Wikidata:Property proposal/reduces

From Wikidata
Jump to navigation Jump to search

reduced by[edit]

Originally proposed at Wikidata:Property proposal/Generic

   Not done
Descriptionaction or phenomenon that caused or causes the subject to be diminished in quantity, size, or severity
Data typeItem
Example 1well-being (Q7981051)mental depression (Q4340209)
Example 2animal welfare (Q459426)habitat destruction (Q552431)
Example 3environmental change (Q16934857)sustainable development (Q131201)
Example 4social media impersonation (Q110264081)account verification (Q55603125)
Planned useMy colleagues and I are seeking to expand Wikidata's representation of sustainability and well-being, and plan to make contributions in this domain over the next several months.
See alsohas cause (P828), has immediate cause (P1478), has contributing factor (P1479)

reduces[edit]

Originally proposed at Wikidata:Property proposal/Generic

   Not done
Descriptionsomething that is diminished in quantity, size, or severity by this action or phenomenon
Data typeItem
Example 1mental depression (Q4340209)well-being (Q7981051)
Example 2habitat destruction (Q552431)animal welfare (Q459426)
Example 3sustainable development (Q131201)environmental change (Q16934857)
Example 4account verification (Q55603125)social media impersonation (Q110264081)
Planned useMy colleagues and I are seeking to expand Wikidata's representation of sustainability and well-being, and plan to make contributions in this domain over the next several months.
See alsohas cause (P828), has immediate cause (P1478), has contributing factor (P1479)

Motivation[edit]

There are a number of properties that reflect various aspects of causality, described here. However, there are many situations where there is an inverse causal relationship rather than a direct causal relationship. While the absence of a cause can be modeled with Property:P9353, I cannot find a property that effectively captures the concept that one item reduces another item. Therefore, I believe a "reduced by" property would allow for richer representation of causality on Wikidata.

Wmtwmt (talk) 16:38, 9 September 2022 (UTC)[reply]

Discussion[edit]

  •  Strong support AntisocialRyan (Talk) 18:46, 9 September 2022 (UTC)[reply]
    account verification (Q55603125) reduces social media impersonation (Q110264081). AntisocialRyan (Talk) 18:49, 9 September 2022 (UTC)[reply]
    Thanks! Great example! Wmtwmt (talk) 19:24, 9 September 2022 (UTC)[reply]
  •  Comment @AntisocialRyan, Wmtwmt: I think this needs more discussion before it is ready to be created. For example, can we use existing properties? (I have added several to see also)? Also, existing properties in Wikidata property for causes and effects (Q28746712) are all modeled as <effect>has cause<cause> but the proposed property is used in the form <cause>reduces<quantity>. Why is that preferable to <quantity>reduced by<cause>? The latter seems like it would be more useful, since it would, for example, put a list of items that harm animal welfare on animal welfare (Q459426). Also, if we create a 'reduces' property than shouldn't we also make the opposite 'increases' property? — The Erinaceous One 🦔 22:28, 18 September 2022 (UTC)[reply]
    I agree with these statements. has cause (P828) is pretty close to "increased by" but if it isn't close enough then I think it should be a property as well. AntisocialRyan (Talk) 23:22, 18 September 2022 (UTC)[reply]
    Thank you for your comments, @The-erinaceous-one and @AntisocialRyan. I agree that this proposal is part of the broader goal of modeling causation, and hence relates to the properties you added to see also. With regard to the directionality of documenting causality, I feel like it is useful to document these relationships in both directions; a list of 'causes' statements, and a list of 'caused by' statements, would both be helpful in understanding how an item fits into its broader context. I wonder if modeling causality would be a good context in which to use the "inverse constraint" functionality, such that the addition of a 'has cause' statement should also elicit the addition of an inverse statement 'causes' and vice versa? (I recently encountered inverse constraints when I added a bunch of my doctoral advisees and Wikidata requested that each of them "should also have the inverse statement doctoral advisor".) Similarly, I feel it would be helpful to have pairwise 'reduces' and 'reduced by' relationships. I don't want to foster an unnecessary proliferation of redundant information, but I feel there is value in having explicit lists of both the things that X reduces, and the things that X is reduced by. (Perhaps even better, but probably unfeasible, would be to have a computational infrastructure where adding a 'reduces' or a 'reduced by' property automatically populates the inverse property, or where a single internal representation manifests as two properties.) I do recall having found instances, in the sustainability space, where there are only-partially-overlapping lists of 'has cause' and 'has effect'. For example, global warming (https://www.wikidata.org/wiki/Q7942) 'has cause' greenhouse gas (https://www.wikidata.org/wiki/Q167336), and greenhouse gas has a bunch of 'has effect' relationships, but global warming is not currently among them. While it's easy to fix that specific instance, it seems like this is an instance of a class of issues that would be well-suited for computational automation.
    For the discussion at hand, to be consistent with 'has cause' and similar properties, I agree that 'reduced by' is a better formulation than 'reduces'. Is it okay for me to edit the text of this proposal to change its title and flip its examples, or is it better for me to start a new proposal for 'reduced by' (and if so, would it be okay if I set its status flag to 'ready' since the required week-long period has already occurred)?
    Even though I agree with @AntisocialRyan that 'increased by' is similar to 'has cause', it does feel like 'increased by' could be useful for some statements ("sea level increased by global warming", "global temperature increased by global warming', 'crop yield increased by fertilizer'). Would you both be supportive if I were to make an 'increased by' proposal as well?
    Thank you again for your help thinking this all through! Wmtwmt (talk) 20:28, 19 September 2022 (UTC)[reply]
    @Wmtwmt: You can definitely edit this proposal however you want and it'll be fine, even if the page title is wrong. Many proposals do that. I don't believe we're allowed to flag our own proposals as ready, though I'm not sure.
    Your increase examples are good, I would support an "increased by" property too. AntisocialRyan (Talk) 20:33, 19 September 2022 (UTC)[reply]
    Great, thanks! I changed it to "reduced by". I also added your example (from your initial message) to the list of examples. Hope that's okay! Wmtwmt (talk) 20:56, 19 September 2022 (UTC)[reply]
    I also just created an increased by (https://www.wikidata.org/wiki/Wikidata:Property_proposal/Generic#increased_by) proposal. Wmtwmt (talk) 21:14, 19 September 2022 (UTC)[reply]
    Awesome AntisocialRyan (Talk) 21:27, 19 September 2022 (UTC)[reply]
  • Do we want to make the complementary properties for "reduced by" and "increased by" (i.e., "reduces" and "increases")? It might be a good idea to add them to this property proposal instead of waiting until a later date. — The Erinaceous One 🦔 07:39, 21 September 2022 (UTC)[reply]
    The proposal is now currently only "reduced by" and "increased by", and no longer "reduces" nor "increases". I do like both approaches though. AntisocialRyan (Talk) 14:01, 21 September 2022 (UTC)[reply]
    I think it makes sense to have all four (potentially with inverse constraints connecting each pair). Since I'm relatively new here, is it better to change this proposal back to "reduces" (to match its title), and then make two more for "reduced by" and "increases" (since I made an "increased by" earlier)? Wmtwmt (talk) 15:55, 21 September 2022 (UTC)[reply]
    I don't think it's important that the name of the proposed property matches the title of the proposal. I'll copy and modify the property proposal template at the top of this page so that we have one for each proposed property (I'll let you do the same thing for the "increased by" proposal that you made). — The Erinaceous One 🦔 08:27, 25 September 2022 (UTC)[reply]
  • @Wmtwmt: I added the template for reduces and modified the descriptions of both to increase clarity. I am second guessing, however, the idea to add both a reduces and reduced by properties. I think this just leades to duplication of information. Perhaps it would be be better to just stick to reduced by. (@ChristianKl: I seem to remember you having an opinion on the creation of inverse properties in the past. Do you want to weigh in here?) — The Erinaceous One 🦔 08:49, 25 September 2022 (UTC)[reply]
    I have been thinking about this over the last few days, and doing some reading in the Wikidata discussion of how best to model causality. Here are some bits that I have found that I believe are relevant.
    ---
    https://www.wikidata.org/wiki/Help:Modeling_causes
    "The properties has cause, has immediate cause and has contributing factor are related to other properties:
    Inverse properties: has effect (P1542) and does not have effect (P7167), immediate cause of (P1536), contributing factor of (P1537)"
    ---
    ---
    https://www.wikidata.org/wiki/Property_talk:P828#A_better_way_to_model_causation
    "we would probably benefit from having inverse properties [for 'has cause', 'has immediate cause', and 'has contributing factor']"
    ---
    These arguments support the proposal to have both 'reduced by' and 'reduces', and both 'increased by' and 'increases'.
    Therefore, I would still be interested in having both pairs, with inverse property relationships for each pair.
    Does that sound okay? Wmtwmt (talk) 11:58, 28 September 2022 (UTC)[reply]
    Sounds fine to me. — The Erinaceous One 🦔 09:20, 30 September 2022 (UTC)[reply]