Wiktionary:Votes/pl-2018-12/Allowing attested romanizations of Sanskrit

Allowing attested romanizations of Sanskrit
Voting on: When citations can be provided showing that a romanization of a Sanskrit word is attested in a string of transliterated Sanskrit text (used to convey meaning in permanently recorded media in at least three independent instances, spanning at least three years; see, e.g., ), a Sanskrit entry for that romanization consisting of the modicum of information needed to allow readers to get to the native-script entry should be included, as a minimum. This proposal makes no statement about whether more than modicum should be included; it ensures that, as a minimum, modicum can be included.

Schedule:
 * Vote starts: 00:00, 7 January 2019 (UTC)
 * Vote ends: 23:59, 6 April 2019 (UTC)
 * Vote created: Dan Polansky (talk) 10:51, 31 December 2018 (UTC)

Discussion:
 * [[Image:Wikt rei-artur3.svg|20px]] Votes/pl-2014-07/Allowing well-attested romanizations of Sanskrit
 * [[Image:Wikt rei-artur3.svg|20px]] WT:RFDE
 * [[Image:Wikt rei-artur3.svg|20px]] Votes/pl-2011-09/Romanization of languages in ancient scripts 2
 * [[Image:Wikt rei-artur3.svg|20px]] Beer_parlour/2018/January
 * [[Image:Wikt rei-artur3.svg|20px]] Wiktionary talk:Votes/pl-2018-12/Allowing attested romanizations of Sanskrit

Support

 * 1) . As someone who has looked up romanized Sanskrit words in English text (not being used as English words) and not found them on Wiktionary, I think this would be helpful. Andrew Sheedy (talk) 18:09, 11 January 2019 (UTC)
 * Wut? If I search for, शब्दकोश is the first entry to come up. -- 10:07, 18 January 2019 (UTC)
 * A similar search for mahā does not fare so well: see Wiktionary talk:Votes/pl-2018-12/Allowing attested romanizations of Sanskrit. --Dan Polansky (talk) 18:35, 19 January 2019 (UTC)
 * You consider a result positively matching the first entry in Sanskrit not faring well? -- 23:41, 21 January 2019 (UTC)
 * It seems you are confusing Sanskrit with Devanagari. How is the reader of a Latin-script text such as this going to know that the script which they do not know and found in search results is to be associated with Sanskrit? --Dan Polansky (talk) 07:55, 26 January 2019 (UTC)
 * I'm referring to the Sanskrit entry of महा. I see no problem in searching for Sanskrit words using Latin transcriptions. -- 08:03, 28 January 2019 (UTC)
 * How is the reader of a Latin-script text [...] going to know that the script which they do not know and found in search results is to be associated with Sanskrit? I mean the script used in महा entry. --Dan Polansky (talk) 11:16, 23 February 2019 (UTC)
 * 1)  for IAST romanisations (those usually encountered), but not for other or ad hoc schemes. —Μετάknowledge discuss/deeds 20:12, 11 January 2019 (UTC)
 * Metaknowledge: Can you please clarify whether this conditional or restricted vote that you cast is in accord with your understanding of proper voting procedure, as you understand it? (I am not objecting at all; I am merely trying to confirm what I see.) --Dan Polansky (talk) 12:05, 12 January 2019 (UTC)
 * Yes, I think this kind of vote is acceptable, although I acknowledge that some disagree. I would probably be more strategic and leave out conditions if this were a well-written vote that had a good chance of passing, but unfortunately it is neither of those things. —Μετάknowledge discuss/deeds 18:17, 12 January 2019 (UTC)
 * Thank you. I am looking forward to see your vote about Sanskrit, one which you think has a good chance of passing. In fact, the proposal of the present vote nearly passed multiple years ago, so I don't think the chances are that bad. --Dan Polansky (talk) 05:37, 13 January 2019 (UTC)
 * I wish I had seen some discussion of this vote beforehand so I could help. (Perhaps you posted in the BP and I missed it?) Unfortunately, there tends to be some vote fatigue, so we'll have to wait a while before bringing the same question to a vote once again. —Μετάknowledge discuss/deeds 06:04, 13 January 2019 (UTC)
 * I think we do not need to wait before we bring an amendment of a running vote to Beer parlour discussion. To the contrary, the subject of the vote is activated in people's minds anyway so there is some economy of cognition in doing so. In any case, I think the restricted vote you cast is a very productive and unbureaucratic way of doing things, and I appreciate it. --Dan Polansky (talk) 06:34, 13 January 2019 (UTC)
 * 1)  Support per the rationale on the talk page. In sum, let us create the best experience for our readers that we know how, and let us make no artificial restrictions on the WT:CFI's general principle that "A term should be included if it's likely that someone would run across it and want to know what it means", with the use of the core evidence-based (as opposed to analysis-based) principle of CFI, which is the attestation requirement. --Dan Polansky (talk) 05:39, 13 January 2019 (UTC)
 * 2)  Words should be entries that can be looked up here, no matter what the script.--Prosfilaes (talk) 23:31, 21 January 2019 (UTC)
 * 3)  I would prefer allowing IAST romanizations of all Sanskrit words, regardless of attestation (i.e. treating Sanskrit like Gothic, Chinese, and Japanese), but if this weakened proposal is the only one that can get consensus, I'll support it. —Mahāgaja · talk 11:58, 25 January 2019 (UTC)
 * 4)  per Andrew Sheedy and others. --Droigheann (talk) 19:29, 17 February 2019 (UTC)
 * 5)  - if a reader is likely to come across something like this in print (which is what attestation is a test of), then we should provide a definition for it. bd2412 T 01:12, 19 February 2019 (UTC)
 * 6)  per Andrew Sheedy and Mahagaja. This may be useful e.g. when there is an entry in a different language with the same spelling as the romanisation. I would also support IAST romanisations without an attestation requirement and am open to backing a more stringent attestation requirement for non-IAST romanisations. ←₰-→  Lingo Bingo Dingo (talk)  14:41, 23 March 2019 (UTC)
 * Allowing all and only IAST romanizations would make far more sense, but this is better than nothing. Hölderlin2019 (talk) 18:34, 1 April 2019 (UTC)
 * Struck as the user is ineligible to vote as per Voting policy. &mdash; surjection &lang;?&rang; 18:44, 1 April 2019 (UTC)
 * I've been editing the Sanskrit entries rather conscientiously for a while, primarily as here. I've only registered this account recently at the explicit request of various editors on the discord. Hölderlin2019 (talk) 04:03, 2 April 2019 (UTC)
 * Doesn't matter. You don't meet the minimum requirements yet to vote. -- 04:50, 2 April 2019 (UTC)

Oppose

 * 1)  Because of being coupled to attestation. As I said: “I would understand it if one created just mechanically, that is by bot, all romanized forms, but attesting romanizations I do not understand, this siphons off the limited attention of editors.” It will be a cringe experience if editors start to attest romanizations instead of the words themselves. I would perhaps ignore this vote if it were just about mechanical additions like with Pinyin or with Serbo-Croatian where one script entails the other and there being two scripts does not double the attestation requirements, but this vote’s differentiation has no rationale I can follow. Fay Freak (talk) 20:29, 11 January 2019 (UTC)
 * The attestation requirement is linked to the rationale: if the form is actually used, then it is included. In that sense, the proposal is merely a confirmation of what is already entailed in WT:CFI as currently written. If this vote passes (a big if), we could create another vote that lifts the attestation requirement for IAST. The main thing is to have a clear statement that editors are allowed to create the best experience for our readers that we know how, as long as they consider that to be the wise use of their resources. To prevent editors from helping our readers so that the editors would have their resources conserved is an undue patternalism, a violation of Mill's principle. --Dan Polansky (talk) 12:06, 12 January 2019 (UTC)
 * @Fay Freak: Can you please indicate what is your position toward automatic IAST Sanskrit? Above, it seems you would be okay with it; is that accurate? --Dan Polansky (talk) 12:52, 26 January 2019 (UTC)
 * Yes. Okay I would be with it. I would not be particularly for it, since as I said on the verso page the cost-benefit assessment does not lean well into the positive sector. If I support it can only be appeasement – weak support perhaps. But apparently most editors prefer to gently push into getting off the Latin script and using the actual scripts, which is sane. Fay Freak (talk) 13:36, 26 January 2019 (UTC)
 * : See Votes/pl-2014-07/Allowing_well-attested_romanizations_of_Sanskrit. -- 20:36, 11 January 2019 (UTC)
 * @Victar: Which of the reasons for opposing in Votes/pl-2014-07/Allowing well-attested romanizations of Sanskrit do you agree with? Can you please identify at least of them? --Dan Polansky (talk) 11:27, 12 January 2019 (UTC)
 * 1)   —*i̯óh₁n̥C[5] 23:44, 14 January 2019 (UTC)
 * @JohnC5: Would you care to provide a rationale for your oppose vote? --Dan Polansky (talk) 09:14, 19 January 2019 (UTC)
 * 1)  When Romanisations are allowed as entries, they are eventuall treated as regular native scripts, native word, a replacement for difficult script, which is wrong. For example 犬, 狗, いぬ, イヌ are all acceptable Japanese words/spellings but [[inu]], written in Roman letters is not a Japanese word, it's only a romanisation. Similarly, महा is a Sanskrit word but [[mahā]] is not. My position hasn't changed since the last vote on Sanskrit romanisations. --Anatoli T. (обсудить/вклад) 11:46, 19 January 2019 (UTC)
 * "When Romanisations are allowed as entries, they are eventually treated as regular native scripts": I don't think that's true. Gothic aflagjandans is clearly marked as a "Romanization of 𐌰𐍆𐌻𐌰𐌲𐌾𐌰𐌽𐌳𐌰𐌽𐍃", not as a "Latin spelling of 𐌰𐍆𐌻𐌰𐌲𐌾𐌰𐌽𐌳𐌰𐌽𐍃" (so there's no way to mistake it for a native-script entry), and it's going to stay that way.
 * Also, I think it's more wrong to treat as an English word than as a Romanized Sanskrit word. Per utramque cavernam 12:50, 19 January 2019 (UTC)
 * Let me first thank you for a reasoned oppose. Now, as for "महा is a Sanskrit word but mahā is not": In fact, a Sanskrit word is an auditory phenomenon that can be put down on paper, screen or other medium either as महा or mahā, whichever is considered to be more convenient. The proposal of the vote has a link to a source that chooses mahā as the means of representation of the word. Strictly speaking, neither महा or mahā are Sanskrit words but rather means of recording or showing a Sanskrit word in a medium. For the sake of convenience and brevity, we often say things like 'cat is an English word', and in that sense, both महा or [[mahā]] are Sanskrit words, alternative forms. --Dan Polansky (talk) 14:30, 19 January 2019 (UTC)
 * But is IAST really less ‘native’ than e. g. Devanāgarī? Sanskrit has been written down in an insane number of writing systems, all of which its core predates by centuries. Sanskrit texts have been traditionally transmitted orally anyway. Sure, Devanāgarī is a kind of academic standard; but so is IAST. Guldrelokk (talk) 18:20, 19 January 2019 (UTC)
 * 1)  non-IAST Romanisation entries.  otherwise (see below). Per utramque cavernam 12:33, 26 January 2019 (UTC)
 * 2)  --Vahag (talk) 08:47, 13 February 2019 (UTC)
 * 3)  because it's a completely inconsistent and useless policy. I see no one asking the same for Old Church Slavonic or Avestan (or Ugaritic as Fay Freak mentioned on the talk page), which use scripts that are just as or even more likely to be replaced by transliterations in modern texts. When Dan Polansky starts making arguments such as "a Sanskrit word is an auditory phenomenon that can be put down on paper, screen or other medium either as महा or mahā, whichever is considered to be more convenient" and "How is the reader of a Latin-script text such as this[3] going to know that the script which they do not know and found in search results is to be associated with Sanskrit?" I think I can safely say he has no clue what he is talking about in the context of Sanskrit and that this vote is a waste of time especially for people who would rather be working on making actual Sanskrit entries. It seems like the purpose of this vote is to improve the user experience for people who mainly use the Latin script, but that rationale can be applied to so many other language that a vote for Sanskrit only doesn't really make sense. Victar already made the point that searching for the IAST usually brings up the Devanagari entry among the tope few results which means finding Sanskrit entries through only Latin script is not cumbersome at all. Also why is this vote going to last 3 months? This whole thing just reeks pointless bureaucracy to me. My position is unchanged from the last discussion. TLDR: Lol, no in the words of JohnC5. —AryamanA (मुझसे बात करें • योगदान) 17:57, 19 February 2019 (UTC)
 * [[Image:Emoji u1f44f.svg|20px|link=]] -- 18:51, 19 February 2019 (UTC)
 * You don't see anybody asking for other languages because you're not paying attention and your example languages aren't as major as Sanskrit. We have romanizations for Japanese and Gothic, so adding it would be more consistent.--Prosfilaes (talk) 04:49, 23 February 2019 (UTC)
 * As for "I think I can safely say he has no clue what he is talking about in the context of ...": Instead of addressing the arguments quoted (playing the ball), you play the man. My arguments are left without response. This vote contains multiple references to texts using Latin Sanskrit. The supporters have provided both arguments and evidence; the opposers have provided hand waving. The 3 months are there because the previous vote lasted 6 months to get enough voters in the first place. The vote is obviously not pointless bureaucracy; it is an attempt to see whether consensus has changed. --Dan Polansky (talk) 11:10, 23 February 2019 (UTC)
 * 1)  per various arguments above. —Suzukaze-c◇◇ 02:20, 21 February 2019 (UTC)
 * 2)  DerekWinters (talk) 02:17, 25 February 2019 (UTC)
 * 3)  Allahverdi Verdizade (talk) 10:53, 1 April 2019 (UTC)

Abstain

 * 1) . I tend to, as I would prefer this to having a bunch of English entries for words that aren't English (, , etc.). However:
 * 2) * I agree with Fay Freak that it would make more sense to have romanisation entries for everything, not just attested romanisations.
 * 3) * The fact that all the contributors to Sanskrit content have disagreed with this must count for something. Per utramque cavernam 10:13, 18 January 2019 (UTC)
 * I can't help but find it unfortunate that an inactive user who has little to no connection to the issue at hand, like Prosfilaes, can have equal voting power to people that work in the area every day. -- 03:01, 22 January 2019 (UTC)
 * Great way to encourage people to work on Wiktionary. If you want to build your own little Sanskrit dictionary, go elsewhere; Wiktionary is a multilingual dictionary where there should be a consistent style and what's done in one language affects what's done in another language.--Prosfilaes (talk) 02:08, 24 January 2019 (UTC)
 * I certainly encourage you to contribute to the languages you vote on. -- 02:47, 24 January 2019 (UTC)
 * Let me add that I've just remembered a collection has mentioned in a previous discussion: the . These are bilingual English-Sanskrit editions of Sanskrit classics, and the original text is entirely written in IAST (an example here). To me, this is a rather strong argument for why we should even have alternative spelling entries (I wouldn't push for that though; I'd be fine with Romanisation entries). Per utramque cavernam 12:33, 26 January 2019 (UTC)
 * I don't object to automatic IAST Sanskrit, but let me point out that the text you brought forward (for which I thank you) pertains to attestation. You say, look, here is a complete text. And the idea is, when it is attested in use, we include it. Let me again emphasize in use, as contrasts to mention. --Dan Polansky (talk) 12:43, 26 January 2019 (UTC)
 * Yes, I agree. What I meant is that it seems a mistake to me to ignore entire books published in IAST. Per utramque cavernam 12:49, 27 January 2019 (UTC)
 * abstain per above. —Suzukaze-c◇◇ 19:45, 24 January 2019 (UTC)
 * Vote changed to oppose. Per utramque cavernam 14:11, 22 February 2019 (UTC)
 * 1)  Not too familiar with Sanskrit and issues surrounding it, so I'm reluctant to vote. But I agree with a lot of points raised by the proponents of this vote, as well as the opposition to the attestation requirement raised by both proponents and opponents of the vote. It seems to me that IAST-based romanization entries w/o attestation requirement would be ideal (mirroring the romanizations of Gothic, Japanese etc. which have perfectly fine romanization systems as has been brought up already in the discussion above), so long as they can be generated automatically based on an unambiguous reading of the native script (usually impossible with abjads, but then Devanagari is a syllabic script so I think we should be fine). I'm not sure why these votes so often hang on the attestation requirement: romanizations are essentially soft redirects, they're basically there to compensate for an imperfect search engine. The attestation requirement should imo only be for the entry in the native script(s), not the romanization that links to it which should only exist if the native script entry exists anyway. (If there should exist somewhere a romanization entry without a native-script entry, such an entry could be categorized automatically, as with Category:Gothic romanizations without a main entry, and the reverse should also be possible.) — Mnemosientje (t · c) 16:40, 25 January 2019 (UTC)
 * I think the larger important question is whether we can be sure to include Latin-script Sanskrit at all. If we can get attested Latin-script Sanskrit to pass (a big if indeed), my hope would be that it shouldn't be too hard to pass an automatic IAST Sanskrit via a follow-up vote. My guess is that most opposers of Latin-script Sanskrit oppose both proposals. As for native script, is Devanagari really native script of Sanskrit? Sanskrit tells us that "Sanskrit texts dated to the 1st millennium CE were written in the Brahmi script, the Nāgarī script, the historic South Indian scripts and their derivative scripts.", and has it sourced. --Dan Polansky (talk) 07:27, 26 January 2019 (UTC)
 * 1) . If there is no universal policy stating that any script is allowed for any language as long as it is attested, then I would prefer to leave questions of Sanskrit up to Indian people. פֿינצטערניש (talk) 23:41, 26 January 2019 (UTC)
 * The thing is, this is the English Wiktionary, serving above all native English speakers and speakers who have English as the second language. These people are necessarily acquainted with Latin script, but not necessarily with other scripts. The dictionary does not serve exclusively readers from India. --Dan Polansky (talk) 09:58, 27 January 2019 (UTC)
 * And as a native English speaker who isn't from India and uses Wiktionary, I want the quality of Wiktionary to be as high as possible, which is why I think it's a good practice to refrain from making it worse through voting out of ignorance. פֿינצטערניש (talk) 21:50, 13 February 2019 (UTC)
 * I do not see how the proposal reduces the quality of Wiktionary or makes it worse; to the contrary, on the talk page I argued that the proposal makes Wiktionary better by improving its usability, which I showed in greater detail for a particular use case. --Dan Polansky (talk) 10:01, 16 February 2019 (UTC)
 * per some of the above. ←₰-→  Lingo Bingo Dingo (talk)  10:47, 28 January 2019 (UTC)
 * Vote changed to support. ←₰-→ Lingo Bingo Dingo (talk)  14:41, 23 March 2019 (UTC)

Decision
8-10-3, Fails. &mdash; surjection &lang;?&rang; 19:07, 8 April 2019 (UTC)