Wiktionary:Votes/2019-05/Excluding self-evident "attributive form of" definitions for hyphenated compounds

Voting on: Wiktionary policy should be not to include entries that define hyphenated compounds merely as an attributive form of the individual components. In the simplest case, these are definitions of "X-Y" as "attributive form of X Y", as, for example, the recently deleted entry "periodic-table: attributive form of periodic table". Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations, from "aardvark-skin handbag" to "zebra-stripe stockings" to "never-to-be-repeated opportunity". Their construction is transparent once the simple underlying principle is understood, and a dictionary should not trouble itself with defining each one individually.

Schedule:

• Vote starts: 00:00, 24 May 2019 (UTC)
• Vote ends: 23:59, 22 June 2019 (UTC)
• Vote created: Mihia (talk) 17:37, 17 May 2019 (UTC)

Discussion:

• Wiktionary talk:Votes/2019-05/Excluding self-evident "attributive form of" definitions for hyphenated compounds

1. Support DCDuring (talk) 03:35, 25 May 2019 (UTC)
2. Support — SGconlaw (talk) 07:57, 25 May 2019 (UTC)
3. Support — John Cross (talk) 08:01, 25 May 2019 (UTC)
4. Support - TheDaveRoss 16:39, 25 May 2019 (UTC)
5. Support, although I would have like for my remarks on the talk page to be addressed. Btw @Mihia, it's obvious you're in favour of the proposal, but you can cast a vote yourself. Canonicalization (talk) 16:41, 25 May 2019 (UTC)
6. Support Andrew Sheedy (talk) 22:48, 26 May 2019 (UTC)
7. Support as vote creator. Mihia (talk) 19:22, 27 May 2019 (UTC)

   @Mihia: Can you please clarify where the rationale is? Is it the following part on the vote page: "Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations, from "aardvark-skin handbag" to "zebra-stripe stockings" to "never-to-be-repeated opportunity". Their construction is transparent once the simple underlying principle is understood, and a dictionary should not trouble itself with defining each one individually." --Dan Polansky (talk) 08:20, 31 May 2019 (UTC)

   Assuming now that the above is the rationale, the rationale is countered by pointing out that 1) only attested forms can be included, which puts a limit on the combinations, and 2) only not sum of parts can be included as per current WT:CFI, and therefore, aardvark-skin mentioned in the rationale is already excluded (aarvark skin is not includable), which is another effective limit on the combinatorial explosion. --Dan Polansky (talk) 09:05, 31 May 2019 (UTC)

   The rationale is that hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations according to a simple mechanical rule, and that not only is it not necessary for a dictionary to individually define these (no other dictionary does, as far as I am aware), it is positively invidious since it gives the false impression that there is something special about the ones that are included. The user is better off understanding the underlying principle, that they can then apply to any similar case that they encounter, rather than depending on looking up individual cases in a dictionary, and imagining that other cases don't exist when they are not specifically listed in a dictionary. Attestation is unimportant when any (reasonable) coinage that is made according to the rules is automatically correct English. Are you sure that the present CFI excludes hyphenated compound modifiers whose unhyphenated form is SOP? Where does it say that? I have seen the opposite argued in RFD discussions: that hyphenated compounds are to be treated as single words, and so should all be includable irrespective of SOP rules. In fact, there cannot be a general such exclusion for hyphenated compounds, otherwise we would exclude e.g. "fig-leaf" on the grounds that "fig leaf" is SOP, which surely we would not want to do (in fact that isn't the greatest example since "fig leaf" also has, um, fig-urative meanings, but you get what I'm saying). Mihia (talk) 20:43, 31 May 2019 (UTC)

   Thank you. As for "according to a simple mechanical rule", so are English -s plurals. When black hole is inclusion worthy (as it is), we also include the plural black holes even though it is created via a simple mechanical rule. As for "Are you sure that the present CFI excludes hyphenated compound modifiers whose unhyphenated form is SOP?": It does not say that explicitly; it is my reading; since, per WT:CFI#Idiomaticity, "An expression is idiomatic if its full meaning cannot be easily derived from the meaning of its separate components." The full meaning of aardvark-skin can be obtained from its separate components: 1) aarvark, 2) hyphen, 3) skin. As for "I have seen the opposite argued in RFD discussions: that hyphenated compounds are to be treated as single words", I have seen supporters of that, but they are a minority; that would be handled by a vote to the effect of "Component words separated by a space or a hyphen are considered to be separate for the purpose of idiomaticity". --Dan Polansky (talk) 05:55, 1 June 2019 (UTC)

   Hmm. My reading of that CFI rule is that "expression" refers to two or more words separated by spaces. If it referred to hyphenated words then, as I mentioned, any ordinary (non-attributive) hyphenated word, say a hyphenated noun, that means the sum of its parts would be excluded. Is that actually the intention? It seems as if this point may need clarifying in a separate discussion somewhere. Mihia (talk) 10:33, 1 June 2019 (UTC)

   Yes, hyphenated sum-of-parts adjective-noun and noun-noun compounds are liable to be deleted, baring WT:COALMINE. Thus, apple-tree could be deleted but is protected by WT:COALMINE since appletree is attested. Some people disagree and say "apple-tree" is a single word since it does not use space for separation, but from what I have seen, they are a minority. --Dan Polansky (talk)

   OK, I raised this here. Mihia (talk) 19:21, 6 June 2019 (UTC)

   Policy changes are usually discussed in Beer parlour, and therefore, Wiktionary talk:Criteria for inclusion#Treatment of SOP hyphenated compounds is rather uncustomary and is less likely to draw attention --Dan Polansky (talk) 07:48, 7 June 2019 (UTC)

   Thank you, yes, after I added the question on the CFI talk page I noticed the surprisingly low traffic there, so I also added a link at BP. Mihia (talk) 23:27, 7 June 2019 (UTC)

8. Support DTLHS (talk) 21:17, 28 May 2019 (UTC)
9. Support İʟᴀᴡᴀ–Kᴀᴛᴀᴋᴀ (talk) (edits) 14:24, 30 May 2019 (UTC)
10. Support Aabull2016 (talk) 16:06, 3 June 2019 (UTC)
11. Support We don’t arbitrarily includes segments of words. And the waste is particularly terrible since if a user ever searches such fragments he will be redirected to the hyphenless form. If the hyphenated form is not used as such then it just isn’t anything. Fay Freak (talk) 17:13, 5 June 2019 (UTC)
12. Support  --Lambiam 21:29, 13 June 2019 (UTC)
13. Support Equinox ◑ 21:34, 13 June 2019 (UTC)

1. Oppose except of course for such terms as are SOP (viz, as their base forms with spaces instead of hyphens are SOP) or unattested.​—msh210℠ (talk) 20:27, 25 May 2019 (UTC)

   Or, in other words, oppose per Dan P., who expressed it better, below.​—msh210℠ (talk) 12:26, 3 June 2019 (UTC)

2. Oppose 1. Some non-attributive hyphenated compounds are only included with attribute-form definition templates, e.g. snapping-turtle. Though my reading is that this proposal only disallow attributive forms and allows non-attributive hyphenated compounds as alternative forms, "not to include entries that define hyphenated compounds merely as an attributive form of the individual components" leaves the possibility open that those non-attributive alternative-form entries may be removed even though they are (also/only) attestable as non-attributive simply because a wrong or incomplete definition-line template was used. I don't think that bulk deletion of this material is a good way forward. 2. I think there should be a general presumption against including attribute forms as definitions, but that other considerations (frequency, multiple lemmings, a word also occurring as a hyphenated compound) should override that. ←₰-→ Lingo ^Bingo _Dingo (talk) 10:29, 27 May 2019 (UTC)

   @Lingo Bingo Dingo: To respond to point 1, the intention is definitely not to exclude "alternative form of" entries for hyphenated compounds. I don't really see how the possibility that some entries may be misdefined or incompletely defined is grounds for voting against the proposal. However, please note that I am not advocating automated bulk deletion, so each case, if manually deleted when encountered, can be assessed on its merits in respect of this point. Mihia (talk) 19:31, 27 May 2019 (UTC)

   Thanks for clarifying, so that point is more or less moot then. To be honest I wouldn't exactly mind bulk deletion of attributive forms that are clearly of the "-" type. ←₰-→ Lingo ^Bingo _Dingo (talk) 10:08, 28 May 2019 (UTC)

3. Oppose If the base term is inclusion worthy, so should be the hyphenated forms. We do include nominal (noun) hyphenated variants of nominal compounds; for instance, we include apple-tree in addition to apple tree. An example: the base term periodic table is not sum of parts and is worthy of inclusion, and therefore, periodic-table should be included as well. By contrast, green leaf is a sum of parts, and therefore, green-leaf should not be included even if attested. On whether periodic-table is a sum of parts: it is only a quasi sum, of 1) periodic table and 2) an operator that takes a space-separated nominal compound and turns it into a hyphenated attributive form that behaves like an adjective. It is quasi since it does not meet the CFI definition of sum of parts, per WT:CFI#Idiomaticity. That said, the likening to possessives like cat's has some force. As for "Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations", I only support inclusion of attributive forms of includable base forms, so most of the hyphenated combinations would not be included even if attested, like aardvark-skin, and indeed, aardvark-skin actually is a sum of parts.

   For reference, this vote was created from RFDE on intensive-care, which uses {{attributive form of}}. The search insource:/\{\{attributive form of* / finds over 500 entries, including decision-making, criminal-law, sea-lion, trade-union, etc. --Dan Polansky (talk) 07:48, 31 May 2019 (UTC)

4. Oppose A hyphenated word is still a word, and we include ALL of those. SemperBlotto (talk) 10:36, 1 June 2019 (UTC)

   @SemperBlotto: no, we don't. We don't include most words, in fact. Any word which is typically lower-cased but which is capitalized at the beginning of a sentence is not included in its capitalized form, which would nearly double the count of allowable words. Similarly most nouns can be formulated in the possessive, but we do not include every noun + 's. Most verbs have an infinitive form which is prepended with to, we don't include all of those, nor the several formulaic verbal forms which are multi-word forms in English (present progressive, etc.). We specifically exclude all of these grammatical forms, and instead focus on lexical forms, because we are a dictionary and not a grammar. - TheDaveRoss 00:56, 3 June 2019 (UTC)

5. Oppose Per Dan Polansky. Jjamesryan (talk) 05:42, 3 June 2019 (UTC)
6. Oppose because the attributive form often corresponds to a different lemma in the translation section Ketiga123 (talk) 05:41, 17 June 2019 (UTC)

   Oppose a little reluctantly only because the wording is too vague. I would support excluding the examples given, but this is said to be "the simplest case". What other cases might there be? I would not want to exclude (what I consider) genuine adjectives whose meaning is transparent from their component parts, but I would agree that we can generally exclude specific attributive hyphenated uses of noun phrases. Ƿidsiþ 06:15, 31 July 2020 (UTC)

   Unindented: this is a late vote. PUC – 10:09, 31 July 2020 (UTC)

1. Abstain —Suzukaze-c◇◇ 10:15, 28 May 2019 (UTC)
2. Abstain I'm tempted to cast my vote in opposition, because I just don't see on what basis we would exclude one hyphenated compound (if it were only attested in attributive positions, or even, if it were an alternative spelling in non-attributive positions in two books, but not three) but include another (if it were attested in non-attributive positions in three books, not just two). OTOH, we do exclude some other "regular" "alt" forms, like we don't have an entry for These saying it's a sentence-initial form. I will say that in any situation where a form is attested both as an attributive form and as an alternative form, I tend to think there should only be one definition line, something like "{{lb|en|chiefly|attributive}} {{altform|foobar}}". - -sche (discuss) 05:30, 4 June 2019 (UTC)

   @-sche: Can you stick an {{abstain}} template in your vote? The tracker isn't currently counting you as having voted. If you would rather not the tracker isn't terribly important. - TheDaveRoss 12:49, 5 June 2019 (UTC)

    Done. - -sche (discuss) 16:53, 5 June 2019 (UTC)

3. Abstain. I am undecided, and don't really see the need for a vote at all. DonnanZ (talk) 12:47, 4 June 2019 (UTC)

   The vote is good for enabling adoption of a systematic approach rather than doing things on case-by-case basis and repeating the same arguments and counterarguments again and again in RFDE. The latter approach is workable, but a vote is even better. This is a good vote to have. --Dan Polansky (talk) 09:18, 7 June 2019 (UTC)

4. Abstain While I probably wouldn't trouble to create such pages myself, neither do I feel a particular need for policy against such things. Cnilep (talk) 06:19, 21 June 2019 (UTC)

Passes 13–6.​—msh210℠ (talk) 22:38, 22 June 2019 (UTC)

Indeed. What is depressing is that this vote passed with a demonstrably false rationale using "aardvark-skin handbag" as an example, and the support section votes in the great majority of cases contain no evidence that the voters even noticed there is a key distinction to be made and they actually have reasons to disregard that distinction or what the reasons were. Well, the loss is not that great and the proposal has no unbearable consequences, but the level of discussion is depressing. --Dan Polansky (talk) 11:37, 28 June 2019 (UTC)

@Dan Polansky: there was discussion, you did not participate in it. Also, just because supporters didn't explicitly mention your concerns in their vote doesn't mean they were unaware; I for one don't agree at all that there was "false rationale" involved in the vote. I voted in support because I don't think we ought to have "attributive form" entries for any terms, whether or not the non-attributive form is inclusion worthy. That makes your quibbles about the "aardvark-skin" example entirely moot. Please don't assume that people who disagree with you (which seems to be the majority in almost every vote) are ill-informed or lazy; your concerns are your own and nothing says any other voter must share or even address them. - TheDaveRoss 12:47, 28 June 2019 (UTC)

Your vote does not say "per discussion on the talk page" so it does not provide any traceability to any specific arguments and counterarguments. The argument of the rationale "Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations ..." is obviously wrong since attestation is a limit and sum of parts criterion is a limit, and indeed aardvark-skin would be deleted as a sum of parts even without this vote and therefore, it does not support the vote's proposal. And the two arguments that you responded to in that discussion do not pertain to the distinction that I mean, viz aardvark-skin (sum of parts) vs. periodic-table (not sum of parts), so your response is more evidence of people still not taking note of this distinction, like I said. That said, the vote has passed and I do not dispute the passing; we need functional and timely decision making even at the risk of passing suboptimal things. --Dan Polansky (talk) 13:12, 28 June 2019 (UTC)

I now noticed that msh said "Hyphenated compound modifiers can be created in combinations neither arbitrary nor limitless but only when attested", on 15 May 2019 in RFDE for intensive-care (later at Talk:intensive-care), and this vote was created on 17 May 2019. The arguments that msh made before the start of the vote were conveniently ignored and the same blatant falsity was repeated in the vote's rationale even though it was refuted before. None of the supporter chose to comment on the blatant falsity at the top of the page. I think the reader can understand that this is frustrating. --Dan Polansky (talk) 17:06, 28 June 2019 (UTC)

@Dan Polansky: Where does it say that a voter must provide rationale for their vote? That is a standard which you have created, and nobody is bound by your opinions on how votes ought to be run. Votes are not debates, where the voters must support their votes, it is the prerogative of each voter whether or not to expand on their vote with rationale or clarification.

Also the "blatant falsity" is based on a willful misreading of other people's statements. The statement at the top of the vote says that Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations, from "aardvark-skin handbag" to "zebra-stripe stockings" to "never-to-be-repeated opportunity". - nowhere does it say that all of the myriad possible combinations would meet the criteria for inclusion. That is your own reading of the statement, and you read it incorrectly. Created in this case means formed grammatically, not entered into the dictionary per the criteria for inclusion. If you claim that the statement as actually written rather than as you misinterpreted it is fallacious, well, then you are just wrong.

Further, you commit the same sin you accuse others of, you did not address any of the supporting arguments in your opposition vote. I think all of this arguing you are doing after the fact is just unnecessary histrionics and acting the martyr. - TheDaveRoss 17:35, 28 June 2019 (UTC)

I am not sure this discussion after the vote closure is a good thing, but I am not sure it is a bad thing either. I am not a martyr in relation to this vote; I suffered no personal loss. I wish the votes were more like requests for comments, and I wish we could introduce either a requirement or strong recommendation that people leave comments on their votes, like the opposers did in this vote. I do not claim such requirement already exists. I do think those who claim that votes are evil might agree with me that cast votes should better either state the rationale or link to rationale, like people sometimes do when they say "per nom", "per Joe Hoe", "per discussion so and so", etc. Again, this is not a present formal requirement, and I am not saying as much. As for "you did not address any of the supporting arguments", I did address the thing with the possessive that someone raised and I addressed "Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations" as you can see above; I may have failed to address some arguments. I posted a writeup rather than silence, so I did at least part of my homework. Let's be accurate. --Dan Polansky (talk) 17:48, 28 June 2019 (UTC)

Even if we admit that I misconstrued "Hyphenated compound modifiers can be created in arbitrary and virtually limitless combinations", it is still blatantly wrong as an argument since the direct analogue would be "-ness nouns can be created from adjectives in virtually limitless combinations"; the sentence does not say anything to the effect that hyphenation is what makes all the difference as contrasted to combinations created without space and hyphen. --Dan Polansky (talk) 17:52, 28 June 2019 (UTC)