Wikidata:Property proposal/stated in reference as

stated in reference as edit

Originally proposed at Wikidata:Property proposal/Generic

Done: object stated in reference as (P5997) (Talk and documentation)

Description	for use in a reference section, to indicate how the object of the statement was stated in reference, if this is different to how the object of statement would have been referred to in the statement's original context
Data type	String
Example 1	Systema vegetabilium, 14th edition (Q6136933)author (P50)Carl Linnaeus (Q1043) ref: BHL bibliography ID (P4327) 549 / stated in reference as: "Linné, Carl von, 1707-1778" / reference vocabulary: Library of Congress Name Authority File (Q18912790)
Example 2	CARTE DU PARAGUAY DU CHILI DU DETROIT DE MAGELLAN &c (Guillaume de L'Isle, 1703-1708) (Q56759863)depicts (P180)Strait of Magellan (Q48365) ref: British Library system number (P5199) 018640954 / stated in reference as: "Magellan, Strait of (Chile and Argentina)" / reference vocabulary: Library of Congress Subject Headings (Q1823134)
Example 3	[relevant view from BL 004920568 ] genre (P136) worm's-eye view (Q16966122) ref British Library system number (P5199) 004920568 / stated in reference as: "Worm's-eye views" / reference vocabulary: Library of Congress Genre/Form Terms (Q47537953)
Source	online catalogues and other external secondary sources
Planned use	I would use this in the current project I am working on, some 18th-century engravings and maps from the British Library; and also go back and change some uses of object named as (P1932) for authors of texts in the Biodiversity Heritage Library to use the new property.
See also	object named as (P1932), reference vocabulary

Motivation edit

I'm a big fan of object named as (P1932), to record exactly what was actually given in a source, when we have translated that text into a Q-number to use as the object of a statement: for one thing it seems a properly accurate reflection of the source, for another it makes things easier to unpick if there has been a mistake in the matching -- and may tip off someone reading the item if such a mistake has been made.

But recently working on items for some of the texts in the Bioheritage Diversity Library (BHL), I found myself becoming increasingly uneasy with statements like Systema vegetabilium, 14th edition (Q6136933)author (P50)Carl Linnaeus (Q1043) object named as (P1932) "Linné, Carl von, 1707-1778", because that is not how his name was recorded on the title-page of the text. If we click through to BHL 549 and then on to the title-page, we can see that how it states his name is "Caroli A Linné". The formulation "Linné, Carl von, 1707-1778" is a standardised indexing term used by the Library of Congress, and libraries that align their catalogues with it; but it seems odd to say that this was how his name was stated, in the context of a statement of authorship of this work.

A second, more practical difficulty is what happens when we want to additionally reference another source, that states his name in a different form -- eg a source that uses one of the other forms listed at VIAF; or, perhaps, the titlepage itself. Two different values of object named as (P1932) can't be sensibly combined on the same statement, so we end up with two statements that the author was Carl Linnaeus (Q1043) -- something humans may just about understand, but it's just the kind of thing to confuse bots and queries.

So I suggest that, where a name in a reference has been standardised to some controlled vocabulary, then instead of using qualifier object named as (P1932), we instead move that information to the reference, using this newly proposed sister property "stated in reference as", and if possible specify the vocabulary that the name has been standardised to as well (the suggested "reference vocabulary" property, proposed below). This also has the advantage that different references can specify different such forms, but still without confusion all be used on the same statement.

I would retain the existing object named as (P1932) if the name is given as it occurred in the original context of the statement, or there is no strong indication that it has been standardised.

One disadvantage is the value would not be as visible in a reference as it currently is in a qualifier, giving less clue to somebody casually coming to the Wikidata item, if there has been a mistake and it has been mismatched. But it still would be there and accessible for people actively investigating the statement. Jheald (talk) 21:41, 30 September 2018 (UTC)[reply]

Discussion edit

Proposed. Jheald (talk) 21:41, 30 September 2018 (UTC)[reply]
Neutral ~~Yes please.~~ I too keep fixing works with metadata from BHL with "stated as" that are basically nonsense and leave off the actual way the author is stated in the work. ("Andrea Murray" in the case of that particular edition of Systema Vegetabilium.) At least with "stated in reference", I don't have to delete the statement to replace it, I can just create a proper "stated as" on top of it! Circeus (talk) 23:20, 30 September 2018 (UTC)[reply]
- Switching to Neutral since I wasn't aware of subject named as (P1810), which seems to be pretty much exactly what this wants to be, except with a different, less clear name and slightly broader scope (i.e. also for identifier statements). How would this be different from that property? Circeus (talk) 22:33, 4 October 2018 (UTC)[reply]
Support David (talk) 08:07, 1 October 2018 (UTC)[reply]
Support Salgo60 (talk) 12:31, 2 October 2018 (UTC)[reply]
Support Jodi.a.schneider (talk) 04:12, 3 October 2018 (UTC)[reply]
Oppose Slightly, because you can use the "Alias" field at the top to add alternative names of the concept. --Zerabat (talk) 13:40, 3 October 2018 (UTC)[reply]

@Zerabat: Sure, but how does that help you know how the source recorded the statement? Jheald (talk) 14:24, 3 October 2018 (UTC)[reply]

How about Wikidata:Property proposal/subject stated as, predicate stated as ? --- Jura 18:19, 4 October 2018 (UTC)[reply]

@Jura1: "subject stated as" looks to be already covered by subject named as (P1810). "predicate stated as" suggests a property to indicate how the predicate (ie the property) was stated. Neither are about what this proposal relates to. This proposal is about how the object of the statement was stated in the source, normally currently given using object named as (P1932), but there are issues with that (as described above) if the way the statement-object is given in the referenced source is different from how it would have been stated in the original context. Jheald (talk) 19:03, 4 October 2018 (UTC)[reply]

@ديفيد عادل وهبة خليل 2, Jodi.a.schneider, Zerabat, Salgo60, Circeus, Jheald: @Jura1: Done: object stated in reference as (P5997). − Pintoch (talk) 19:40, 24 October 2018 (UTC)[reply]