Wiktionary talk:Votes/pl-2010-03/Change to first lines of CFI

Idiomatic

• The problem now becomes what you mean by "idiomatic". Since it really means "characteristic of a given language", almost anything can be considered idiomatic, and as a matter of fact it is usually only meaningful when applied relative to something else (ie another language). An example: "he walks with a stick" is idiomatic English in the sense that it is a perfectly natural and fluent way to express the idea. (Unlike, say, French, where you have to say "he helps himself with a stick".) That doesn't (necessarily) mean we should have an entry for walk with a stick. That particular example happens to be difficult to translate, but many phrases which are idiomatic English are also uncomplicatedly sum-of-parts and not dictionary material.
• To sum up: "idiomatic" is not the opposite of "sum-of-parts", despite what some editors apparently believe. Ƿidsiþ 14:20, 25 March 2010 (UTC)
  • I agree apart from the now bit; why wasn't that a problem before? Mglovesfun (talk) 14:24, 25 March 2010 (UTC)
    • Note the vote hasn't started yet, so you can still change the text. Mglovesfun (talk) 14:25, 25 March 2010 (UTC)
      • • Well, it was. But I assumed this wording change was supposed to resolve the problem somehow (I was trying to work out what the difference between the two versions was, and the only substantive one seemed to be adding idiomatic to the top line.) Ƿidsiþ 14:26, 25 March 2010 (UTC)

See User talk:Mglovesfun/CFI/1. IMO (but you can disagree) idiomatic in terms of Wiktionary policy should be defined in CFI, but maybe not in these first lines. In fact, I think it already is, just not all that well. Mglovesfun (talk) 14:34, 25 March 2010 (UTC)

A good solution would be to change attested and idiomatic to attested and can be considered as a term, as a building block, of the language. Lmaltier 06:28, 26 March 2010 (UTC)

A language dictionary is not only for definitions

The sentence A term should be included if it's likely that someone would run across it and want to know what it means. assumes that a dictionary is used for definitions and only for definitions. This assumption has caused many unnecessary discussions in RfD. I would change it to A term should be included if it's likely that someone would run across it and want to know what it means, or want some linguistic information about it (such as pronunciation, etymology, etc.).. Lmaltier 06:28, 26 March 2010 (UTC)

Compounds

I have removed "compounds" from the list of singled out cases of "term", as compounds such as "headache" are single words in the usual sense of "word". Sorry for the late editing. --Dan Polansky 11:56, 29 March 2010 (UTC)

Wanting to know what it means

I am starting a discussion thread for the proposed deletion of the following text:

A term should be included if it's likely that someone would run across it and want to know what it means. This in turn leads to the somewhat more formal guideline of...

I support removal of this sentence. The real criteria that are applied in the RFV and RFD processes are the attestation and idiomacity. To introduce the core policy with the words "This in turn leads to the somewhat more formal guideline of..." seems absurd to me, and confusing to newbies. The section "General rules" should get to the point as soon as possible, instead of starting with preamble-like proclamations that do not serve to regulate inclusion and explusion of terms and senses. --Dan Polansky 08:12, 30 March 2010 (UTC)

I don't think "This in turn leads to the somewhat more formal guideline of" introduces the core policy. The core policy is that a dictionary serves to provide information about words and phrases to people who want that information, and I think that is nicely summed up in the first "A term should be included if it's likely that someone would run across it and want to know what it means". The stuff after that should just explain what sorts of things people are likely to run across and want to know about. IMO, at least. What's your core policy? L☺g☺maniac ☃ 14:07, 30 March 2010 (UTC)

What I have referred to as "core policy" is something like "include nonSoP attestable terms". That is the policy that is actually used to decide the inclusion of most entries; there are exceptions for proper names. This is the rule we mostly abide by. It is this rule that is the client for the sections that define "term", "attestable", and "nonSoP" (AKA "Wiktionary:idiomatic").

"A term should be included if it's likely that someone would run across it and want to know what it means" is something I have rarely seen invoked in RFD and RFV discussions, if ever at all. The way the sentence reads, to get a term included I only need to prove that at least one person is likely to want to lookup the term. But if a person comes to Wiktionary and enters a term, he can then attest to being himself the person who wants to lookup the term and know what it means. He will say: "I did not know what the term meant, so I did some research, and stored the result to Wiktionary." It will be not only likely, it will be certain that at least one person wants to lookup the term. So the formula seems wholly unfit as a policy item, as far as I can see and argue. --Dan Polansky 16:45, 30 March 2010 (UTC)

It should be invoked in RFD and RFV discussions more often.

Is there a problem with having a term because one person would likely want to look it up? Not for obviously sum-of-parts things like "red blanket" or "yellow wallpaper", of course, but other things. I guess it should be "if it's likely that people would run across it and want to know what it means" or something like that. I'm just really uncomfortable with deleting that sentence because it seems that is the only sentence in CFI that reflects our attempt at being focused on serving users, which really should be our main focus.

I still don't see exactly why you are uncomfortable with that sentence. Can you elaborate further? L☺g☺maniac ☃ 01:37, 31 March 2010 (UTC)

CFI is not a mission statement; it is a set of practical regulations to be applied in the RFV and RFD processes. The discussed sentence could be in the very first paragraph of CFI, that first paragraph that is a preamble of sorts anyway, featuring the slogan "include all words in all languages". But in the section "General rule" or "General rules", there should only be actually applied rules. If the community actually agrees that the discussed sentence is superordinate to "include nonSOP attested terms", let us make that very clear. Because that would probably be an attack on the nonSOP requirement. We have repeatedly deleted items that people were likely to lookup merely on the ground that they were sum-of-parts. The attestability requirements seems to be less endangered, but even that requirement would be questioned from time to time, when people would want to look up terms that are found in hundreds of occurrences in non-durably archived sources.

To me as a person who has taken part on a host RFDs and RFVs, the presence of the sentence in CFI seems hypocritical. The part "This in turn leads to the somewhat more formal guideline of..." is a joke; what follows this part is not a "somewhat more formal guideline" but rather a formally and rather strictly enforced policy. --Dan Polansky 07:31, 31 March 2010 (UTC)

Alright; fine; I defer to the will of the active community and the judgment of more experienced editors. L☺g☺maniac ☃ 14:44, 31 March 2010 (UTC)

Another proposal

I have come up with a proposal that seems to meet the objections raised in the vote:

Broadly, a term should be included if it is likely that someone would run across it and want to know what it means. This broad requirement leads to the following two practically decidable specific requirements, both of which have to be satisfied for a term to be included.

1. Attestation: The term has to be attested, typically by quotations from permanently recorded media, conveying meaning, in at least three independent instances spanning at least a year.
2. Idiomacity: If the term consists of more than one word, it has to be idiomatic, meaning that the meaning of the term cannot be easily deduced from the meanings of its constituent words.

The two requirements are detailed in the following sections.

The proposal makes a compromise with the editors who would miss the broad requirement. It avoids nested conditionals by breaking the requirements into separate sentences. --Dan Polansky 22:39, 2 April 2010 (UTC)

That's better, but I don't like "practically decidable", I think you mean "in a practical way" rather than "virtually". Nor do we include all words in all languages. We include less than that, as we often exclude "specific entries" like towns, fictional characters etc. We also include more, as we should include all meanings of those words, not just the words themselves. Nobody would support to have only having one definition. Mglovesfun (talk) 00:04, 10 April 2010 (UTC)

Thank you for the feedback. According to one of its points, two contrasting terms used in the formulation need improvement: "broad requirement" and "practically decidable specific requirement". Alternatives that come to mind:

• (a) "broad requirement" and "specific requirement" (dropping the "practically decidable" part)
• (b) "generic requirement" and "specific requirement"
• (c) "vague requirement" and "sharp requirement"
• (d) "vague requirement" and "definite requirement".

I am looking forward to hearing specific proposals for the change in the wording. I am a non-native speaker, and it shows. --Dan Polansky 11:25, 18 April 2010 (UTC)