Talk:Race (human categorization)

Wiki Education assignment: Evolution of the Genus Homo

This article was the subject of a Wiki Education Foundation-supported course assignment, between 29 March 2022 and 3 June 2022. Further details are available on the course page. Student editor(s): WordlyWaleed (article contribs).

T 1845/11 (Asian race/MERCK SERONO) of 26.11.2015 - Decision of the European Patent Office

I think this appeal decision should be mentioned somewhere in the Race article.

In essence, the term "Asian race" was found to be unclear by the Board of Appeal.

[1] https://www.epo.org/law-practice/case-law-appeals/recent/t111845eu1.html — Preceding unsigned comment added by 185.73.25.204 (talk) 11:22, 6 May 2022 (UTC)

This is an interesting piece of anecdotal information, but it should appear in a secondary source before we can cite it here. Anyway, there's no dearth of quality sources which explain that the category of human races (which is still perpetuated as a social construct, predominantly in countries with a longstanding segregationist history) has no evidential base in biology. –Austronesier (talk) 11:17, 15 May 2022 (UTC)

Gichoya et al. 2022

This new Lancet article finds that races (Black, White and Asian) are more than just "social constructs": https://www.thelancet.com/journals/landig/article/PIIS2589-7500(22)00063-2/fulltext Please help incorporate it into the article.--Pakbelang (talk) 10:50, 15 May 2022 (UTC)

How so, when the authors explicitly write: "In this modelling study, we defined race as a social, political, and legal construct that relates to the interaction between external perceptions (ie, “how do others see me?”) and self-identification, and specifically make use of self-reported race of patients in all of our experiments. We variously use the terms race and racial identity to refer to this construct throughout this study" ?? Be assured, the Lancet will not that easily promote fringe views. –Austronesier (talk) 11:00, 15 May 2022 (UTC)

The ability of an AI to recognise the social categories that it was trained by humans to recognise in no way invalidates the social construct theory. Regardless, this is a primary source and shouldn't be incorporated into the article when we have a wealth of secondary sources to draw on. – Joe (talk) 07:59, 17 May 2022 (UTC)

I agree with Pakbelang that the reputably published Lancet article helps stir the conversation away from rhetorical nonsense and towards the objective reality of non-discrete biogeographical clusters (which generally correspond to socially recognized racial classification or self-identified geographical ancestry). I also agree that the reputably published Lancet article should definitely be incorporated. That I could not find any mention of artificial intelligence on the page was very surprising to me. The article would benefit from a section dedicated entirely to the inference of human geographical ancestry (as non-discrete categories) from visual information which now includes medical imaging, and which is an active field of research highly relevant to science and engineering. C. M. Belanger Nzakimuena ^talk 22:34, 21 May 2022 (UTC)

Given that our article isn't about AI, there is nothing surprising about AI not being discussed. And furthermore, the Lancet article cited above says nothing to support the existence of 'non-discrete biogeographical clusters'. That isn't the subject of the study, and the authors make it perfectly clear both what they are investigating, e.g. "we defined race as a social, political, and legal construct that relates to the interaction between external perceptions (ie, “how do others see me?”) and self-identification, and specifically make use of self-reported race of patients in all of our experiments", and what they are not describing, e.g. "Race... often incorrectly conflated with biological concepts (eg, genetic ancestry)". AndyTheGrump (talk) 22:54, 21 May 2022 (UTC)

(edit conflict) Can you provide a quote from the Lancet article that actually supports the idea of an objective reality of non-discrete biogeographical clusters? We have already given one that indicates the contrary. The "inference" you are talking about is simply not made in paper, unless wilfully misread into it. The fact that AI can reproduce things that humans do (including irrational decisions) doesn't mean these things are meaningful categories in evidence-based science. –Austronesier (talk) 22:58, 21 May 2022 (UTC)

While they generally correspond (Rosenberg et al, 2002; Bamshad et al, 2004; Jorde & Wooding, 2004; Tishkoff & Kidd, 2004) and as the reputably published Lancet article highlights, I believe it can be important not to conflate (i.e. combine into one) genetic ancestry and self-reported human geographical ancestry also referred to (in the article in particular) as race (e.g. Asian). It's impressive that evidence-based science (or objective reality, as described in the reputably published Lancet article) supports that biological images contain "model decipherable information related to racial identity". Clearly in this instance (unless one completely lacks common sense), ascribing a purely socially constructed (or irrational) character to things that humans do (rather than "more a social construct than a biological construct", which is the more balanced view the reputably published article espouses) is rhetorical nonsense. I suppose I could agree that instead of a section in the present 'Race (human categorization)' article, a separate Wikipedia article solely dedicated to 'Race (machine classification)' may be more appropriate and would help leave out rhetorical nonsense. Thank you for all feedback (above). C. M. Belanger Nzakimuena ^talk 01:44, 22 May 2022 (UTC)

Have you ever read Wikipedia:No original research? AndyTheGrump (talk) 00:48, 22 May 2022 (UTC)

In case the inquiry was not rhetorical, I added references supporting the previous introductory statement. As pertains more directly to improvement of the present article, you can let me know if you support adding a 'Race classification algorithms' section or if you support the creation of a separate Wikipedia article solely dedicated to 'Race (machine classification)'. C. M. Belanger Nzakimuena ^talk 01:44, 22 May 2022 (UTC)

No. And No. AndyTheGrump (talk) 02:25, 22 May 2022 (UTC)

Concealing evidence-based science relevant to the present article without justification does not seem appropriate or consistent with a neutral point of view. Current talk page discussion objections to the addition of any content (including secondary sources) relating to race classification algorithms (both supervised and unsupervised by humans; the vast majority of which are centered on the topic of race and unconcerned with healthcare or medical imaging, the reputably published Lancet article representing a notable and relevant exception) do not seem to favor compromise. It appears that with respect to any changes, other processes should be sought beyond discussion as outlined. Thank you for your comment. C. M. Belanger Nzakimuena ^talk 16:02, 22 May 2022 (UTC)

Previous version on 03:56, 22 May before refactoring: "Concealing evidence-based science relevant to the present article without justification does not seem appropriate or consistent with a neutral point of view. Current talk page discussion objections to the addition of any content relating to race classification algorithms do not seem to favor compromise. It appears that with respect to any changes, other processes should be sought beyond discussion as outlined. Thank you for your comment." Mathsci (talk) 16:25, 22 May 2022 (UTC)

This isn't a productive discussion because you are ignoring what everyone else is telling you. The Lancet article says the opposite of what you are saying. I'll try to summarise it very simply for you: AI algorithms can detect a person's socially assigned race because they are unconsciously trained by humans to do so. It shows bias in the training datasets used in medical imaging, not "objective reality", and therefore is a problem for the application of AI in medicine. This information could be useful for articles like artificial intelligence in healthcare or medical imaging (though these are subject to WP:MEDRS, which is even more strict about avoiding primary sources) but it's not particularly relevant here and absolutely does not show that the widely-accepted social construct model of race is "rhetorical nonsense". As the authors of the Lancet article repeatedly emphasise, the ability of AI to predict racial identity is itself not the issue of importance. And you definitely can't write a whole article based on a handful of cherry-picked and misinterpreted primary sources. – Joe (talk) 09:17, 22 May 2022 (UTC)

Humans cluster by ancestry, variation clusters, and AI detects this. Nobody "unconsciously trains AI to detect race", that's just something you made up. A five year old child can detect race. You have to be pretty well educated to delude yourself otherwise. Verena Boddenberg (talk) 17:49, 27 June 2022 (UTC)

The same problems with the use of phrases like "the objective reality of non-discrete biogeographical clusters" previously resulted in a page ban from Ahmose-Nefertari. The archiving of user talk page requests by Generalrelative suggests similar bludgeoning, which runs contrary to the mainstream consensus of race as a "social construct". The edits here unduly promote a minority fringe view of "machine classification" or "classification algorithms". Mathsci (talk) 11:20, 22 May 2022 (UTC)

Racism definition in the lead, too specific

When defining racism in a sentence, it’s almost certainly going to be wrong and be unsatisfactory to someone. The one used here is just that, but knowing the difficulties I think it could be made better and more helpful. It now reads as:

“The concept of race is foundational to racism, the belief that humans can be divided based on the superiority of one race over another.”

is that really it? I would think it could be rephrased to emphasize the differentiation based on racial characteristics rather than here which confines it to superiority, and thus omits a vast amount of beliefs that might be considered racist but do not fall int that narrow definition I’d rephrase it as:

“ The concept of race is foundational to racism, the belief that humans can be differentiated or distinguished socially or politically on the basis of physical characteristics common to identifiable racial groups.”

this much better since the current description is so simple and confining as to be largely useless. Sychonic (talk) 18:46, 10 July 2022 (UTC)

Your proposal would make the sentence inaccurate: humans can, in fact be differentiated or distinguished socially or politically on the basis of physical characteristics common to identifiable racial groups. They can also be differentiated or distinguished socially or politically on the basis of any number of other arbitrary qualifications, such as ice cream flavor preference or their aesthetic opinion of the word "moist".

The 'superiority' clause which your proposal does away with is fundamental to the concept of racism, as seen in the well-sourced opening sentence of Racism. Happy (Slap me) 13:39, 11 July 2022 (UTC)

Consensus

Is it tenable to claim there's a consensus when Dawkins, Pinker and Coyne disagree?[1] Bogestra Bob (talk) 17:37, 25 July 2022 (UTC)

I don't believe you understand what Dawkins is saying in that tweet. But even if you were correct, three scientists' dissent would not undermine the consensus of literally thousands of their fellows. Happy (Slap me) 21:38, 25 July 2022 (UTC)

In the opening section it says "Modern science regards..." then links to a couple of opinion pieces. I don't see a survey of biologists. Bogestra Bob (talk) 08:56, 26 July 2022 (UTC)

Neither of the two linked sources are opinion pieces.

I strongly suggest you read the notice that Doug Weller provided you on your talk page and familiarize yourself with the subject (there happens to be a comprehensive encyclopedia article on the subject close at hand, to get you started) as well as the norms of modern science before you continue to advocate for changes to this page based (as the two comments you have made thus far demonstrably are) on your misunderstandings around the subject.

P.S. You should also read the notice on your talk page about discretionary sanctions, including all of the linked terms. That is highly useful, practical information about how to go about editing in contentious areas, and editors who edit without that information tend to quickly find themselves subject to sanctions. Happy (Slap me) 12:29, 26 July 2022 (UTC)

What I'm seeing here is all condescension and nothing about my point. Where is your survey of biologists? We're supposed to believe Dawkins and Coyne, both extremely distinguished biologists, think race isn't a "social construct", while the entire rest of the field does? Whether a concept is biological is a question for biology. Bogestra Bob (talk) 15:52, 26 July 2022 (UTC)

Yes. Jerry Coyne has called himself an 'outlier' on this subject. Consensus does not mean universal agreement. MrOllie (talk) 16:01, 26 July 2022 (UTC)

My dear friend, please answer my question. Where are you getting this "consensus" from? It appears to be assumed out of hand here. Bogestra Bob (talk) 17:34, 26 July 2022 (UTC)