Wikidata talk:WikiProject Taxonomy/Archive/2021/10

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Guidelines for adding references for taxonomic names

WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. After some discussion with @Totodu74: @Succu: @Ambrosia10: @Christian Ferrer: about how to improve referencing of taxonomic names on Wikidata (the discussion took place on my talk page), we'd like to propose the following set of guidelines for when we edit taxa. Most of this has been implicit in what a number of people have been doing, but we think it can be tidied up a little:

1. Where possible every taxon name (P225) should have at least one reference to a source for that name (ideally this should be the case for any Wikidata statement). These references could be external databases, but ideally would also include scientific publications (which themselves have been added to Wikidata).
2. If the reference is the publication that first described the taxon (e.g., the publication includes phrases such as "sp. nov.", "gen. nov.", etc.) then we can add the qualifier reference has role (P6184) with the value first valid description (Q1361864) to the reference. This should only be used for the first time a name has been published. We recommend that the property P5326 (P5326) not be used.
3. If taxon name (P225) is a "new combination", for example, a species moved from one genus to another resulting in a new name, then for the reference that moved the species (e.g., the publication will include phrases such as "comb.nov") we can use the qualifier reference has role (P6184) with the value recombination (Q14594740). We also encourage editors to connect the Wikidata item being edited to the item that has the original taxonomic name using either basionym (P566) (botanical names) or original combination (P1403) (zoological names). This way we can trace the name changes and help keep track of synonyms.
4. The taxon name (P225) being edited might be the result of a nomenclatural change, such as changing the spelling, it is a replacement name for a homonym (Q902085), etc. For these names we can use reference has role (P6184) with the value nomenclatural act (Q56027914). More precise terms can be used, e.g. replacement of junior homonym (Q108040629) if the reasons for the name change are known. If using replacement of junior homonym (Q108040629), we encourage the Wikidata item to be linked to the homonym being replaced using replaced synonym (for nom. nov.) (P694).
5. If the reference doesn't change the name of the taxon, but is otherwise taxonomically significant (e.g., it is a taxonomic revision, redescription, or treatment) then if the reference is a publication we could use one of revision (Q2146881), taxon redescription (Q42696902), or taxonomic treatment (Q32945461) as the value for reference has role (P6184) . If the reference is a itself a taxonomic treatment (Q32945461) (for example it corresponds to a Plazi (Q7203726) treatment) then there is no need to use the qualifier.

I'm hoping that these guidelines are clear, although looking at them there seems to be more text that I'd hope :( However, the key message is (a) please add references for taxonomic names, and (b) where possible to use reference has role (P6184) paired with one of the qualifiers discussed so that anyone looking at a taxon name (P225) can know where to go to understand the history of the name. Having some clarity and consensus on how to add references to taxon name (P225) will hopefully make it easier to add these references, whether it is single edits, or bulk uploads of data. Rdmpage (talk) 14:58, 11 August 2021 (UTC)

@Rdmpage: @Succu: @Ambrosia10: @Christian Ferrer:: I am currently replacing all the P5326 (P5326) statements I have made for various taxa (2'462 items) by stated in (P248) combined to reference has role (P6184): first valid description (Q1361864). I am using this query to extract all items concerned with P5326 (P5326), then the following code in QuickStatements:

QITEM	P225	"taxonLabel"	S248	QPUBLICATION	S6184	Q1361864
-QITEM	P5326	QPUBLICATION

Let me know if you agree with that method, so we can go forward and remove P5326 (P5326) for good. Totodu74 (talk) 10:49, 31 August 2021 (UTC)

• Ok for me, thanks you. Christian Ferrer (talk) 11:14, 31 August 2021 (UTC)
• @ Totodu74: Looks good to me too. Rdmpage (talk) 11:20, 31 August 2021 (UTC)
  • Ok, I am now doing it for all other 1'557 items (I mean not only the ones I was responsible for using the soon-to-be-deprecated P5326 (P5326)). This property should soon be empty or will concern items which escaped the aforementioned query. I will take a look at it tomorrow. Does any one knows how to deal with property removal? Totodu74 (talk) 16:26, 31 August 2021 (UTC)
    • As expected, there is still over 300 items using P5326 (P5326), sometimes in unexpected ways. For instance on Euretaster insignis (Q2338686) this property is used in the reference for taxon name (P225). Any help would be welcome to deprecate for good P5326 (P5326). Totodu74 (talk) 07:57, 2 September 2021 (UTC)
      • Ok, I done a few, and I will do others in the next days. Christian Ferrer (talk) 20:46, 2 September 2021 (UTC)

• @Totodu74: @Succu: @Ambrosia10: @Christian Ferrer: There is still the issue of what to do with recombination (Q14594740). This is our preferred term to flag that a reference has published a new combination. However, it is also often used as a qualifier for a taxon name (P225), see for example Peribatodes secundaria (Q33312). This causes a constraint violation: "The property instance of should not be used in this location (as qualifier). The only valid location for this property is as main value." How do we feel about this? I can see the value in flagging a name as a new combination, although arguably we can also do that using recombination (Q14594740) as a qualifier for a reference, and also whether there is a basionym linked to the item. However having a simple way to say "this is not the original name" seems useful, and it has bene used 59,000 times already, see query https://w.wiki/3zSG, so seems well established. If we're going to keep using this it would be nice to tidy this up so that we don't have a constraint violation.Rdmpage (talk) 12:39, 31 August 2021 (UTC)
  • A possible solution is to replace instance of (P31) with object has role (P3831), I thnk there are some usages already. --Succu (talk) 15:35, 31 August 2021 (UTC)
    • Yes, I'd say instance of (P31) is never a good fit in that place. +1 for object has role (P3831). has characteristic (P1552) works too, but object has role (P3831) makes even more clear which side of the statement the qualifier is about. --Azertus (talk) 15:54, 31 August 2021 (UTC)
      • In fact the presence of original combination (P1403) should be enough to generate the parentheses for a zoological name. --Succu (talk) 16:25, 31 August 2021 (UTC)
        • I agree that flagging the name as a new combination has value and also that object has role (P3831) is likely the best fit. However this too appears to throw up a constraint violation which would be nice to have sorted. --Ambrosia10 (talk) 19:47, 31 August 2021 (UTC)
          • Please note Module talk:Taxobox#Unknown value throws error. --Succu (talk) 20:00, 31 August 2021 (UTC)
            • Ok to use object has role (P3831) if that can help. Christian Ferrer (talk) 19:07, 1 September 2021 (UTC)
• This proposal looks clear to a non-expert. Could you amend or add the model item (P5869) on taxon (Q16521) to show a couple of these cases in action? The model item statements could be tagged using has characteristic (P1552) to note which scenario's they depict. --Azertus (talk) 15:54, 31 August 2021 (UTC)
  • @ Azertus: Sounds like a good idea, I'll investigate. Rdmpage (talk) 17:00, 31 August 2021 (UTC)

• OK, based on discussion so far I'm starting to bulk add links between plants names with IPNI plant ID (P961). I'm starting with original descriptions, so my code adds a link to the publication and the qualifier first valid description (Q1361864). @ Succu: I've done a query to find links you have already added so I don't duplicate them (in cases where you have additional details such as page numbers Quickstatements will create a second reference so we get the same reference twice(!). I think I've avoided this. I'm just doing a small batch at the moment linking names published by Brian Laurence Burtt (Q3286276) Rdmpage (talk) 17:00, 31 August 2021 (UTC)
  • Handling IPNI could be tricky. I added those statments for Taxon (Q2003024) some years ago. At the moment I recheck IPNI entries with my succulent plants literature. If we have an item for the paper I'm using this. Otherwise I'm using the journal/book item and adding sometimes the title of the publication. For older publication I'm try to find the corresponding BHL page ID (P687). --Succu (talk) 18:32, 31 August 2021 (UTC)
    • @Succu: I have a mapping between IPNI and publication-level identifiers (such as DOIs) that I've constructed over the years and I'm using that to create the links here. Rdmpage (talk) 18:56, 31 August 2021 (UTC)
      • I'm aware of this. Unfortunately DOI based items often have only a start page and missing the end page... --Succu (talk) 19:08, 31 August 2021 (UTC)
        • Yes, I know. I've spent some time getting that information from other sources where possible. Tedious work... Rdmpage (talk) 19:35, 31 August 2021 (UTC)
• Under #2 there's a recommendation not to use P5326 (P5326). Does that extend to taxon author (P405) and year of publication of scientific name for taxon (P574) as well, since those can be inferred from the reference? --Azertus (talk) 08:46, 4 October 2021 (UTC)
  • No, they can't be inferred from the reference (item) in all cases. --Succu (talk) 17:42, 4 October 2021 (UTC)
• @JerryL2017: Hello, I see that you reverted me, I just like to point you this discussion where we all agree to remove P5326 (P5326) and instead to use "stated in" + "reference has role". Christian Ferrer (talk) 04:23, 8 October 2021 (UTC)

  @Christian Ferrer OK, thanks for pointing out, I wasn't aware of this discussion, it makes sense. I have corrected the few times I have used Property:P5326 I think. JerryL2017 (talk) 12:35, 13 October 2021 (UTC)

Great, thanks you. Christian Ferrer (talk) 16:30, 13 October 2021 (UTC)

• @Infovarius: Hi, I noticed that you reverted me, I removed P5326 (P5326) after to have done this, you may be interested by this discussion. Christian Ferrer (talk) 19:21, 11 October 2021 (UTC)
  • Infovarius, please note this discussion. Thx. -Succu (talk) 19:51, 13 October 2021 (UTC)
• Hello Totodu74, Succu and Rdmpage, it is done, P5326 (P5326) is no longer used anywhere, see [1]. What is the next step? should someone nominate it for deletion? Christian Ferrer (talk) 16:26, 23 October 2021 (UTC)
  • Great! Good job and thank you very much for this. Indeed the next step world be Wikidata:Properties for deletion. --Succu (talk) 18:30, 23 October 2021 (UTC)
    •  Done see Wikidata:Properties_for_deletion#publication_in_which_this_taxon_name_was_established_(P5326) Christian Ferrer (talk) 13:10, 24 October 2021 (UTC)

Recombinations

Hi, a bit above it is suggested to use as qualifier object has role (P3831) instead of instance of (P31) to indicate a recombination. In order the the taxobox works well and add parentheses I suggest:

to add in the Module:Taxobox: local P_OBJECT_HAS_ROLE = "P3831" between the lines 78 and 107

to replace in the Module:Taxobox at line 601: P_INSTANCE_OF by P_OBJECT_HAS_ROLE

• We can always test that and revert in case that this don't works. However be aware that if we do that then the current items using "instance of recombination" will stop to display parentheses. Christian Ferrer (talk) 19:23, 19 September 2021 (UTC)
• @Rdmpage, Succu: I tested it and that works fine, see Goniosigma tasmanensis (Q105408368), however as I excpected the way "instence of recombination" don't work anymore. We can revert if needed. If we want to use object has role (P3831) we have also to allow that property to be used as qualifier for "taxon name". Christian Ferrer (talk) 21:03, 2 October 2021 (UTC)

Now the both works fine, see Goniosigma tasmanensis (Q105408368) for use with "object has role" and Goniosigma singletoni (Q105408360) for use with "instance of". Now to avoid the use of "instance of" as qualifier, hence constraint violations, if we want we can replace all the "instance of" by "object has role". It will be good if someone is able to do that automatically with a BOT or in a semi-automatically way. Christian Ferrer (talk) 05:01, 3 October 2021 (UTC)

There are currently 61980 "taxon name" properties that are "instance of recombination". I will file a request for User:William Avery Bot to make the change to "object has role", unless somebody else wants to do it. William Avery (talk) 11:12, 25 October 2021 (UTC)

Request has been filed at WD:BRFA#William Avery Bot 2. William Avery (talk) 17:21, 27 October 2021 (UTC)