I can sometimes get her to go counterclockwise in my peripheral vision, but as soon as I look right at her, bam, clockwise every time.
I checked carefully after enlarging the gif, and her foot actually passes higher when going rightward than leftward, so the normal perspective assumption that we’re above the ground makes it look clockwise. Try assuming that she’s spinning on a glass floor slightly above you, viewed from slightly below (or equivalently that the axis she’s spinning about is tilted backwards rather than being vertical — hard to do without support): for me that reliably reverses it, and what would give correct perspective for the path of her foot.
So your clockwise model is actually more veridical, under the reasonable assumptions that no glass floors or messing with gravity are involved.
Neat! I have confirmed your observation about the relative positions of the two feet being different in the two directions.
It’s pretty subtle though—the “right cue variant” in the show/hide box here (direct links: left, right) doesn’t look obviously “from below” to my eyes. Maybe a bit? Hard to say.
So your clockwise model is actually more veridical
I disagree with this part: I would say “So your clockwise model actually has a stronger prior”.
As I wrote: “A veridical model of the real-world thing you’re looking at would feel like a 2D pattern of changing pixels on a flat screen—after all, nothing in the real world of atoms is rotating in 3D!”
Schröder’s stairs likewise has the property that one of the options tends to be stronger-prior than the other. (Most stairs are on the floor not ceiling.)
I checked carefully after enlarging the gif, and her foot actually passes higher when going rightward than leftward, so the normal perspective assumption that we’re above the ground makes it look clockwise. Try assuming that she’s spinning on a glass floor slightly above you, viewed from slightly below (or equivalently that the axis she’s spinning about is tilted backwards rather than being vertical — hard to do without support): for me that reliably reverses it, and what would give correct perspective for the path of her foot.
So your clockwise model is actually more veridical, under the reasonable assumptions that no glass floors or messing with gravity are involved.
Neat! I have confirmed your observation about the relative positions of the two feet being different in the two directions.
It’s pretty subtle though—the “right cue variant” in the show/hide box here (direct links: left, right) doesn’t look obviously “from below” to my eyes. Maybe a bit? Hard to say.
I disagree with this part: I would say “So your clockwise model actually has a stronger prior”.
As I wrote: “A veridical model of the real-world thing you’re looking at would feel like a 2D pattern of changing pixels on a flat screen—after all, nothing in the real world of atoms is rotating in 3D!”
Schröder’s stairs likewise has the property that one of the options tends to be stronger-prior than the other. (Most stairs are on the floor not ceiling.)
Fair enough, if this silhouette was actually generated from a mathematical 3D model rather than a real dancer, as I assume it was!