Blue/black vs. white/gold isn’t the point. How #dressgate has opened up many more interesting questions about vision and the brain.
It’s not often that the Twitterverse gets excited about how light, optics, and the brain interact to produce perception. If this enthusiasm had to be drawn out by a fashion mystery, so be it. The scientific explanation for the blue/black vs. white/gold debate – a failure of a perceptual mechanism called color constancy – has been presented well, for example by Adam Rogers over at Wired, who thankfully got some top-notch vision scientists on record. Buzzfeed of course got more than its share of clicks from the uproar, and also included a handy poll in its post, showing that 75% of people see the dress as white and gold.
Here’s what I think most of the coverage has missed: while there has been plenty of chatter about the “deceptive brain”, no one has explicitly stated that what most people see – a white/gold dress—is actually the right answer in this situation, regardless of whether or not the dress “objectively” is blue/black. The brain and the visual system evolved to make use of the only major light source available (at least until electricity came along), the sun. This illuminant has predictable properties – an output spectrum and modestly varying sunrise and sunset times – that the brain internalized. This is how color constancy came about: the ability to see a stop sign as red or grass as green regardless of what kind of light is illuminating it. But when the brain is confronted with impoverished input, it makes an educated guess.
The best guess in the case of the photo of the dress is white/gold. A lot of natural cues about the illumination situation have changed or been taken away. First, it’s a photo, so it’s second-hand information. People have problems interpreting photos in other domains too, for example depth. Second, the original illumination situation is murky. Was a camera flash used and was the dress under fluorescent lighting? Third, we are of course all viewing the photo of the dress on screens. These might be HD displays or smartphone screens with various degrees of backlighting, under different lighting sources (the fluorescents at your office, the gloom outside or the solitary incandescent in your basement). Or maybe you printed the photo, altering the viewing conditions yet again.
The point is, if you see white/gold, you are correct in this viewing situation. All the information your brain has received screams “white/gold”, so there actually couldn’t be any other answer. The question of what color the dress actually is is secondary and boring. Photoshop sleuths have broken down the image in exhaustive pixelated detail. Those patches are indeed blue and black. The far more interesting questions are
It would be great to see journalists follow up with these kinds of questions, with input from scientists and the scientific literature. (I also see a lot of fascinating experiments for a dissertation, but that’s the recovering scientist in me talking.) Most of the public never even considers that the experiential phenomenon of color is constructed by the brain, and that light itself is not colored, though we often erroneously refer to it that way. (Ferris Jabr got it quite right in this Twitter conversation.) You’d think a story that delved deeper into color perception would get a few clicks – posts about the amazing eyes of the colorful mantis shrimp, for example, never get old.
So thanks, Taylor Swift, Kim Kardashian, and all the rest, for getting people excited, even if only momentarily, about vision science!
Further reading:
Original post on Tumblr by Caitlin McNeill
NPR: How Animals Hacked The Rainbow And Got Stumped On Blue