Hacker News Reader

Comment by cetalingua | Hacker News Reader

Another fascinating topic is how our language determines how we see different colors. The Himba tribe does not distinguish between blue and green, but is able to discriminate among shades of green that basically look the same to us, Westerners.

https://6thfloor.blogs.nytimes.com/2012/09/04/its-not-easy-s...

View on HN

That probably never happened. http://languagelog.ldc.upenn.edu/nll/?p=17970

From the actual study: "Himba participants show categorical perception only for their own linguistic categories and not for either the supposed universal categories of English"

Roberson, D., Davidoff, J., Davies, I. R., & Shapiro, L. R. (2005). Color categories: Evidence for the cultural relativity hypothesis. Cognitive psychology, 50(4), 378-411.

On top of that there is also Berinmo tribe that " marks a color boundary that is absent in English (i.e., a boundary within the English green category), but also fails to mark other boundaries that are found in English (e.g. the boundary between blue and purple (Roberson, Davidoff, Davies, & Shapiro, 2005)"

That reporting is completely factually incorrect [1]. There is no evidence that the tribes are unable to distinguish colors or are able to see colors that other people can't.

[1] http://languagelog.ldc.upenn.edu/nll/?p=18237

Actually, this statement is not factually correct either. While there might be a debate about Himba tribe and colors, there are other tribes like Karajá and Waorani, as they both are counterexamples to universals of color naming, which could imply a possibility of perceptual differences as well ( see Regier et al., 2009)

http://muse.jhu.edu/article/369843

could imply a possibility of perceptual differences

isn't quite the same as

is able to discriminate among shades of green that basically look the same to us

If you read the paper cited you will see it is even more dramatic than that: "Thus, the color-naming system of Waorani, like that of Karajá, does not fit the structure of perceptual color space. Although neither Waorani nor Karajá has been regularly held up as a counterexample to universals of color naming, they both are genuine counterexamples"

Also, from the actual Himba study (by the way,the study was conducted and published) "Himba participants show categorical perception only for their own linguistic categories and not for either the supposed universal categories of English" Roberson et al., 2005 (full citation in another comment)

Huh? I just ran a color picker over the squares in the article's pic [1]. They're all RGB 0/199-200/0 except for a left mid-upper one, which is 0/206/0 ... is that what sticks out to this tribe?

Then I follow the link to the original [2] chart, which claims RGB values of 80/186/15, except for the odd one which is 97/192/4. But my color picker is showing the same as before: everything is 0/199-206/0.

Something is not right here.

Edit: Here are the two colors he intended, based on the labeling of [2]. I can tell them apart.

https://www.webpagefx.com/web-design/color-picker/50BA0F

https://www.webpagefx.com/web-design/color-picker/61C004

[1] In Mac OS X you can pull up the "digital color picker" from spotlight search.

[2] had to to change to http from https ... should be safe for this context: http://origin.boingboing.net/wp-content/uploads/2011/08/ring...

This is why this has always been confusing to me.

Those aren't the same colors. They're both greens, but different in both exact hue and in brightness. This has nothing to do with words, but with the precision of the human eyeball.

That is the experiment. All the boxes are the same color except for one (the oddball), which the subject is supposed to identify. The original study claimed that the Himba were (slightly?) faster to pick the oddball when it was a slightly different shade of green than UK English speakers. When the oddball was blue, UK English speakers were faster than the Himba.

Not to mention the...unusual choice of highly-compressed-JPG for an image containing only solid blocks of similar colors.

"Digital Color Meter" in my version of macOS.

Reply on news.ycombinator.com