Wikidata:Property proposal/lower rank than

next lower rank ; next higher rank edit

Originally proposed at Wikidata:Property proposal/Organization

Done: next lower rank (P3729) (Talk and documentation)

Description	higher/lower rank or level in a ranked hierarchy like sport league, military ranks. If there are several possible, list each one and qualify with "criterion used" (P1013), avoid using ranks and date qualifiers. For sports leagues/taxa, use specific properties instead.
Data type	Item
Domain	instances of police rank (Q19476593)   or position (Q4164871)   and its subclasses such as military rank (Q56019), police rank (Q19476593), ecclesiastical occupation (Q11773926) (honorific title rank; noble titles; ... any class of organisation whose members are hierarchically ordered)
Allowed values	same as domain
Example	⟨ General of the Army (Q1331037)    ⟩ next higher rank (P3730) ⟨ General of the Armies of the United States (Q3100539)    ⟩ criterion used (P1013) ⟨ special rank ⟩ ⟨ General of the Army (Q1331037)    ⟩ next higher rank (P3730) ⟨ no value Help ⟩ ⟨ Knight of the Legion of Honour (Q10855271)    ⟩ next lower rank (P3729) ⟨ Officer of the Legion of Honour (Q10855195)    ⟩ ⟨ cardinal (Q45722)    ⟩ next higher rank (P3730) ⟨ pope (Q19546)    ⟩ Subproperties for specific use cases and communities: parent taxon (P171) ; league level below (P2500) /league level above (P2499)

A pair of two inverse properties.

Motivation

I think this property is missing. Right now followed by (P156) and co. are used as a substitute, but it's not like this relation is a temporal sequence or the like, it's a hierarchical order, which is of very different nature. Maybe needed : a qualifier to link the statement to the item about the whole order. author TomT0m / talk page 17:22, 3 August 2016 (UTC)[reply]

Discussion

@TomT0m: We already have parent taxon (P171). Also, your proposal is headed "higher rank than", but the page is at "lower rank than". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:03, 3 August 2016 (UTC)[reply]
- Actually parent taxon (P171) is probably a (subproperty/duplicate) of metasubclass of (P2445) -> an instance of kingdom is a superclass of an instance of species. This proposal is probably of both. author TomT0m / talk page 19:58, 3 August 2016 (UTC)[reply]
partial support as Andy notes, we have this functionality for Taxonomic ranks so this should not be used for that domain. We also have ranking (P1352) which is similar but not really quite what you looking for here as not all ranks will be used in all places at all times and so the level will vary even if the immediately higher/lower ranks are the same. I think it would be best if we have a pair of properties here - "next higher rank" and "next lower rank" in a similar way to next crossing upstream (P2673) and next crossing downstream (P2674). Thryduulf (talk) 19:10, 3 August 2016 (UTC)[reply]
- modifications required acknowledged and done. author TomT0m / talk page 19:49, 3 August 2016 (UTC)[reply]
- On the biological taxonomy front, what I said to Andy. (I'll once more say that I think that parent taxon (P171) should be replaced by subclass of (P279). A claim
  ⟨ some taxon ⟩ instance of (P31) ⟨ taxon ⟩
  on any subclass is probably already on every taxon.)author TomT0m / talk page 19:58, 3 August 2016 (UTC)[reply]
Question Could that replace league level above (P2499) and league level below (P2500)? --Pasleim (talk) 19:12, 8 August 2016 (UTC)[reply]
Question I'd support a more generic pair of properties "subordinated to" and "superordinated to". These could be used as super-property and where no specialized propery such as league level above (P2499)/parent taxon (P171)/... and league level below (P2500)/... exists and to prevent inflation of similar properties. -- JakobVoss (talk) 13:02, 15 August 2016 (UTC)[reply]
Comment by reading the English labels of P2499 and P2500 at some point I thought we already had this. I don't think the labels of P2499 and P2500 should stay that way. In any case, I think we need properties for this: follows/followed by isn't ideal.
--- Jura 16:31, 27 August 2016 (UTC)[reply]
- I updated the labels of P2499/P2500. English labels now include "league" as well.
  --- Jura 11:04, 5 September 2016 (UTC)[reply]
I think that "higher league" and "higher rank" could use the same property. Are there any situations where this wouldn't work? --Arctic.gnome (talk) 15:48, 12 October 2016 (UTC)[reply]

Scope edit

@JakobVoss, Pasleim: Maybe we need to think a little bit of the formulation to see what the goof level of genericity is and what (usable) definition we can come up to. On subordination : per http://www.dictionary.com/browse/subordinated this does not seem to apply, as it seem to apply to instance, not to rakns themselves. You're subordinated to your boss, the Lieutenant concept is not subordinated to the General one.

The rank concept, however, seem to fit a few case per a french dictionary http://www.larousse.fr/dictionnaires/francais/rang/66440?q=rang#65696 for example :

Position occupée par quelqu'un dans une hiérarchie, et en particulier dans la hiérarchie sociale, du fait de sa naissance, de sa fonction, de sa fortune, etc.

in english "position of someone in a hierarchy, in particular in a social hierarchy, given by it's origin, function, wealth, ..." if we generalise a little bit, we can abstract "someone" by "something", and think of a rank as a name given to a special depth of a hierarchy. This work for taxons, kingdoms are highest levels taxons in the taxon hierarchy, for example. This also works for sports leagues I guess. author TomT0m / talk page 13:46, 15 August 2016 (UTC)[reply]

The lieutenant concept is subordinated to the General concept in a broader sense of being hierarchically below. I'd favour such general property while you propose a more specific property for "next higher level or rank" (see echelon for another word). We already have properties like parent taxon (P171), league level above (P2499), metasubclass of (P2445) ... for special cases but not for more general application. -- JakobVoss (talk) 18:33, 15 August 2016 (UTC)[reply]

P.S: I proposed to delete metasubclass of (P2445). If the new property also covers the intended use of P2445, I fully support it! -- JakobVoss (talk) 19:09, 15 August 2016 (UTC)[reply]

@JakobVoss: A link so that I can oppose ? This is a subproperty but it carries more meaning as meta-subclass to is defined wrt. instance of (P31) AND subclass of (P279) whereas this property would not have any reference to this. author TomT0m / talk page 20:03, 15 August 2016 (UTC)[reply]

Comment metasubclass of (P2445) appears not to have been proposed for deletion. @JakobVoss: do you still intend to do so? If not, how does this proposal interact with it? Thryduulf (talk) 14:57, 4 November 2016 (UTC)[reply]
See Wikidata:Properties_for_deletion#metasubclass_of_.28P2445.29. I support both, deletion of P2455 and creation of this new pair of properties, independently. By the way how about naming them "next lower rank / next higher rank"? -- JakobVoss (talk) 19:22, 6 November 2016 (UTC)[reply]

Domain and values edit

See ranking (Q526719) and possibly position (Q4164871). Direct subclasses of the former include military rank (Q56019), police rank (Q19476593), and taxonomic rank (Q427626). As all Wikipedias (and by thus Wikidata) do not differentiate between "ranking" (the whole) and "rank" (one part of a ranking), it is difficult whether to apply instance-of or part-of. Any domain or value item should be either "instance of a (subclass of) ranking (Q526719)" or "part of an instance of a (subclass of) ranking (Q526719)", depending on context. -- JakobVoss (talk) 19:38, 6 November 2016 (UTC)[reply]

Sorry, I don't get a word of what you mean here. ranking is not the whole, it's more the relationship beetween someone and its rank (or the act of assigning a rank to someone). I don't really see how Wikipedias are relevant here. If they don't provide us with enough items, we just have to create our own. The rest I don't understand. author TomT0m / talk page 19:46, 6 November 2016 (UTC)[reply]

Moving forward edit

@TomT0m, JakobVoss, Pasleim, Pigsonthewing, Thryduulf, Jura1: Reading the above I feel this pair of properties is close to a consensus on creation; I've modified the labels in English as JakobVoss suggested. Can any more of you voice clear support for this as is, or opposition to the creation if you still have reservations? Thanks! ArthurPSmith (talk) 16:57, 6 January 2017 (UTC)[reply]

Oppose Duplicates parent taxon (P171). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:10, 6 January 2017 (UTC)[reply]

@Pigsonthewing: I don't understand. It's clear that it's a generalisation of "upper taxon", certainly not a duplicate. Can you be more precise on you concern ? author TomT0m / talk page 19:48, 6 January 2017 (UTC)[reply]

I've edited the description and domain fields in the proposal to remove taxonomic rank as an option here. The intent is clearly not to replace parent taxon (P171) but to generalize it. Other objections? ArthurPSmith (talk) 21:21, 6 January 2017 (UTC)[reply]

Sports league should also be removed from the description and examples since we have league level above (P2499). --Pasleim (talk) 00:18, 7 January 2017 (UTC)[reply]

Or just added as subproperties. The rationale is that they are subproperties :) I think the examples should remain as to note that they are usecases, but the real subproperties should be added. Nevertheless, the misuses are just automatically fixable with no problem with petscan : from items which are instance of taxon that uses "highet rank", substitute "higher rank" with "parent taxon". We can generate a link from petscan with all this coded it petscan parameter URI. author TomT0m / talk page 09:47, 7 January 2017 (UTC)[reply]

And, just as usual, I don't think such minor issues should be blockers. This can and will move later when the properties will be used and community will start to rationalize the uses. But the right way to do can only be done afterwards and hardly guessed on paper. author TomT0m / talk page 09:49, 7 January 2017 (UTC)[reply]

please never apply PetScan replacements since you loose all sources and qualifiers.
Given that one can computationally determine the right "lower rank" property, why are multiple "lower rank" properties necessary? What is the disadvantage of generalizing the scope of league level below (P2500)? --Pasleim (talk) 18:27, 8 January 2017 (UTC)[reply]

I would not care to have only one property. Maybe taxonomists here would for reasons I don't understand. But if you generalize "higher or lower league", I'd be happy with that until we discover a usecase that don't fit. author TomT0m / talk page 18:43, 8 January 2017 (UTC)[reply]

For 1) : I guess we need a feature reques : ~~and it's here~~ here. author TomT0m / talk page 18:47, 8 January 2017 (UTC)[reply]

Comment Given that this discussion has gone in all sorts of directions, it might be worth formulating a new proposal that better defines the intended scope.
--- Jura 00:02, 9 January 2017 (UTC)[reply]

Comment I have further modified the proposal - it has changed so much since it initial form so we should better start a new discussion. -- JakobVoss (talk)

I wonder about the advantages (or disadvantages) of limiting it to the next lower/higher rank. Don't most hierarchies eventually add intermediate steps? These can remain or disappear later.
--- Jura 11:30, 10 January 2017 (UTC)[reply]

We will deal with this with qualifiers and rank, as usual. Preferred rank for the most actual values, end date and so on. author TomT0m / talk page 19:38, 10 January 2017 (UTC)[reply]

Do we have to do that? Wouldn't you just attempt to do that with the sample and realize that it might not work? Currently we have:

< General of the Army (Q1331037) > next higher rank < lieutenant general (Q2046665) >
< Catholic priesthood (Q17351675) > next higher rank < Catholic bishop (Q611644) >

I don't think either matches the definition. Maybe a qualifier "criterion used" would work better. --- Jura 07:06, 14 January 2017 (UTC)[reply]

I don't get a word of what you mean. author TomT0m / talk page 10:28, 14 January 2017 (UTC)[reply]

The samples either make no sense or are badly chosen.
--- Jura 10:45, 14 January 2017 (UTC)[reply]

Support. This property could be used in many and various cases, I suppose. It could replace a lot of specific properties.-- Andrew Krizhanovsky (talk) 05:27, 14 January 2017 (UTC)[reply]
Support. I am currently looking to figure out how to establish a rank between different art prizes ("second prize of..." lower than "first prize of..."), and see this kind of use for many other cases... Thanks, --Daehan (talk) 23:27, 25 January 2017 (UTC)[reply]

Comment I fixed the second sample as well and expanded the description.
--- Jura 09:22, 26 January 2017 (UTC)[reply]

Neutral Could use follows (P155)/followed by (P156) with criterion used (P1013) = order of precedence (Q1479879) or whatever other system of ranking is being followed. --Swpb (talk) 18:11, 15 February 2017 (UTC)[reply]
@Swpb, Jura1, Daehan, AKA MBG, TomT0m, Pasleim: Done next lower rank (P3729) and next higher rank (P3730)ChristianKl (talk) 08:16, 3 March 2017 (UTC)[reply]