Several of my recent edits have been commented upon by @vermillon, which raised some questions about what language constitutes and whether the language can be readily inferred from the tracklist. According to the release style guidelines, the release language should be set according to the tracklist, not the lyrics of the songs in that release. Below are the two edits where we have discussed the issue:
Although I feel that @vermillon was needlessly assuming bad faith, I am not trying to seek support here. I just want to see what the community thinks about the style issue.
Iāll reply here and start by apologizing about assuming bad faith. The issue of languages in the Chinese language family is particularly difficult and politically connotated, so I perceive forcing a generic āChineseā when a specific language was provided as suspicious. Iāll try here to start on a fresh basis, sorry again.
Iāve re-read the guidelines, and I think they have some problems:
in the case where a title could be read as being several languages (we could take example of the many words that are shared among European languages, Iāve provided āfaceā as a random example of word that exists in many languages), the guideline fails to answer what to do. Silentbird suggests, and I think this is a good idea, to look for context (in the lyrics, in the other songs), in other words, to use common sense. In the example, āfaceā would be interpreted as English if it can be easily found that the song is in English, French if in French, etc.
with regards to Chinese, silentbirdās argument is that they canāt decide from the title alone what Chinese language the title is written in. Fair enough, but why is their proposed common sense strategy not employed there? Wouldnāt it make sense to go beyond the title and see in what language the song is sung, then? If not, how is it different from refusing to mark as English a song called āFaceā? And why would an album called ę ¹ be marked as āChineseā and not āI donāt know, could be Japanese tooā?
comes the question of what is the purpose of MusicBrainz, what service is it providing to its users? My take is that I want to be able to search for songs and albums in Min Nan Chinese. Thankfully, MusicBrainz lists such a language, and so people can label albums with it, when appropriate. Rejecting the label āMin Nan Chineseā on the basis that āoneā editor canāt decide effectively makes the database less useful for all the people who want to look for music in Min Nan Chinese, Cantonese etc. What gain is made? I only see a net loss.
finally, comes the question of how to treat other editorsā work; in any collaborative effort such as MusicBrainz, we are bound to have local domains of expertise, and nobody can know everything. When an editor with specific knowledge of something (e.g. āthis song āFaceā is in English, this can be looked up anywhereā) adds data (or in this case, even created the entire entry), what benefit is there to align the database on the lack of knowledge of another editor and make the database less useful for users? Claims to āaccuracyā donāt make much sense to me: first, because - in the current example - Min Nan Chinese is correct and more precise, but also - on general grounds - because dumping the entire database would obviously improve accuracy too (since it would then contain 0 error and be 100% accurate).
Perhaps it means that the guidelines (letās note that these are āguidesā, suggesting common sense should still be used) should be made more precise, in light of this kind of discussions. I donāt want it to be personal (although my initial reaction was bad, because work I genuinely consider to have been good has been damaged, in my perception), I would simply like to point that there is tremendous value in allowing people to specify which language is used in a song, even when the title alone could be construed as ambiguous.
Unrelated to the rest of this topic, but: for lyrics, we do have a specific ālyrics languageā field at the work/composition/song level where the language being sung can be indicated
Iām a little busy IRL at the moment so Iāll participate in the discussion later.
Yes, according to my understanding of languages and Chinese. If I have not misunderstood @vermillonās proposal, most Release Language:Chinese will be converted to Release Language:Mandarin Chinese, even for Cantonese (Yue Chinese) albums because the absolute most lyrics of Cantonese songs are written using Mandarin grammar but sung in Cantonese.
But most is not all.
I donāt mean much with my remark. You are fully qualified to find proper use of release Chinese language I am not.
I just noticed your most word and that left me little bit uncertain.
No, this is not my proposal, mainly because I do not believe in chasing entries and applying a cookie-cutter approach.
(I didnāt comment specifically on that bit because I donāt think itās particularly relevant, but saying a language ācan only be spokenā is of course a nonsense: any language can be written, and Mandarin can be written.)
Effectively, the reasoning should go the other way around: any entry āChineseā could be specified further (provided that itās in one of the Chinese languages that MusicBrainz support). I wouldnāt personally bother chasing them, and I think any effort of that sort is bound to annoy people and do more harm than good (e.g. the present case).
The example of Cantonese is particularly ill-suited, because indeed, a very large number of Cantonese songs are really written with a Mandarinized grammar and Cantonese pronunciation (to the point that reading the lyrics, it can be quite difficult/impossible to detect itās Cantonese), but this isnāt always the case in Cantonese, and is never the case in Hokkien (/Minnan), which tends to keep to its native syntax, vocab etc.
After some thought, I think my own personal take is:
If the written text (release title + track titles) could match multiple languages, then the Release Language should be set to the most generic option. This could be āChineseā in this case, but could also be empty/unset if itās e.g., a Danish/Norwegian/Swedish mixup.
If there are other qualifiers giving further detail, then itās fine to set it to the more specific language (but this shouldnāt be required, similar to āUse of special charactersā). Additional sources could be actually listening to any singing and determining the lyrics language, asking the artist, information in the booklet etc. (Burden of proof for the extra specificity lies on the editor marking it as such and should be given in the edit note setting it to this.)
Based on these two guidelines (that are in no way official and just something I made up just now), @silentbirdās edits should be cancelled. @reosarevok, do these seem sensible to you?
I have cancelled my edits so that discussions can happen here. Please tell me if I have missed any.
@Freso, can you tell me more about this? Whatās the situation there and why canāt they be distinguished from each other?
Itās relevant because written Mandarin is synonymous to standard written Chinese since the 20th century. So regardless of region, āstandard Chineseā is Mandarin (see Wikipedia).
Just for the record, my position is flexible. I just need to know what to do when adding releases and that the solution makes sense linguistically and in terms of the schema. I agree that Cantonese is indeed an oddball, but we still have the account for it. Tell me what you think of this: a Mandarin album originally released in Taiwan (where most people speaking Mandarin) is re-released in Hong Kong (where an absolute majority speaks Cantonese) and the tracklist printed on the album doesnāt change because both populations read traditional Chinese. Should I mark the Hong Kong re-release to have a Cantonese tracklist? Is there artist intent that the tracklist be read in Cantonese in Hong Kong but in Mandarin in Taiwan, despite the tracklists being identical?
If we decide that the artist intent is always that the tracklist follows the lyrics language, then what is the point of having release language? Itās duplicate data, because the lyrics language is already stored at the work level.
Generally the three languages (four written, BokmĆ„l and Nynorsk variants for Norwegian) are possible to be distinguished between, but there are also a lot of words that are the same (with same or different meanings) across the languages. The example was not much different than @vermillonās
we could take example of the many words that are shared among European languages, Iāve provided āfaceā as a random example of word that exists in many languages
Either way, further discussion about the similarities (or not) between the Scandinavian languages are off-topic for this topic, so letās move this to another topic if you want to discuss further.
I donāt see anyone saying this (ie., that is a strawman). A release with track titles clearly in English but lyrics in Russian should still have its language set to English.
However, in the case where the tracklist can be one (or more) of multiple languages based on how itās written, either set it to the most specific one possible (in this case āChineseā, in other cases could be unset/empty or maybe some other language). However, if listening to lyrics (or through some other investigative measure) it becomes clear that the language is a more specific one, then it should be fine to set it to that.
If itās a release of erhu instrumentals, there will obviously be no lyrics, and thus the release language will not be able to be determined by lyrics and youāll have either take the tracklist at face value (Chinese) or investigate more (e.g., ask the artist). On that same note, I wouldnāt take place of publication(/release) to mean anything for release language. If a Taiwanese artist makes a release in Mandarin Chinese by artist intent (but indistinguishable from Cantonese) and itās reāreleased in Hong Kong, Iād say the artist intent is still that the release is in Mandarin.
Thatās why I added āifā . So for releases where a Mandarin song is included with a Min Nan song, should I mark that release as [multiple languages]? This still bugs me conceptually, because if a person reads a Mandarin song title in Cantonese, itās not āwrongā. Itās just different sounds coming out of their mouth, but the meaning conveyed is identical.