Property talk:P407

From Wikidata
Jump to navigation Jump to search


language of work or name
language associated with this creative work (such as books, shows, songs, or websites) or a name (for persons use P103 and P1412)
Descriptionlanguage associated with this item (for persons use native language (P103) and languages spoken, written or signed (P1412))
Representslanguage (Q34770)
Data typeItem
Domainwork (Q386724) (note: this should be moved to the property statements)
Allowed valuesany languoid (Q17376908), mainly language (Q34770) (note: this should be moved to the property statements)
ExampleAutobiografia di Alice Toklas (translation) (Q15741500)Italian (Q652)
Tracking: sameno label (Q42533405)
Tracking: usageCategory:Pages using Wikidata property P407 (Q20989967)
See alsooriginal language of film or TV show (P364)
Proposal discussionProposal discussion
Current uses7,566,518
Search for values
[create] Create a translatable help page (preferably in English) for this property to be included here
Value type “languoid (Q17376908): This property should use items as value that contain property “instance of (P31)”. On these, the value for instance of (P31) should be an item that uses subclass of (P279) with value languoid (Q17376908) (or a subclass thereof). (Help)
Exceptions are possible as rare values may exist. Known exceptions: multiple languages (Q20923490)
List of this constraint violations: Database reports/Constraint violations/P407#Value type Q17376908, SPARQL, SPARQL (new)
Conflicts with “instance of (P31): Wikimedia disambiguation page (Q4167410), human (Q5): this property must not be used with the listed properties and values. (Help)
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P407#Conflicts with P31, SPARQL, SPARQL (new)
Qualifiers “writing system (P282), start time (P580), end time (P582), translator (P655), platform (P400), applies to part (P518): this property should be used only with the listed qualifiers. (Help)
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P407#Allowed qualifiers, SPARQL, SPARQL (new)
Pictogram voting comment.svg conflict with company
instances of company (Q783794) should not use this property. Generally, companies don't have a language. (Help)
Violations query: SELECT DISTINCT ?item WHERE { ?item wdt:P407 []; wdt:P31/wdt:P279* wd:Q783794 }
List of this constraint violations: Database reports/Complex constraint violations/P407#conflict with company
Pictogram voting comment.svg conflict with physico-geographical object
instances of physico-geographical object (Q20719696) should not use this property. Generally, physico-geographical objects don't have a language. (Help)
Violations query: SELECT DISTINCT ?item WHERE { ?item wdt:P407 []; wdt:P31/wdt:P279* wd:Q20719696}
List of this constraint violations: Database reports/Complex constraint violations/P407#conflict with physico-geographical object
Pictogram voting comment.svg conflict with administrative territorial entity
instances of administrative territorial entity (Q56061) should not use this property. Generally, administrative territorial entities don't have a language. (Help)
Violations query: SELECT DISTINCT ?item WHERE { ?item wdt:P407 []; wdt:P31/wdt:P279* wd:Q56061}
List of this constraint violations: Database reports/Complex constraint violations/P407#conflict with administrative territorial entity
This property is being used by:

Please notify projects that use this property before big changes (renaming, deletion, merge with another property, etc.)

On different use scenarios[edit]

Hi, I've been sticking this property to a lot of video game pages, and now I started pondering my use of it, and various use scenarios. So, the original use case was for a specific translated edition of a book – say, the Septuagint would have language as Greek and original language as Hebrew and Aramaic, am I correct? That's a pretty limited use, considering each work should have its language as "original language". Now, the way I saw it and have been using it for games (granted, without considering it this thoroughly) is, I've put "original language" as, well, the original language, and "language" as whatever languages the game was released in. For example, Pokémon Gold and Silver has original language as Japanese and language as "Japanese, English, German, Spanish, French, Italian, Korean", as these are the languages the games were released in and are available. Is that appropriate use of this property? Even if it isn't, I hold that the different language releases are information that should/could be covered. For games it's straightforward, as they're usually just released in a set of languages (often on the same copy!) and not translated over time, as publishers become interested, into a huge load of languages (like books). Opinions? Enoirditalk 08:34, 11 May 2013 (UTC)

Hi, I'm not sure if I've understood you correct. "Original language" P:P364 should be used only for the original language (normally should be one), thats right (that just for clarifying). But "language" P:P407 should only contain the translated languages. So "Japanese" in your example would be not correct, because it's now stated twice. But I'm not really sure if that was really your question..? Greetings, --#Reaper (talk) 12:57, 12 May 2013 (UTC)
Are you sure this property is meant to list translations? What I meant is that I'm myself not sure of how this property should be used – should it list translations like you suggested or be used only where the item is about a translation, not an original work (e.g. translation Septuagint; language: Greek; original language: Hebrew, Aramaic). I'm calling for this property to be defined; for there to be a guideline for use to follow. Also, I added Japanese to both because I thought of a situation where the language information would be used on a infobox on Wikipedia: of course it should list also Japanese as one of the languages the game is available in. Enoirditalk 14:08, 12 May 2013 (UTC)
Um, no one else seems to be interested in this topic..? I see no problem to "reuse" this property for this case, I think this property should be a generic one. Until there are no votes against this way of usage you could do so. If necessary it could still be changed by a bot. Your proposal (on the bottom) is an interesting one, but maybe the wrong place, we should try it there: WD:PP (there are also more people than here). Because of the problem with the infoboxes: It's possible to concatenate (add) both values (original language and languages) so there is no need to add Japanese as language as well. Best regards, --#Reaper (talk) 21:44, 14 May 2013 (UTC)
PS: I've completely forgotten that there is already a proposal: Wikidata:Property proposal/Creative work#translated into --#Reaper (talk) 20:09, 15 May 2013 (UTC)
(This proposal was discarded, see Property_proposal/Archive/11#Translated into. - LaddΩ chat ;) 12:28, 7 March 2014 (UTC)

Alternative arrangement[edit]

  • Property original language for the language of a work. Even if the work is specifically a translation, this would list the language of the translation, as it's the "original language" of that specific edition.
  • Property translated from for items about a translation.
  • Property translated into for items about a work that's not a translation.

A possibility, unless I misunderstood what I'm talking about above. Enoirditalk 14:08, 12 May 2013 (UTC)

Remove Single value constraint?[edit]

I´m not sure if this contraint makes any sense. There are works that are bilingual or multilingual. So the single value constraint does not apply as a general rule. I´d like to remove it.--Giftzwerg 88 (talk) 08:50, 26 March 2015 (UTC)

Removed.--Giftzwerg 88 (talk) 14:48, 1 April 2015 (UTC)
Il existe des œuvres (journaux) écrits en plusieurs langues sans que l'on puisse énumérer les langues. Il faudrait pouvoir mettre dans ce champ une valeur «multilingue». There is scientific revues wrote in too many language. A value like «multilingual» should be authorized . HB (talk) 15:48, 28 August 2015 (UTC)

language of TV/radio broadcast[edit]

[1] @Intgr:, TV/rađio programs/chanels broadcast uses particular language(s) and also called creative works.--Avatar6 (talk) 18:13, 18 April 2016 (UTC)

@Avatar6: Well I reverted partly because the alias had incorrect grammar. I'm not sure I would call channels works, but I agree that out of all the language properties, it's the most appropriate one, so probably not worth debating. Intgr (talk) 12:05, 19 April 2016 (UTC)

Label is confusing[edit]

I just found out that the original language of work that I had input on a large number of publications was changed by a bot to this new property; however, the name of this new property is super clunky. There has to be a better description that isn't so convoluted. It's non-ideal this. -- BrillLyle (talk) 20:36, 21 November 2017 (UTC)

I tried to simplify the description. If it comes to the label, rewriting it to "language" would be in my opinion the best. --Pasleim (talk) 08:38, 24 November 2017 (UTC)

The label is still confusing. In English it is "language of work or name", but this property is being used on both works and editions, as well as other items that are neither works nor names, and where the distinction between what is and is not a work is highly important. --EncycloPetey (talk) 14:54, 5 December 2017 (UTC)


On Wikidata:Properties for deletion#P2439 I have proposed to change the label of this property to "language" but probably that discussion didn't get enough attention, so I start another survey about the label change.
I think the label should be changed because this property (P407) does not only indicate the language of works or names but generally the language of the described entity. For example:

  • on an item about a work, P407 indicates the language of the work
  • on an item about a name, P407 indicates the language of the name
  • on an item about a term, P407 indicates the language of the term
  • on an item about a webpage, P407 indicates the language of the webpage
  • on an item about an edition of a work, P407 indicates the language of the edition
  • on an item about a translation of a work, P407 indicates the language of the translation.

In conclusion, the property should be renamed to language of work, name, term, webpage, edition or translation or simpler language.
Pinging user of the previous discussion on WD:PFD: @Jura1, Matěj Suchánek, Liuxinyu970226, Snipre, ChristianKl, Fralambert: @Arbnos, Hsarrazin, Thierry Caro, Deryck Chan, Pigsonthewing: --Pasleim (talk) 07:25, 25 May 2018 (UTC)

Just language or maybe language used. Matěj Suchánek (talk) 07:48, 25 May 2018 (UTC)
language used is language used (P2936). P407 is not about the language used by an object but about the language the object has. --Pasleim (talk) 08:05, 25 May 2018 (UTC)
It's getting more simple, then :) One more suggestion: in language (but still, I prefer the shortest one). Matěj Suchánek (talk) 08:11, 25 May 2018 (UTC)
We can use "language of item subject". Does it help? Deryck Chan (talk) 09:48, 25 May 2018 (UTC)
This sound quite technical. What about "language of subject"? --Pasleim (talk) 11:10, 25 May 2018 (UTC)
I would agree to "language", with a strict condition "not to be used on Q5 or places items", or there will be a lot of misuses ^^ - as long as previous labels are kept as alias, it is alright for me. -> and merge P2439, or the label will be misleading ^^ --Hsarrazin (talk) 13:33, 25 May 2018 (UTC)
The current wording is fine. It does not have to be exhaustive, but it does need to distinguish this property from other "language" properties, such as languages spoken, written or signed (P1412) and native language (P103) and language used (P2936) and official language (P37). Reducing the label to just "language" creates too many possibilities for confusion. --EncycloPetey (talk) 15:32, 25 May 2018 (UTC)
I'm confused by your statement. In the section above you wrote: The label is still confusing. What changed your mind? --Pasleim (talk) 21:10, 25 May 2018 (UTC)
Nothing changed my mind. The label can be confusing because of the precise meaning used on WD for "work", and this can be coupled with the (erroneous) assumption that if a property says "work" it can't be used for an "edition". But changing the label to "language" merely makes the problem worse, so given the two options, I'd prefer to keep it as it is. --EncycloPetey (talk) 02:42, 26 May 2018 (UTC)
What about the other ideas: "in language", "language of item subject", "language of subject"? --Pasleim (talk) 13:18, 27 May 2018 (UTC)
”language of subject” could be read ss “language spoken by subject”, so I don’t think we should use that. I have no problem with the current label. - PKM (talk) 20:19, 27 May 2018 (UTC)
  • Pictogram voting comment.svg Comment Simply having "language" is too ambiguous. Having it longer may give more context, however, in all similar situations they have been added as aliases and to me that suffices. There are many other such examples where the final label has variations of contexts, eg. subject has role (P2868)
@PKM, EncycloPetey, Pasleim, Matěj Suchánek, Hsarrazin, Deryck Chan:, @Kcoyle, VIGNERON, billinghurst: And what's about "expressed in (language)" ? Snipre (talk) 13:11, 3 December 2018 (UTC)
@Snipre: I'm not convinced. "Language" is a concise label that came out of the long discussion both here and WD:PFD, which isn't perfect but works as a keyword for people to find this property when they type it into the property input box. I don't think any of the proposed labels can avoid both ambiguity and unnecessary constraint. Deryck Chan (talk) 11:28, 5 December 2018 (UTC)
@Deryck Chan: ""Language" is a concise label that came out of the long discussion both here and WD:PFD". The consensus was so large that since 6 months, no change was made in the label item.
More seriously, we have several properties dealing with languages and having a too generic one creates a large risk of misuse. Currently we only have a list of propositions, perhaps the next step it is to start a vote. RfC or a simple section on this talk page is enough ? Snipre (talk) 16:41, 5 December 2018 (UTC)
I think we should have a generic property for any work of human creativity that is expressed in some particular language(s) (P407) and more specialist properties for the particular relationships represented by original language of film or TV show (P364) and languages spoken, written or signed (P1412). The ambiguity you're describing is better solved by writing a good property description and usage notes, rather than trying to squeeze instructions into the property label. Deryck Chan (talk) 17:10, 5 December 2018 (UTC)
I often caution against using a language designation on text for anything other than actual written or spoken prose or poetry. It can be very difficult to attribute a language to short strings, such as a title of a work (cf the "1984", albeit translated into dozens of languages, or a book entitled "Marie Antoinette"), to a person's name (many examples here, including my own "Karen Coyle" which could be said to be Swedish/IRish, but also "Mario Wong" (actual person)), or to company names (Oakland California restaurant "Pasta Cuisine"). Yet, taking a look at Q172850 for Eco's "Nome della Rosa" and in it the editor is encouraged to add labels for the page title "in more languages", which happens to be the title of the book, which in some cases might not represent actual translations, but is given a caution when those same strings are given as titles (because there should only be one title on a work). This has got to be confusing for editors! Title title (P1476) is described both as "title of original" and "title of the work" without resolving if those are the same (they aren't in library data, although I could be convinced that they should be - however, if they are the same there should be only one way to describe it, to avoid confusion). I'd say that there's a bigger problem than language here, and I haven't had a chance to dig into it deeply. Kcoyle (talk) 21:09, 5 December 2018 (UTC)
@Kcoyle: The problem you mentioned is solved if people use the FRBR system as described in WikiProject Books for books: only relevant data should be added to an item and duplicates have to be avoided. A tanslated edition should not provide any information about the original edition or about the work, but only provide a link to the work item and users have to extract relevant data using items relation. A work has to have one item, the original version/edition has to have one item and the translated edition has to have one item. Snipre (talk) 12:53, 6 December 2018 (UTC)
@Snipre: This is a hugely complex topic and I don't know that I can do it justice in a comment, but FRBR was designed as a relational database design, and possibly works well in that context (we don't really know for sure). For it to "work" in Wikidata and its graph structure depends quite a bit on how WD is used. The idea behind FRBR was that none of the Group 1 entities would actually "stand alone" but would need to be used as RDBMS tables are - as parts of a whole within a database and its table-aware applications. I'm OK with moving to work/edition in wikidata but it only makes sense if input and output provide something useful for general users of the data. In other words, input/output would need to be cognizant of the dependencies between works and editions, and users of WD would somehow have to know that retrieving properties on an edition will not be complete without also retrieving the properties associated with a work. As an example, there would be no subjects/topics associated with an edition; those would only be associated with a work. (I'm not sure what you mean above about "an item" but I'm guessing that you are referring to WD entries as "items".)
I guess I should mention here, and I will add to my user page, that I wrote an entire book on FRBR (mostly about what's wrong with it, but quite a bit about how it came about historically), and that book is available as open access on my site and has a minimal entry FRBR, Before and After: a Look at Our Bibliographic Models (Q56487853) on WD. My main message is that while some of the concepts introduced by FRBR are interesting, it should not be assumed to be a viable data design for a specific application, and definitely needs rethinking if used outside of the RDBMS context. Kcoyle (talk) 22:21, 6 December 2018 (UTC)
@Kcoyle: I didn't have the time to read your book, but instead to saying FRBR system we should the application of FRBR on WD. We didn't have the time to fully integrate the FRBR concept into our data model, but we select the distinction between work and expression, in the case of books this leads to the concept of written work (Q47461344) and version, edition, or translation (Q3331189). We don't need to analyze complex cases to define that by separating both concepts we have to separate data according to concept. So original language in the expression concept is duplicate information if we have the language defined in the work concept and if a link is created between the expression item and the work item. We can't justify the use of original language without creating the other properties "original publication date", "original publisher", "original publication place",...
Yes, item = Wikidata entry.
The data model for books is described in Wikidata:WikiProject Books is you want more info about input/output. I am not an expert, we are developing data model from case studies and the model is expanded following discussions on the project talk's page. The description of that system should be better presented, I know, and a summary is under construction. Snipre (talk) 10:17, 10 December 2018 (UTC)
@Snipre: I'll just say here that the aspect of FRBR that is difficult to implement is that the main bibliographic entities (Work, Expression, etc.) are not subclasses of each other, and therefore properties on the "broader" ones are not inherited by the more specific ones. Thus a WD:Edition instance does not obtain (by inferencing) any properties assigned to a WD:Work instance, even if they are linked. The two are separate "things". Because of this there really is no duplication of data between them any more than you would have duplication between, for example, a description of a man and a description of a woman who are connected between "isMarriedTo". If both are said to have brown hair, that is not a duplication. Each entity needs to get the description that meets user needs for that entity. It may be the "original language" isn't needed for editions, but if the information is needed it should be immediately available to the edition instance, because it cannot be inferred from the work instance, even though a specific program could make that link. But we aren't designing for one specific application, so we can't assume that the link will be made. Kcoyle (talk) 19:01, 25 December 2018 (UTC)

other languages[edit]

Independent whether the English label gets changed, we should also come up with a solution for all other languages. Currently, around 40% of all languages don't follow the English label. Instead, the following translated terms are used:

language: be, bs, ca, ce, fi, gl, gu, hy, la, lmo, ms, nds-nl, oc, or, scn, sk, sl, sr-ec, sr-el, te
language of work: be-tarask, da, el, eo, fa, fy, gsw, he, ja, nb, nn, pt-br, sv, tg, yi
language of work or name: ar, ast, bg, bn, cs, cy, es, et, eu, gd, ia, it, ka, ko, lb, lt, mk, ml, ms, nds, nl, pa, pl, ps, pt, ro, ru, sco, sq, sr, ta, th, tl, tr, uk, ur, uz, vi, zh(-hans/-hant/-cn/-hk/-mo/-my/-sg/-tw)
language of work, name or term: de, fr

Should the English label dictate the labels for all other languages, or should each language community be free to choose their most appealing label? --Pasleim (talk) 23:54, 29 May 2018 (UTC)

@Pasleim: I confirm that all zh-* families, Arabic, Bulgarian, Czech, Estonian, Spanish, among others, are translated under third case. Also confirm Japanese translation (作品の言語) matches second. --Liuxinyu970226 (talk) 11:38, 6 June 2018 (UTC)
Anyway, Punjabi "ਕੰਮ ਜਾਂ ਨਾਂਮ ਦੀ ਭਾਸ਼ਾ" is actually about third case "language of works and/or names". --Liuxinyu970226 (talk) 12:02, 6 June 2018 (UTC)
Some others that I'm confused with:
  1. The Bengali translation "রচনার বা নামের ভাষা" seems like about composition, @Mahir256: thoughts?
  2. Central Kurdish "زمان", is this about "time"? Kurmanji Kurdish "zimanê berhemê an jî yê sernav": language of product and/or title?
  3. Hungarian "alkotás vagy megnevezés nyelve": Yandex translate told me "language of creation or denomination", not sure if this matches third or not.
  4. Indonesian "menggunakan bahasa": Using language?
  5. Latvian "darba vai nosaukuma valoda": language of job (or title)?
  6. There's another Tajik translation under tg-cyrl "забон ё номи асар" which by using Yandex translate, it probably means "language or worksheet".
  7. Urdu "تخلیقی زبان": Creative language...
  8. Yoruba "home country or name": Home countries and names???

--Liuxinyu970226 (talk) 12:08, 6 June 2018 (UTC)

@Liuxinyu970226: I rendered "language of work or name" into Bengali and changed "language" to be an alias. @Edgars2007, Şêr, Calak, Csigabi, Ârash, Rachmat04:, who can help with explaining the meanings of the labels in the remaining languages in question. Mahir256 (talk) 13:22, 6 June 2018 (UTC)
Also @BukhariSaeed, Agbalagba: as the previous ping did not seem to work for them. Mahir256 (talk) 13:23, 6 June 2018 (UTC)
✓ DoneBukhari (Talk!) 13:36, 6 June 2018 (UTC)
@BukhariSaeed: What was the original label supposed to say? I think that's what @Liuxinyu970226: was asking. Mahir256 (talk) 13:51, 6 June 2018 (UTC)
The Hungarian translation is correct. Forget Yandex translate. Csigabi (talk) 19:04, 6 June 2018 (UTC)
Is correct to add "language of the reference" label for usage in a reference item for references which don't have an item of their own, as per suggestion --Ogoorcs (talk) 19:25, 16 July 2018 (UTC)

Allow “applies to part” as qualifier[edit]

Parts of the work or edition may be in different languages - for example, a Latin text may have glosses or notes in English. I use applies to part (P518) to indicate this - can we formally allow P518 as a qualifier here? - PKM (talk) 22:07, 30 December 2018 (UTC)