perhaps Hekaheka has references on Finnish with IPA transcriptions and can say what symbols those references use. Checking Google Books, I find only a few English- or German-language references that give IPA-like transcriptions, but it's not clear to me that they are actually IPA and not just regular letters enclosed in IPA-like brackets:
  • Melvin J. Luthy's Phonological and Lexical Aspects of Colloquial Finnish (1973) speaks of "a syllable boundary between all low and mid vowels, e.g. [pa.eta], [kä.etä]."
  • Variation in Finnish phonology and morphology (1997) speaks of "the mid vowels /o, ö/".
- -sche (discuss) 19:47, 26 August 2014 (UTC)
If you ask me, this is an example of the bad Western-European-centric design of IPA. Most Germanic and Romance languages have at least four vowel heights, so for them it makes sense. But globally, far more languages have three heights than four. So having no symbol for a straight mid-vowel is quite a frustrating omission. —CodeCat 20:03, 26 August 2014 (UTC)
You're looking at it wrong. /e/ is what you're describing. /ɛ/ was only added to accommodate languages with four heights. --WikiTiki89 20:18, 26 August 2014 (UTC)
How do you know? —CodeCat 20:47, 26 August 2014 (UTC)
Because "e" has been used for "transcriptions" since Roman times, and they did not have a separate letter "ɛ". --WikiTiki89 20:50, 26 August 2014 (UTC)

I have relied on this source [3]. BTW, to my untrained ear, the Finnish "e" sounds about the same as English /ɛ/. --Hekaheka (talk) 21:42, 26 August 2014 (UTC)

But that's because English has no actual [e]. It has [eɪ] which is a slight closing diphthong, which probably sounds more like the Finnish [e̞i]. —CodeCat 23:05, 26 August 2014 (UTC)
Can you prove it? Lysdexia (talk) 04:05, 29 August 2014 (UTC)

Policy for Translations entries[edit]

Within the definition of most(?) English words, there is a "Translations" section giving corresponding words in various languages. (Like anything else with the word "translation" in it, this is not perfect, but can be immensely helpful.)

I guess there is a policy of not putting a "Translations" section on non-English words, which sounds reasonable since one can look at the English word, but there is an interesting class of exception to this rule -- when the English word does not exist. For example, 何番目 and wievielte are perfectly good Japanese and German words respectively, but the best we can do in English is the non-word whatth. (Personally I would have looked for "how manyth", but that isn't even a single word.) It seems to me it would be helpful to add the translations section to the foreign words, since otherwise a person who recalls that there is a way of saying this in at least two languages has no way of navigating from one to the other. Imaginatorium (talk) 10:08, 25 August 2014 (UTC)

I've thought about this too, and part of the problem is deciding which language to put the translation table in if we're going to add them for non-English languages. For example, a whole lot of languages have a single verb for "to be silent" (schweigen, zwijgen, taire, taceō, etc.), but English doesn't. I've often wanted to be able to put all of those words in a translation table, but where? If we allowed them in non-English languages, we'd have to have the same translation table in each one of the languages where this word exists, and that's a lot more than just the four I mentioned above. —Aɴɢʀ (talk) 11:20, 25 August 2014 (UTC)
Incidentally, to judge from Google Books, how manyth and/or how manieth might actually be attestable. —Aɴɢʀ (talk) 11:26, 25 August 2014 (UTC)
Another example would probably be "double eyelid" and "single eyelid". Wyang (talk) 12:56, 25 August 2014 (UTC)
The only way of adding these would IMHO be to allow sum-of-part translation-only English entries. We have some of them already, see Category:English_non-idiomatic_translation_targets, but there is some risk that these entries are eventually deleted afer discussion on WT:RFD. See also the discussion on Wiktionary_talk:Criteria_for_inclusion#Translation_target Matthias Buchmeier (talk) 17:54, 25 August 2014 (UTC)
Do you guys realize that we are not here to put everything that is interesting (Even if it is linguistically interesting) into articles? Examples are this, dord (and the like), some sum-of-part terms that may have an interesting etymology or maybe it is so widely used that it has become a thing, but... they are still sum of parts.
Anyways, Ill throw an idea (probably a stupid one though xD): let's make a User script, that will fetch (assuming JS can go on external links) the translation table from the language's own wiktionary and insert it into our article formatted in the the same way as English entries do and also apply assisted adding of translations, which actually adds to foreign wiktionary. In this way en:wikt runs the world :D of wiktionaries--Dixtosa (talk) 18:08, 25 August 2014 (UTC)
Hate to be a spoilsport, but we were not able to make automatic generation of entries work properly even within the English wiktionary. In principle we already have a way to find out translations of wievielter to other languages: one clicks "Deutsch" in the "In other languages" -list on the left side of the page and checks the translations in the German wiktionary. In the particular case of wievielter there's the problem that there's no article on it in de-wikt, but that would be a problem in any automated model as well. --Hekaheka (talk) 04:46, 27 August 2014 (UTC)
@Imaginatorium: I've made an entry for what number. --Anatoli T. (обсудить/вклад) 23:21, 25 August 2014 (UTC)
  • This could be solved by wrapping all such sum-of-parts English translations into a special template to track them in FL entries, and then generate (by bot) the appropriate entries in the Appendix namespace. This would really be an awesome feature - a cross-lingual glossary of terms otherwise not directly translatable to English, like that disputed list of terms that there was a big discussion about which I cannot find now (deleted?) containing "to call a mobile phone and let it ring once" and others. Some additional tags would be necessary though (e.g. part of speech). {{translation only}} is absurd - how I am suppose to know the meaning to translate if I knew only English and the missing FL? --Ivan Štambuk (talk) 15:19, 26 August 2014 (UTC)
    • From the entry name of course. —CodeCat 15:30, 26 August 2014 (UTC)
      Ha ha! Except that it's often not that self-explanatory - what number is barely even valid English and could be translated both as "what kind of" and "how much". Those translations referring to concepts that need entire English sentences to translate (like the notorious "to call a mobile phone") are too cumbersome to have as their own entries. --Ivan Štambuk (talk) 08:49, 27 August 2014 (UTC)

Proclitics in Hebrew[edit]

I saw the paragraph Wiktionary:About_Hebrew#Proclitics, and entry הערב. This seems inconsistent, although I don't see any problem to have entries such הערב. — Automatik (talk) 15:14, 26 August 2014 (UTC)

הָעֶרֶב is a different story, since it is idiomatic in the sense of "this evening/tonight". There are a few other similar cases, such as הַיּוֹם (today). The problem with including the non-idiomatic ones, is that there is a very large number of them and they add nothing useful to the dictionary. As I recently pointed out in the WT:GP, we would end up with ridiculous cases such as וּכְשֶׁבְּבֵיתְכֶם (u-kh'-she-b'-veit'-khém, and when in your house). --WikiTiki89 15:28, 26 August 2014 (UTC)

Allow the hyphen to be specified explicitly on Template:suffix and such[edit]

The templates {{suffix}}, {{suffixcat}}, {{prefix}} and so on currently require that you omit the hyphen from the affix parameter. If you put it in anyway, you end up with a double hyphen. This sometimes causes problems mainly because people might not think about it and add the hyphen in the parameter. But I believe that having the hyphen as part of the parameter explicitly is also a bit clearer. So I'd like to ask for two changes:

  1. Instead of never being able to add the hyphen, allow the hyphen to be specified optionally. If the module detects that the hyphen is already present, it will not add another one, but it will add one like before if it's not already present.
  2. If you agree with the first change, I would also like to ask if I may use a bot to convert entries that don't include the hyphen in the parameter, so that they include it. (For example: {{suffix|hyphen|less|lang=en}} > {{suffix|hyphen|-less|lang=en}}) This would effectively establish the use of the hyphen in the parameter as the more common form of use, while leaving the hyphenless form as a possible alternative for backwards compatibility for those who are used to it or prefer to do it this way.
  3. If you agree with the second change, there's the option to convert to {{affix}}. This template requires hyphens, so if we are going to have hyphens in the entries' wikicode anyway, we could also opt to use this template instead.

Please specify whether you agree with just option 1, or with both option 1 and 2. Or alternatively, if you oppose, specify whether you oppose only option 2 or both option 1 and 2. —CodeCat 14:40, 27 August 2014 (UTC)

I strongly oppose two sources producing the same output.--Dixtosa (talk) 16:50, 27 August 2014 (UTC)
Why? —CodeCat 16:59, 27 August 2014 (UTC)
Option 1
Support option 1. I mistakenly save pages with a hyphen all the time. Abstain on option 2. --Vahag (talk) 14:52, 27 August 2014 (UTC)
Support option 1, unlikely to change my mind. —This unsigned comment was added by DCDuring (talkcontribs).
Support option 1. --WikiTiki89 16:52, 27 August 2014 (UTC)
Mild support, since people (including me) do save pages with explicit hyphens sometimes, and it would be useful if the template handled that smoothly. - -sche (discuss) 00:15, 28 August 2014 (UTC)
Option 2
Support option 2, might change my mind. DCDuring TALK 15:32, 27 August 2014 (UTC)

Now that there's also {{affix}}, I've added a third option. I don't expect people will support migrating everything to it only days after it was created, but at least there's the opportunity. —CodeCat 00:37, 28 August 2014 (UTC)

Option 3

I've applied the first change. The templates will now first test to see if the given form already has hyphens where they belong, and if it doesn't, it will add them. So {{suffix|red|-ness|lang=en}} correctly shows -ness and not --ness. But it also means that something like {{suffix|brako|um-|lang=eo}} will have the hyphen applied, giving -um-, as the suffix parameter is a prefix here, and hence not of the expected type. This was also the previous behaviour, so it's necessary for it to work this way to not break anything. More subtly, it means that {{infix|education|-ma|lang=en}} will end up linking to --ma-, because infixes are expected to have hyphens on both ends, and this parameter didn't. —CodeCat 19:20, 2 September 2014 (UTC)

Template:affix[edit]

User:Kephir created this a few months ago but left it unfinished. So I made it usable. It replaces {{prefix}}, {{suffix}}, {{confix}} and {{interfix}}, but not {{compound}}, {{infix}} or {{circumfix}}. I hope it's useful, and if there are any problems or shortcomings, please report them on Module talk:compound. —CodeCat 22:37, 27 August 2014 (UTC)

Module:ar-verb needing testing, work, use of new {{ar-conj}}[edit]

Moved to Module talk:ar-verb. --Anatoli T. (обсудить/вклад) 01:36, 3 September 2014 (UTC)

Away for a few days.[edit]

I will be away for family matters until September 2. Please try to have the dictionary done by the time I get back. I will, however, be going to town on RfD cleanup when I return, unless someone else does it first. See you next Tuesday! bd2412 T 03:15, 29 August 2014 (UTC)

I am back, and I am going to town. Cheers! bd2412 T 13:39, 2 September 2014 (UTC)
I'm sorry to disappoint you, but the dictionary hasn't been completely finished yet. --WikiTiki89 14:09, 2 September 2014 (UTC)
It's closer than it was. I'll take that. bd2412 T 19:37, 3 September 2014 (UTC)

Request for comment: Admin corruption[edit]

When I was blocked I left the below message on my talk page but nobody dealt with or said anything about it. After the term it was no longer appropriate to post it at WT:VIP.

Earlier today Chuck Entz applied his assumption of Grimm's law and wrote a rude comment about me on https://en.wiktionary.org/wiki/Wiktionary:Etymology_scriptorium/2014/August#.CE.BA.CE.B1.CF.84.CE.AC_and_Appendix:Proto-Germanic.2Fgad.C5.8Dn.C4.85. Lysdexia (talk) 03:30, 29 August 2014 (UTC)

Transclude to Wiktionary:Vandalism in progress[edit]

Atelaes (talkcontribsglobal account infodeleted contribsnukeedit filter logpage movesblockblock logactive blocks) and EncycloPetey (talkcontribsglobal account infodeleted contribsnukeedit filter logpage movesblockblock logactive blocks) are ultraconservative English-haters who revert glosses and bully and block the user, me, who adds them. Their NPOV violations are linked at User talk:Lysdexia#2 week block and thereabove. Help:FAQ is exactly the standards I follow where one section tells the editor to use OED and other dictionaries for verification. When I tried to post a OED link to prove my edit to the entry sense, EncycloPetey said he didn't get OED and couldn't read the link; therefore he wouldn't consider my edit and block, even if one can access OED.com for free if one enters anything in the library card field. Atelaes has picked on me the whole time where at last he's reverted a legitimate edit after I had a talk with another admin, Chuck Entz, and blocked me for two weeks after I had another talk with Chuck on another word where I stated my case where he and the other admins or editors all break the NPOV policy and that the policies agree with my edits and disagree with theirs; however, Chuck has not answered and I did not revert this word. These two editors bring a hostile work environment to this project and misrepresent its policies and my edits. Lysdexia (talk) 17:51, 19 May 2014 (UTC)

We know Atelaes and EncycloPetey quite well and we know that they are neither vandals nor English-haters. I don’t know what you mean by ultraconservative. By making exaggerated accusations such as these, your own reputation and credentials suffer. I looked at a couple of the entries you may have meant, and it was clear to me that you think this is Wikipedia. We don’t work like Wikipedia. You think policy and rules must be stated somewhere in black and white. That has never been the case in this Wiktionary. First and foremost, listen to the advice and explanations of the old hands such as Atelaes and EncycloPetey. If you find something written in a policy page that disagrees with them, mention it with a link so that we can correct the policy page. —Stephen (Talk) 18:33, 30 August 2014 (UTC)

Make rhymes pages use / as the separator[edit]

Previous discussion: WT:RFM#Rhymes pages from using : as the separator to using /

In the previous discussion there was no opposition but also not much support either. Mglovesfun suggested removing the hyphen as well. I think that's a good idea, so I propose moving Rhymes:Dutch:-ɑn to Rhymes:Dutch/ɑn, Rhymes:Dutch:-ɑ- to Rhymes:Dutch/ɑ-, and doing the same kind of change to other pages. The {{rhymes}} template would need to be changed to point to the new place. —CodeCat 18:20, 30 August 2014 (UTC)

Should showI3raab be on by default in Module:ar-translit?[edit]

Apparently there is a standard for transliteration functions, involving 3 params text, lang, sc. The Arabic one has a fourth parameter showI3raab that causes final-syllable short vowels to be displayed; otherwise they are omitted. I think this should be the default -- if someone writes in those vowels they should be displayed. If you don't write them in, they won't be displayed, simple as that. Arabic script containing vowel diacritics isn't that common anyway, and when it's there, often those final vowel diacritics are omitted when it's intended that they don't be read. Benwing (talk) 02:51, 1 September 2014 (UTC)

Yes, for automatic transliteration, "showI3raab" should be the default, otherwise verbs won't show full endings. Otherwise, there's no point in displaying classical Arabic grammar endings. The ability to drop the endings in pausa or in a more relaxed MSA should be assumed. So, يُسَالِمُ (yusālimu) should be transliterated "yusālimu", even if it can be pronounced as "yusālim". The module will probably be restricted to conjugation and headers (if implemented), anyway. We have a tradition to drop "i3raab" vowels in translations - "katab", not "kataba". If translations were supplied with full vowels and automated, we could revisit this. --Anatoli T. (обсудить/вклад) 02:57, 1 September 2014 (UTC)
I think it should be disabled by default (mostly for the headword line and for links), but enabled in conjugation and inflection tables. --WikiTiki89 01:31, 2 September 2014 (UTC)
There is no real reason to disable "showI3raab" by default in the headers as well. We transliterate كَتَبَ (kataba) as "kataba", not "katab" and يَكْتُبُ (yaktubu) as "yaktubu", not "yaktub". If case or conjugation diacritics are written, they can also be transliterated, e.g. "يُسَالِمُ" is "yusālimu" with the final ḍamma, "يُسَالِم" is "yusālim", without it. It also matches Arabic grammar references. Textbooks often skip final vocalisation to reflect the actual pronunciations, which also matches transliteration then. Learners who know this feature will have no problem with this, dropping i3raab is a common knowledge, e.g. even if we transliterate "yusālimu", they will know to drop "u" in the less formal environment or in pausa. You can arrive at the abbreviated pronunciation by dropping i3raab, the reverse is more difficult ("kataba" and "yaktubu" transliterations are informative and unambiguous, you can always remove -a and -u). --Anatoli T. (обсудить/вклад) 02:13, 2 September 2014 (UTC)
Module:ar-translit isn't used in the headword line as far as I know; nor in links. Headword translations do contain final vowels in verb forms. I would rather have the default be to translate what is written; this is what all other languages do. If you don't write the final vowels in Arabic, they won't get transliterated, simple as that. Benwing (talk) 05:41, 2 September 2014 (UTC)
Thanks to your improvement in Module:ar-translit (it seems to do the right thing now) it will be possible to use it in the headword (without the override of the manual transliteration). It will only be neccessar to provide all the necessary diacritics and the the translit module will do the rest, if manual translit is missing. A headword containing نَظَافَة (naẓāfa) would transliterate it as "naẓāfa" or "naẓāfatun" if tanwīn is added, e.g. نَظَافَةٌ (naẓāfatun). --Anatoli T. (обсудить/вклад) 05:53, 2 September 2014 (UTC)
Our current practice for nouns is to write مَكْتَبَةٌ (with the tanween), but transliterate it as "maktaba". Do you think this should be changed? --WikiTiki89 12:06, 2 September 2014 (UTC)
I this this current practice makes no sense. Either do or don't write the tanween, but either way, transliterate what's written. I've never seen an Arabic book that uses a practice like what you just described; e.g. John Mace's "Arabic Verbs" writes the i`rāb vowels on verbs but not on verbal nouns or participles, and transliterates accordingly.
BTW Currently Module:ar-translit transliterates final taa' marbūṭah as (t), although it could as well write it as h or nothing. Writing as nothing has the problem that it becomes hard to distinguish feminine endings from the verbal -a ending, and hard to distinguish the long feminine ending -āh from masculine alif maqṣūrah ending . Benwing (talk) 00:33, 3 September 2014 (UTC)
I am all for transliterating the tanween in nouns, but that is not our current practice. If we do that, then we won't have any problem of how to transliterate the ta marbuta. --WikiTiki89 00:47, 3 September 2014 (UTC)
(E/C) @Wikitiki89. Yes, I think it makes sense to change it, especially if we are going to use more the automatic transliteration module. In my translations or occasional Arabic entries I make I usually don't use iʿrāb, so I don't think I didn't follow it.
@Benwing. I think it's better to transliterate tāʾ marbūṭa (in pausa, without iʿrāb) as nothing. It's not possible to avoid all possible confusions and we have no intention to reverse-engineer Arabic script, and we will still need to cater for exceptions in pronunciations, especially loanwords and dialects. alif maqṣūra and final alif (pronouncable) may also end up as "ā". Another issue is initial alif, Mahmudmasri insisted on not transliterating it and I eventually agreed. --Anatoli T. (обсудить/вклад) 00:50, 3 September 2014 (UTC)
Currently we do transliterate an initial hamza when the Arabic script writes alif-hamza, and we don't if the Arabic script has a plain alif (hamzat-al-waṣl). I think this is the right thing to do, and again it keeps the transliteration consistent with the original script. I know that Hans Wehr's dictionary doesn't transliterate initial hamza in any case but I think this doesn't make a lot of sense and it's contrary to the practice of most other Arabic books (e.g. John Mace "Arabic Verbs"). I thought what you had actually agreed was not to transliterate hamza in loanwords; if you really want to implement that, just write the Arabic without an initial hamza. I really think in general the transliteration should not gratuitously lose information, which is what happens if you omit initial hamzas.
BTW the ending اة is commonly pronounced -āt even in pausa (I'm pretty sure at least), so it could be transliterated this way. Benwing (talk) 01:42, 3 September 2014 (UTC)
Re: initial hamza. Well, that was my idea as well, all the time. I conceded to Mahmumasri's demands/requests with reluctance but he is not a very active editor. Hans Wehr doesn't usually use initial hamza, so we can't use it for reference here, anyway. Wikitiki89 seems to favour "ʾ" symbol over nothing "" (for initial hamza) as well. I agree with your point but he (Mahmumasri) had a point too. It's a matter of convention, anyway, subject to change. Yes, اة can be transliterated as "āt". --Anatoli T. (обсудить/вклад) 02:02, 3 September 2014 (UTC)

┌─────────────────────────────────┘
I've turned it on by default. Benwing (talk) 01:12, 4 September 2014 (UTC)