These do seem quite SoP-ish. Still, I'm of the opinion that, if reasonable people can disagree about whether a phrase for which we have an entry is sum of parts, we should keep the entry. That is, I would rather have 100 well-formed entries for phrases that might be sum of parts than miss 1 entry for a phrase that is not sum of parts. Polywords and proper nouns are the two greatest gaping holes in our English coverage, and no-one will ever bother to work on them if each entry is liable to be deleted.
This entry and the two following were created by two intelligent, reasonable editors, and were nominated for deletion by another intelligent, reasonable editor. To me, this seems like prima facie evidence that reasonable people can disagree as to whether these phrases are suitable for inclusion. Thus, without any actual consideration of the merits, I will happily vote keep. :-) -- Visviva 07:44, 30 January 2009 (UTC)
Keep per Visviva. —RuakhTALK 15:41, 30 January 2009 (UTC)
Then we should change WT:CFI, which does not read that way. If we want to continue to increase the quantity of English entries there is literally no end to what we can achieve by exploiting the combinatorial explosion of compound and multiword entries. As more and more corpera are deemed durably archived attestation can be less and less of a barrier. I would be happier if we worked on entry quality, especially for highly polysemic words, and on Proper nouns than pushing in the direction of non-idiomatic collocations. DCDuringTALK 16:04, 30 January 2009 (UTC)
To put it another way, if I might: the editors who created these entries are familiar with CFI, and from what I have seen I am reasonably sure they would not deliberately flout Wiktionary's standards. Thus, I am inclined to assume that they could/would/did compose a reasonable argument in favor of the idiomaticity of these phrases. It is not too hard to see how such an argument would run; I would say the argument is relatively strong for by accident and somewhat weaker for the others. But where there are grounds for reasonable dispute over whether something is or is not idiomatic, it only makes sense to err on the side of inclusion. IMO this isn't a matter of CFI itself -- no one is questioning that phrases that are indisputably sum of parts should be deleted -- but of our implementation of CFI. Implementation naturally varies depending on practical concerns; specifically, I think the actual threat of the project being flooded with sum-of-parts phrases is quite small compared to the threat of our being overrun with protologisms or brand-name spam. -- Visviva 16:38, 30 January 2009 (UTC)
I would vastly prefer a "rule of law", largely because it allows us to be more welcoming of contributions by non-experts. That means to me statute, not precedent (especially not hard-to-find precedent). Still less does it mean rule by judges.
I, too, believe that some prepositional phrases are close to idiomatic (eg, by accident, though I'm not sure why). I am not at all convinced that the de facto rules that seem to be emerging are desirable. In particular, I am troubled by the rule that a phrase combining two low-frequency (in the unsupported opinion of one or three senior contributors) senses of common words makes an includable collocation. I am bothered that it seems to lead to the failure to include senses at the individual word level (eg, sail and steam). The problem is particularly pernicious for prepositions, where missing senses may be used in many phrases.
I would be very surprised if adding these "multiwords" and "phrases" helped us in the slightest in our competition with the other on-line dictionaries. DCDuringTALK 18:53, 30 January 2009 (UTC)
I agree that it doesn't have much effect on our competitiveness, but then not all contributors are motivated by competition with other dictionaries, and their contributions are no less valuable for that. Personally, I'm not very concerned by other dictionaries' competition, although I do look forward to the day when we flatten them... as we will, unless something flattens us first.
Back to the point at hand, I'm not sure if you're objecting to 1) the assumption that the creators of these articles would make plausible arguments in their defense, 2) the principle that if a plausible argument can be made for non-compositionality, the entry should be kept, or 3) both. I can understand the objection to 1) -- it is a bit ad hominem -- but I only meant it as a time-saving device. I mean, why wait for an argument to actually be made, when it can so easily be foreseen? ;-) My intent with 2) is actually to reduce the amount of arbitrariness in these decisions. When an RFD is decided solely on the (often rather erratic) judgment of the (often very small sample of) editors who weigh in, many entries end up being deleted/kept that might have had the reverse outcome if the discussion had happened a month earlier or later. This is untenable, and ends up discouraging useful contributions and contributors. So IMO this sort of "reasonable doubt" principle actually reduces the weight of our "judges" (if you will) vis-a-vis our "law."
At any rate, I didn't mean to drive this particular discussion so far off-course. Perhaps we could discuss this further on the BP. -- Visviva 15:26, 31 January 2009 (UTC)
I think the rub with these terms comes in the bent or method through which you want to define them. When certain phrases have an idiomatic meaning, do you want a separate noun sense at the main word (sail ...5. A sailing vessel or vessels. We traveled by sail.) or take it as an idiomatic phrase by defining 'by sail'... [or both... the breadcrumb theory?]. For accident (below), you would at least have to add at least one more idiomatic sense to the main word (accident ...5. Chance; fortune; lack of intention. I ran into an old friend by accident.) In cases where the idiomatic sense exists only in the phrase, it should have an entry, which as far as I can tell is true for that specific sense of accident, and this sense of sail. Can anyone think of any other uses where the specific idiomatic sense of these main words exist in other phrases, or alone? It would be a good way of understanding the way each of these break down on a sense level, and if any of them have uses outside the phrase. -- Thisis0 18:56, 30 January 2009 (UTC)
I don't think sail means sail-boat so much as sail-power. A person can go by sail, which means he's on a sail-boat, but a boat can also go by sail, which means it is a sail-boat. In days of yore, a single boat might travel sometimes by sail, and sometimes by steam, depending on circumstances. I don't think that's very common any more, but this b.g.c. hit for "using sail" suggests both (1) that it does still happen and (2) that the exact words "by sail" are not necessary to evoke this sense. And that's ignoring variations like "by solar sail" and "by steam or sail" which have the "by" and the "sail" but don't put them together. (All of which are arguments for deletion; my "keep" vote is based entirely on the Visviva Principle.) —RuakhTALK 20:09, 30 January 2009 (UTC)
Idiomatic phrases are not necessarily fixed. A pin in an enormous stack of hay; rain felines, canines, and pachyderms; have your retirement cake and eat it too. DAVilla 06:59, 3 February 2009 (UTC)
I suppose I am a breadcrumbian, with a further belief in "all senses for all words". My particular sect would favor three kinds of breadcrumbs, but view advocacy of the fourth as heresy. The favored ones are:
senses at each component word for every sense of the word used in any idiom (indeed any collocation). This would include senses for the current existence of which we had no evidence except in phrases that are currently idiomatic;
usage examples, which provide a natural means to get a user searching for a common collocation to an appropriate entry.
My disfavored fourth class of breadcrumb, non-idiomatic collocations, is the one that seems wasteful. It can be positively harmful if its presence leads to neglect of the others. DCDuringTALK 16:53, 31 January 2009 (UTC)
Keep. Well I can't of course speak for all non-natives, but I myself see very few things more unuseful, nay, counterproductive (because confusing) for a non-native than lists of senses which appear in only one or two set phrases under separate definitions for the word, rather than under their own headings (cross-referenced with the word's heading of course). Moreover, it increases the danger of the warning "only in this phrase!" being omitted, thus leaving the user under the delusion that the definition is applicable generally (adding an example sentence doesn't solve this at all); and with respect to prepositional phrases (probably some others too) it's bound to make even the Translations tables much less user friendly, as many easily translatable phrases would be turned into intranslatable definitions. --Duncan 18:42, 31 January 2009 (UTC)
Keep (with minor reservations) I disagree. I believe that entries that are eligible for Category:English prepositional phrases are extremely useful and not at all confusing. Many (not to say all) Eng L2 learners are encouraged to investigate and learn these idiomatic preposition + noun (phrase) collocations. BTW, I hope this Category answers your above question, Thisis0. Nearly all prepositional phrases are adverbial or adjectival and somewhat idiomatic, such as by accident below, (which is why DCDuring is having trouble seeing what makes it different). Phrases such as by sail, rail, road, air, sea, etc I believe should have an entry, as their meaning is idiomatic despite being easily understandable. Whereas phrases such as by car, train, boat, bike, plane, etc. should be catered for under by as these are non-idiomatic SoP's. -- ALGRIFtalk 16:34, 1 February 2009 (UTC)
Delete. This is standard use of by (by car, by horse, by boat, by ferry, by jet) with a metonymic use of sail to stand for the whole boat. --EncycloPetey 16:47, 1 February 2009 (UTC)
Precisely, it's always metonymic, not literal. Keep.DAVilla 06:59, 3 February 2009 (UTC)
So which is literal: "by road" or "by car"? "by train" or "by rail"? "by ship" or "by sea"? Travel statements using "by" use metonymic nouns as a matter of course. --EncycloPetey 16:39, 1 March 2009 (UTC)
What's metonymic about "by car", "by train" and "by ship"? --Duncan 22:53, 14 March 2009 (UTC)
Kept per consensus. --Jackofclubs 00:15, 19 June 2009 (UTC)