I think it’s a little simpler than that. Our brain’s default assumption is that light comes from “above”. The very box you’re typing your post into looks indented because the lower-right edge is bright and the upper-left is shaded.
That’s what we see at first here (light from upper-left)–but it doesn’t quite look right. The ridges cast a shadow on the bottom of the plates, making them look wider. Some of the perspective looks slightly off as well. And once we detect this, somehow we can’t easily go back.
The illusion worked for me as described, but I was able to squint a little and make the plates look convex again. And what I noticed is that in that mode, the plates don’t appear to lie on the same plane–they’re tilted at odd angles, both relative to the table and each other. It really “wants” to snap back to the actual interpretation since it makes more sense physically as well.
Septimus, it looks like you are onto something…but I had a hard time following it exactly, because I think maybe you accidentally switched references to right and left, or to same side vs. opposite side?
ETA: I wonder why no one seems to see the plates as facing up right from the beginning?
When I write “Light at right, plate concave” I mean the light source is at right, so the light (unshadowed) part of the plate is at left.
Your wonder-why question is a good one! But note that your eye will not initially be drawn to the little bowls in the left corner that make it obvious the light source is from the right. Instead your eye will be drawn to the oddly shaped plates where the shadowing is ambiguous, but slightly favors belief that the light source is at left.
It’s a version of the Crater illusion - Wikipedia and as Dr. Strangelove says it has to do with innate assumptions about the position of the source of light.
I can sort of switch back and forth, but, as also described, there’s something off about the shapes and angles for the upside down version to persist.
I can absolutely not switch back and forth, which is one of the things I find most bizarre about this.
With those craters: why does flipping it 180 degrees make a difference?
Aha, so my “I wonder why…?” was premature. This makes sense to me, as I am increasingly finding that it’s difficult for me to recapture the original illusion.
(To recover the original illusion, use your hand to block sight of the bowls at lower left.)
Again: The 2 (or 3) ordinary round bowls at lower left are very clear-cut. Once you see them you know the light source is shining from the right. You are so certain of this that you know none of the plates are upside down — the shadows would be wrong.
But before you notice the 2 clear-cut round bowls in lower left, you see ambiguous shadows on complex shapes. For example, consider the round bowl in the middle, far right. When you think it’s upside down, you think it’s a deep bowl; but when it’s right-side up, it’s shallower! The shadows on those complex plates work either way. (They have to work either way, or the illusion wouldn’t operate properly.)
But even though those complex shadows work either way, they work *somewhat better *if those plates are upside-down — so that’s what your brain goes with before it notices the clear-cut bowl shadows at lower left.
My experience with the plates in the illusion above is a simple one. When I first saw the image, all the plates looked convex. After reading the text below the image I figured it all depended on me whether the plates or upside down or right side up. Then all the plates looked concave and stayed that way.
Look at the world in front of you and think of it as a flat picture. Where are the natural light sources likely to be?
The brain, developed to interpret shapes based, among other things, on shade and light in a natural environment, starts with the assumption that things are illuminated from above. Reorient a picture so that the real light source is above and the brain has less of a problem figuring out the real 3D geometry.
I wonder if it would matter if one placed the images on the floor and actually looked down.
Yeah, I keep going back to this illusion every so often, and I can’t make it look any other way. About the only thing is that big platter on the center left–that one I can sorta see upside down if I really want to, but every time I look at the image as a whole and then at the individual parts, I see a face up orientation.
For those of you who have only seen them as concave from the beginning, have you tried covering the left side of the picture before you look at it?
Even when you go back later? Like JAQ, I initially found that it didn’t take long away from the image for it to go back to convex again. But it’s becoming increasingly difficult to get back into that “virgin” mode again. One thing I love about these kinds of illusions is how it gives us a peek into how non-conscious parts of our brains are constantly doing sophisticated stuff we are totally unaware of. We can know that rationally, based on reading about brain science, but this kind of thing really brings it home.
BTW, one little thing that bugs me about the text is its grammar. It isn’t quite accurate, amirite? I wonder if it was written by a non-English speaker, maybe someone East Asian from the syntax.
Yeah, I just tried that. Just put my hand far enough in front of the monitor to block where the post would be and slowly revealed the picture from the right side to the left, and still nada. Look like plates facing up to me.
OK. I’ve been starting at the whole thing for a minute now, and I finally got bits of the right side to “flip” for me by forcing myself to logically think of the shadows as forming the ridge of a peak rather than a valley, but, for whatever reason, when taken as a whole, my brain doesn’t like the perspective or something and wants to make everything concave/facing up in the photo.
My sister said yesterday on Facebook that the illusion didn’t work for her, and I naturally assumed it was something similar. But today she clarified that “they all look face down” no matter how long she looks at it. :eek: How is that possible?!?
Maybe it’s one of those gold dress/blue dress things. Or that 3D model pirouetting clockwise vs counterclockwise. For the longest time, I could only see that one spinning in one direction. Gold dress vs blue dress I can’t get to flip, though, no matter how hard I try. The 2D illusion in the original OP here, though, is one of the most mind-blowing ones I’ve seen.
If I squint enough to fuzz out the smaller details of the graphic. The major axis become straight. I suspect it is the slants and curls of the smaller details that cause you to start mentally shifting everything to align, when they can’t. Make the details less apparent and effect diminishes. Less things to try and correlate.
It may also play on our ability to quickly identify patterns. We process the two patterns and try and blend them into one pattern with similar aspects. One gets bent to fit.