Wiktionary:Votes/pl-2018-04/Image policy

Voting on: Adopting the following image policy, to be placed to WT:EL:

Rationale: on the vote talk page.

Schedule:

• Vote starts: 00:00, 28 April 2018 (UTC)
• Vote ends: 23:59, 27 May 2018 (UTC)
• Vote created: Dan Polansky (talk) 09:03, 21 April 2018 (UTC)

Discussion:

• Wiktionary:Beer parlour/2018/April#Vote: CFI and images
• Wiktionary talk:Votes/pl-2018-04/Image policy

1. Support per vote talk page: Make it clear there is an express agreement on having images both in English and non-English entries, and regulate certain cases that recently appeared in the mainspace and are being removed, including diff. --Dan Polansky (talk) 10:30, 28 April 2018 (UTC)
2. Support. My rationale is in the Beer Parlour convo. edit: Images should help to clarify what the word means or refers to, or serve as an attestation of the word. If they don't do that, then they're out of place. --Per utramque cavernam 08:56, 29 April 2018 (UTC)

   @Per utramque cavernam: Which convo do you mean? I see nothing in Wiktionary:Beer parlour/2018/April#Vote: CFI and images. --Dan Polansky (talk) 09:10, 29 April 2018 (UTC)

   Whoops, I was referring to Wiktionary:Beer parlour/2018/April § Image captions and descriptions of the representing objects such as paintings but that's about another issue. --Per utramque cavernam 09:31, 29 April 2018 (UTC)

3. Support. In SG's example of Aristotle's picture at Aristotelian, this policy need not exclude it, because "another rationale for inclusion applies" (he is the referent of the term, not merely the etymology) -- though I'd have thought Aristotle to be the better entry for that image. Equinox ◑ 09:13, 29 April 2018 (UTC)

   Why not in both places? :) — SGconlaw (talk) 13:11, 29 April 2018 (UTC)

   Same reason for putting a fruit pic at orange vs. orangey. It depicts the former more closely than the latter; anyone who looks up orangey and doesn't know what an orange is will be heading to that entry anyway. Equinox ◑ 13:19, 29 April 2018 (UTC)

   With orange, an orangey object may look like an orange. With Aristotle, an Aristotelian object does not look like Aristotle; rather, the adjective points to Aristotelian thought, which is far from easy to depict. --Dan Polansky (talk) 14:32, 29 April 2018 (UTC)

   If that's the criterion then it seems your original vote should have been phrased more along those lines ("images must look like the thing named by the headword"). Equinox ◑ 14:46, 29 April 2018 (UTC)

   My mistake, then. The proposal does some useful things as is, even if its application to Aristotelian is less than clear. --Dan Polansky (talk) 14:58, 29 April 2018 (UTC)

   Hm. About "he is the referent of the term": I don't actually think so. A referent of "Aristotelian" is anything that is Aristotelian, I'd say. While it is above all nouns that have referents, we can make a trick with adjectives by considering "an Aristotelian thing", and taking the referent of the adjective to be the referent of that phrase. An alternative is to take the referent to be the same as "Aristotelianness", the property of being Aristotelian. In either case, Aristotle does not seem to be the referent of the adjective. --Dan Polansky (talk) 06:40, 1 May 2018 (UTC)

4. Support DTLHS (talk) 00:58, 30 April 2018 (UTC)
5. Support —Suzukaze-c◆◆ 05:57, 1 May 2018 (UTC)
6. Support —Μετάknowledge^{discuss/deeds} 20:01, 1 May 2018 (UTC)
7. Support. Let's be parsimonious when it comes to images, so that they don't distract from the important stuff (senses and the like). This, that and the other (talk) 10:32, 9 May 2018 (UTC)
8. Support --Daniel Carrero (talk) 03:45, 11 May 2018 (UTC)
9. Support --Victar (talk) 06:30, 14 May 2018 (UTC)
10. Support --Makaokalani (talk) 08:54, 27 May 2018 (UTC)
11. Support as a good start, but I'd prefer an image policy that allows only images of the relevant definiendum. —Mahāgaja (formerly Angr) · talk 16:27, 27 May 2018 (UTC)
12. Support - -sche (discuss) 16:35, 27 May 2018 (UTC)

1. Oppose: I feel this is unnecessarily narrow, and there is no particular reason why we should not include an image, for example, of a person (1) whose name or work is the source of a word (e.g., Aristotle for Aristotelian); or (2) who has coined a word (e.g, George Orwell for newspeak). It seems to me that it is an arbitrary choice to rule such images to be "irrelevant". — SGconlaw (talk) 12:44, 28 April 2018 (UTC)

   The reason is that such images are not lexicographical, pertaining to the services rendered by a dictionary. A key service rendered by a dictionary is to answer the question What is it? The question is answered by definitions. However, a definition often leaves some lack of clarity, which can be gained by showing an image of the referent. By contrast, once the etymology or source of coinage is stated, there is no lack of clarity remaining that can be removed by an image. Allowing etymological images in general except perhaps for some limited set of cases would lead to pretty bizarre outcomes, IMHO; for instance, we would add an image of circle to encyclopedia entry since the word stems from κύκλος (kúklos, “circle”). We can also check what lemmings (other dictionaries) are doing. I know of no dictionary containing images of people coining a term. By contrast, here is Duden:mops, showing an image of pug (dog). On a different note, 𐤉𐤓𐤅𐤔𐤋𐤉𐤌 contains an attestation image for a word for which it is likely to be hard to find a normal textual attesting quotation; that is eminently lexicographical. --Dan Polansky (talk) 08:34, 29 April 2018 (UTC)

   That's my point. We define lexicographical to mean "the art or craft of compiling, writing and editing dictionaries". I don't see why it isn't relevant to that art or craft to provide an image of a person that is related to the etymology of an entry rather than to its meaning. That is an unnecessarily narrow definition of lexicographical. As an online dictionary, we do not always have to be constrained by the limitations of print dictionaries. Also, the so-called "bizarre outcomes" scenario can always be dealt with by discussion on a case-by-case basis. — SGconlaw (talk) 09:59, 29 April 2018 (UTC)

   I did not talk of print dictionaries, which you mention; I gave an example of an online electronic dictionary, Duden. My argument is not about lack of space; it is about lack of user focus. User attention is a very scarce resource; we should not waste it. Each element on screen, part of field of vision, consumes a user's attention. Visuals are more prone to consuming attention; therefore, they need to be well justified and useful; they should not be there just for fun or curiosity. Users often consult a dictionary when they are trying to understand a piece of text in domain relevant to them; they already have their head full of their own stuff, and do not need stuffing to make their dictionary experience more interesting. This is not the only case where you are adding irrelevant or less relevant elements to Wiktionary mainspace, making it harder to use for its increased consumption of the reader's scarce resources. Your excessive identification of sources, to the point of being utterly ridiculous by my lights, is another example. My general message to you would be, stop wasting the reader's resources with omittable clutter. --Dan Polansky (talk) 10:27, 29 April 2018 (UTC)

   We get it – Wiktionary should be shaped exactly as you like it, and that anything going against that is "utterly ridiculous" and "wasting the reader's resources". I see no evidence either way as to what users want or would find useful, so unless some sort of survey is conducted, speculating about that seems pointless to me. Anyway, I remain guided by consensus on the issue. — SGconlaw (talk) 13:10, 29 April 2018 (UTC)

   That needless elements burn attention is an obvious fact, not a matter for discussion. What is for discussion is the cost-benefit analysis, that is, what borderline relevant element is still worth it. You did not seem to have made any such cost-benefit analysis, or else you would not be cluttering Wiktionary. I do not know for sure what our readers find worthwhile, but I do know that very few contributors of images contributed SG-style images, which is at least suggestive. --Dan Polansky (talk) 14:30, 29 April 2018 (UTC)

   (What are SG-style images? — SGconlaw (talk) 14:42, 29 April 2018 (UTC))

   Unfortunately Duden does have clutter in the form of advertising, in common with many other online dictionaries; Duden seems to be the worst. DP conveniently overlooks the fact that Wiktionary mercifully does not have advertising clutter. DonnanZ (talk) 13:32, 29 April 2018 (UTC)

2. Oppose: I agree with SGconlaw. I also have had images I added removed by DP, coats of arms in particular, probably with the "irrelevant" excuse. Fortunately some still survive, but I'm not going to be silly enough to point them out. Finding a suitable image on Commons can be quite a challenge at times, and I get great pleasure when I find a good one that hasn't been used anywhere else. DonnanZ (talk) 16:33, 28 April 2018 (UTC)

   There was consensus in Wiktionary:Beer parlour/2017/October#Removing images of coats of arms; based on that, I indeed proceeded to remove such images in volume. --Dan Polansky (talk) 07:05, 29 April 2018 (UTC)

   I was too late on the scene. They were relevant to the entries concerned. All you have succeeded in doing is making Wiktionary a duller place. DonnanZ (talk) 09:17, 29 April 2018 (UTC)

3. Absolutely Oppose. Why should Pythagorean theorem not contain an image of Pythagoras? Chinese: 勾股定理 (gōugǔ dìnglǐ), nothing to do with Pythagoras at all. For English learners, this name could be confusing as hell, and an image of Pythagoras could well help dispel the confusion and improve the entry's usefulness. Also, the amount of abuse directed at SGconlaw on this topic in various discussions is absolutely unbelievable. Wyang (talk) 06:11, 2 May 2018 (UTC)

   No comments allowed, thank you. This is a vote, not a discussion. Wyang (talk) 06:17, 2 May 2018 (UTC)

   This is a vote-cum-discussion, as per long-standing English Wiktionary practice. Comments are generally welcome. --Dan Polansky (talk) 17:05, 4 May 2018 (UTC)

4. Oppose, per most of the points made above. I find that the proponents of this vote have a too narrow conception of what an online dictionary can be and become (vis-à-vis print dictionaries) and of what constitutes 'usefulness' in a dictionary. — Mnemosientje (t · c) 18:15, 5 May 2018 (UTC)

   Then, if these images are "useful" for some sense of the word, what is the dictionary use case that these images serve? I posit that "I am bored" is a not a lexicographical problem and use case. By contrast, "I am in an explorative mood, not searching for anything specific" is a valid lexicographical use case, and we serve it with a variety of nagivation aids, including topical categories, clickable etymologies, lists of descendants, etymologically related terms, thesaurus, word of the day, you name it. --Dan Polansky (talk) 07:59, 6 May 2018 (UTC)

5. Oppose -Xbony2 (talk) 22:15, 7 May 2018 (UTC)
6. Oppose. Some of my rationale is below. I want to wait on this--it doesn't seem like a pressing issue, and I'd rather this be freer until just the right constraints are formulated. --SanctMinimalicen (talk) 22:39, 14 May 2018 (UTC)

   @SanctMinimalicen: I don't understand. Below, you write "I think that relating to the meaning of the word can (and should) be interpreted broadly; ...", but the proposal does not say anything about "meaning". The proposal is worded to be rather narrow in its exclusion; it excludes images included only via coinage or etymology and is even explicit in saying "unless another rationale for inclusion applies". --Dan Polansky (talk) 08:46, 15 May 2018 (UTC)

   @SanctMinimalicen: Could you please state for the record whether you agree with the removal of the image of Orwell from newspeak in diff? --Dan Polansky (talk) 07:17, 17 May 2018 (UTC)

   As I said below, agree. I realize that this is consistent with the proposed policy--my hang-up is somewhere in the word "etymology". I've been thinking about the Aristotle example again, with Aristotelian. I would want to include an image of Aristotle there, because of it's bearing on the definition; however, in that case, the definition is inextricably linked to the etymology, and I wouldn't want a policy to exclude that. I think I'd be totally fine with this policy if it was only a matter of coinage, leaving etymology up to debate. Or, perhaps, if it had an additional clause allowing for images linked to etymology if the etymology still bore a direct and obvious effect upon the meaning of the word (e.g. Aristotle -> Aristotelian). --SanctMinimalicen (talk) 23:55, 17 May 2018 (UTC)

   @SanctMinimalicen: Thank you for your answer. Could you prehaps try to explain why the "unless another rationale for inclusion applies" does not address your concern? The Aristotle -> Aristotelian case seems covered by that, doesn't it? I apologize if your above answer was intended to answer that, but it is not clear to me from reading the answer that the caveat does not handle that case. --Dan Polansky (talk) 10:52, 20 May 2018 (UTC)

   It technically covers it, it seems, but I think the wording causes it to be slanted against such cases. Because "etymology" is explicitly stated in the prohibitions, but "etymology that weights directly and heavily on the meaning" (or something to that effect) is not explicitly stated in the caveats, it will place an unduly burden of proof on the part of the latter which I don't think should be there; rather, I feel that the burden of proof should rest on the side of exclusion of images, not inclusion. --SanctMinimalicen (talk) 17:13, 20 May 2018 (UTC)

1. Abstain. My inclination is that images be lexicographical (in the "strict sense", as it were, i.e. relating to the meaning of the word specifically), and so I lean towards the side of laying constraints; that said, I think that relating to the meaning of the word can (and should) be interpreted broadly; for example, I would support the inclusion of an image of Aristotle at Aristotelian (as discussed above) because the meaning of the word is, more or less, "pertaining to Aristotle". In contrast, I wouldn't be inclined to support the inclusion of an Orwell image at newspeak because he is only in the etymology/coinage, not in the meaning. It feels encyclopedic. For comparison, I would happily include an image of a page from 1984 that includes newspeak, which does illustrate the meaning. In the end, I think a constraint is a good idea but I'm not sure this is the right policy for the job. --SanctMinimalicen (talk) 19:58, 29 April 2018 (UTC)

1. @SanctMinimalicen: For Aristotelian, Equinox points out above the part "unless another rationale for inclusion applies", and its application. You cannot include an image of Aristotle on account of etymology, but you can do so on account of semantics. I will probably support exclusion of the image of Aristotle from Aristotelian on the grounds that the adjective refers to Aristotle's thought, to which his appearance has no bearing, but the present proposal does not necessarily ensure that exclusion, as Equinox pointed out. --Dan Polansky (talk) 05:48, 1 May 2018 (UTC)

   Personally, I don't see any harm in an image of George Orwell at newspeak, as such an image may be inappropriate at Orwell, George Orwell being a pseudonym and not his real name. DonnanZ (talk) 10:03, 1 May 2018 (UTC)

   Which brings to mind that I have added at least one image of a famous bearer of a name. DonnanZ (talk) 10:40, 1 May 2018 (UTC)

2. Abstain. I'm still not convinced this is necessary, but I have no objections to it. —Granger (talk · contribs) 12:13, 4 May 2018 (UTC)

   @Mx. Granger: Could you please state for the record whether you agree with the removal of the image of Orwell from newspeak in diff? --Dan Polansky (talk) 07:16, 17 May 2018 (UTC)

   I'm neutral. I can understand the argument for removing it, but I also don't think the image does much harm. Hence my abstention: I don't really care whether we adopt this rule or not. —Granger (talk · contribs) 15:28, 17 May 2018 (UTC)

What do we do here? A previous vote with exactly two-thirds support (4 support/2 oppose) was extended to allow more participation. The difference with this vote is that 19 users have voted, which is a relatively high turnout, and extending the vote might not achieve anything. This, that and the other (talk) 02:13, 28 May 2018 (UTC)

For the case that this vote does not get extended (indeed, the turnout is good, which it was not at 4:2), my position is that 2/3 at this turnout is a pass. An example vote that was closed as a pass at even a lower threshold than 2/3 is Wiktionary:Votes/pl-2014-12/Making simplified_Chinese soft-redirect to traditional Chinese. There are probably other examples. --Dan Polansky (talk) 06:14, 28 May 2018 (UTC)

I also think 2/3 should be closed as passed—my understanding was that 2/3 is the threshold for passing, though strangely I don't see it written down in the voting policy. I'll close it that way in a day or two unless someone objects or beats me to it. —Granger (talk · contribs) 12:06, 28 May 2018 (UTC)

So is 2/3 the official threshold? I was wondering where the lines between pass/fail and no consensus lie, but I couldn't find documentation with a definition. --SanctMinimalicen (talk) 17:53, 28 May 2018 (UTC)

I too wonder why we don't have that in the voting policy. Perhaps we will need to vote on it :) But I don't object to Granger's plan. This, that and the other (talk) 08:47, 29 May 2018 (UTC)

Passes 12-6-2. —Granger (talk · contribs) 12:51, 29 May 2018 (UTC)