Wikidata talk:Requests for comment/How to capture negative results in Wikidata?

Question from Photocyte after the closing of the RfC

Photocyte (talk) 23:51, 12 March 2022 (UTC) @Andrawaag, Egon Willighagen, TomT0m, ArthurPSmith: Hello folks. I happened to find this discussion via Google & am wondering if a consensus or a best practice has been reached since 2019? My naive view was a qualifier could be used to negate a certain statement. I.e., using a biological example, all known fireflies (Lampyridae) are bioluminescent in at least the larval life stage, so "has quality: bioluminescence" and "uses: bioluminescence" on the Lampyridae (firefly family) Wikidata item is accurate. But let's say, a single species of firefly is discovered which lacks all ability for bioluminescence. I was imagining putting the "has quality: bioluminescence" and "uses: bioluminescence" on the Wikidata item for this "non bioluminescent firefly" species, but negating the statement with a qualifier like a logical NOT. But wasn't sure if that was "accepted" & wasn't able to find a good list of the extant qualifiers...

Hi @Photocyte:, I moved your comment because the RfC has been closed. It’s imho a bad idea to add a « not » as a qualifier because the qualifiers are removed if the statement is taken as a « truthy » statement in the query service for example, qualifiers are not taken into account. So you risks to inverse the meaning and make the « truthy » form unusable if you are not careful enough with ranks.

To capture negative results we have does not have characteristic (P6477) or does not have part (P3113). It’s true that it does not make automated treatment of inheritance trivial either, however. author  TomT0m / talk page 10:59, 13 March 2022 (UTC)

Thanks TomTom, On your opinion that the RfC being closed means further discussions shouldn't happen on the Project page. My comment on your talk page (https://www.wikidata.org/wiki/Topic:Wrp6a70659u0fa3j) addressees how I disagree with this - if you think this is "the law", the guidance on the page that I highlight on your talk page is ambiguous & should be updated. When I wrote my question on the Project page, the discussion tab was not active. For discoverability of these discussions via Google (How I found the RFC in the first place), I feel it is better to keep follow up discussions on the project page. For example, to now put your response in the context of the other discussions, I have to switch back and forth between two tabs, which is annoying. But regardless, I appreciate your feedback. It seems like there hasn't been a consensus on how to do this in a few years time... Photocyte (talk) 20:17, 13 March 2022 (UTC)

This seems like an especially relevant article to the discussion in context: "Negative Knowledge for Open-world Wikidata" https://dl.acm.org/doi/10.1145/3442442.3452339 . It seems the "Open world assumption", is at play... Edit: These two articles were useful as well. "Negative statements considered useful" https://www.sciencedirect.com/science/article/pii/S1570826821000366 , "A Minimal Deductive System for RDFS with Negative Statements" http://arxiv.org/abs/2202.13750 Photocyte . My take away from the latter article is RDF & RDFS can't handle negative statements, so, I assume Wikidata would have a hard time as well, given its choice be compatible with RDF for export (see: http://link.springer.com/10.1007/978-3-030-00668-6_23) (talk) 21:27, 13 March 2022 (UTC)