Perception · about 3 minutes
One At A Time
Twelve straight lines, drawn flat on a flat screen, not moving. There is no depth in them and never was -- depth is something you are adding.
You are about to add it two different ways, involuntarily, over and over, and you will not be able to do both at once no matter how hard you try.
Look at the cube. Which face is nearest you -- the lower-left one, or the upper-right? Keep looking. It will change its mind. Tap the button every time it flips.
Thirty seconds, when you're ready
0
Now two instructions, in order. First: hold both at once. See it as both cubes simultaneously -- near-face-low and near-face-high, together, for one second. Take as long as you need.
Then: make it stop. Pick whichever cube you like and hold it there. Do not let it flip. Stare it into submission.
It is not just cubes. Here is Edgar Rubin's figure from 1915 -- a vase, or two faces about to argue. Same edges. Same ink. You will get one at a time, and you will get whichever one your brain has currently decided the black belongs to.
A vase. Or two profiles. Never a vase being looked at by two profiles.
The input never changed
This matters more than it sounds, so hold onto it: nothing on the screen moved. The photons leaving that cube were identical throughout. Every flip you counted happened entirely on your side of the glass, in a system that was receiving a constant signal and kept changing its answer about what the signal meant.
The drawing is ambiguous -- it is a perfectly good projection of two different three-dimensional objects, and the flat image contains no information that could settle which. Your visual system is not built to report ambiguity. It is built to deliver a scene, and a scene has to commit. So it commits, gets restless, and commits the other way.
This is the same machinery as Exhibit № 31, caught in the act of not being able to decide. Perception is your brain's best guess about what is out there -- and when the evidence genuinely supports two guesses equally, you get to watch the guessing happen.
The two things you couldn't do
Both failures are more interesting than the flipping.
You could not see both at once. Not "it was hard" -- there is no version of the experience where both cubes are present together, in the way you can hear two notes at once or feel warmth and pressure at once. Conscious experience of an ambiguous figure appears to be strictly exclusive: one interpretation occupies the slot, and the other is not dimmed but absent.
That exclusivity is not a quirk. It is written into one of the leading theories of consciousness as a founding postulate -- integrated information theory (Exhibit № 29) holds that an experience is always definite, one thing rather than a superposition of candidates, and points at exactly this sort of demonstration when asked why.
And you could not hold it still. You can bias it -- most people can hold a preferred interpretation longer by attending to a particular corner -- but you cannot lock it. Somewhere below the level where your intentions operate, something gets bored of its own answer and reopens the question. You are not consulted. You are informed.
What it costs to be shown a world
Every room in this wing has been telling you the same thing from a different angle: what you experience is not the input. Exhibit № 1 showed you an invention where there was no signal at all. Exhibit № 2 showed you motion invented between two stills.
This one shows you the invention changing its mind while you watch, which is the closest any of these rooms gets to letting you see the workings. The cube is not flipping. You are.
Take it further
The cube is Louis Albert Necker's, from 1832, and he noticed it in drawings of crystals -- he kept getting the wrong face and wrote to a colleague about it. Rubin's vase is 1915. The modern research programme uses binocular rivalry, where each eye is shown a different image and consciousness alternates between them; it is one of the main tools for hunting the neural correlates of consciousness, because the input stays fixed while experience changes -- which is exactly the experiment you want and almost never get.
One honest note: your flip count is not a measurement of anything about you. Rates vary hugely with how you look, what you are told, how tired you are, and whether you are trying to influence it -- which most people cannot help doing once they are counting.