What to see
One the right you see the silhouette of a spinning figurine. Does she spin clockwise or counterclockwise?
What to do
After you have decided which way she spins, mentally try to make her spin the other way. Yes, this is possible. But difficult. It may help to blink, or to look a special features. People differ markedly here. While it may be very difficult to voluntarily flip the direction, it may also occur spontaneously.
Note: Silhouettes are ambiguous – they can turn either way and it would look exactly the same.
But it is very difficult to flip the spin direction. Tick the checkbox “Spheres”. With a little effort, e.g. “lock hand to the spheres”, you can make the silhouette go either way too.
If you press the button “🚶 ↺⇄” the silhouette itself reverses motion direction. Wait, I just said silhouettes are ambiguous, yes, but if you tick “Lighting” you will see a full 3D model which totally disambiguates the rotatation direction.
If you tick the Eyes checkbox, you will see black eyes when the model looks at you. That also disambiguates the direction, although less then via Lighting.
- All silhouettes are ambiguous. Our brain tries to reconstruct the third dimension (space) from the flat image in our eyes, adding information which is usually realistic, but not really there. And in the case of a silhouette, there are two equally likely interpretations, leading to perceptual rivalry.
Actually, they are not completely identical in likelihood, because on left rotation the 3D arrangement is such that one looks from below – one looks at the sole of the foot. That may explain the statistical preference for rightward motion.
To unconfound the possible influence of the view angle I am now using a fully symmetrical model (see original Kayahara dance at bottom right).
- This illusion made the rounds in the Internet late 2007. And two aspects I find very obnoxious:
- The original author’s (Nobuyuki Kayahara) copyright was removed. Naughty! Here the material is reproduced with permission (granted Oct 2007, thanks again!).
- Second, it was accompanied with misguided and erroneous comments about the right and left brain hemispheres, insinuating that a given spin direction was associated with dominance of one of the two hemispheres, topping this with totally exaggerated interpretations of hemispheric specialisation, which is strongly overstated in populist claims anyway.
- During a recent presentation, I performed an informal ad-hoc experiment with the audience: I showed the silhouette for a few seconds, and then asked the audience which way they had seen the silhouette spinning.
Of the 61 women, 49 saw it turning clockwise, 12 anti-clockwise.
Of the 60 men, 50 saw it turing clockwise, 10 anti-clockwise.
Thus there was a close 50:50 sex distribution, and 20% of either saw anti-clockwise rotation.
I conclude: there is a strong preference for initial clockwise rotation, without any marked difference between the sexes.
Website of the original author, Nobuyuki Kayahara
Troje NF, McAdam M (2010) The viewing-from-above bias and the silhouette illusion iPerception 1:143–148 [PDF]
Nice variation by Marcel de Heer
I thank Vijai in Irvine, who pointed me to threejs, which sped up loading by a factor of 50!