Do we perceive reality? The checker shadow illusion.

January 8, 2011 • 10:11 am

I recently finished Steve Pinker’s The Blank Slate (recommended!), and in one chapter was taken by his discussion about whether human senses perceive a real, external reality or whether that reality is somehow “constructed” socially or by our senses.  If you’ve read the book, you know that Pinker comes down on the “it’s real” side (this solution is obvious to all but a moron*)—but not always: what we perceive as “real” is sometimes distorted by our expectations.  That, of course, is the basis of optical illusions.

One of Pinker’s examples was the famous “checker shadow illusion,” which completely bamboozled me. In case you haven’t seen it, I’ll reproduce it here.

Take a look at this checkerboard, and at squares A and B.  They’re different shades, right?

Nope—they are exactly the same color and shade!

Don’t believe it? (I didn’t.)  Check out this video, which puts a black background behind the squares to show that they’re the same color:

If you’re still not convinced, there’s another demonstration here.

This famous illusion was produced by Edward Adelson, a professor of vision science at MIT.  (I’ve given references to two of his papers below, the second of which has some other cool video illusions.)  According to Pinker (and I buy his point), these illusions are hitchhiking on evolved adaptations of our visual system.

Here’s Adelson’s explanation for the checker shadow effect, which involves how our visual system distorts external reality as way to compensate for how things should look under situations of shadow and local contrast.

Why does the illusion work?

The visual system needs to determine the color of objects in the world. In this case the problem is to determine the gray shade of the checks on the floor. Just measuring the light coming from a surface (the luminance) is not enough: a cast shadow will dim a surface, so that a white surface in shadow may be reflecting less light than a black surface in full light. The visual system uses several tricks to determine where the shadows are and how to compensate for them, in order to determine the shade of gray “paint” that belongs to the surface.

The first trick is based on local contrast. In shadow or not, a check that is lighter than its neighboring checks is probably lighter than average, and vice versa. In the figure, the light check in shadow is surrounded by darker checks. Thus, even though the check is physically dark, it is light when compared to its neighbors. The dark checks outside the shadow, conversely, are surrounded by lighter checks, so they look dark by comparison.

A second trick is based on the fact that shadows often have soft edges, while paint boundaries (like the checks) often have sharp edges. The visual system tends to ignore gradual changes in light level, so that it can determine the color of the surfaces without being misled by shadows. In this figure, the shadow looks like a shadow, both because it is fuzzy and because the shadow casting object is visible.

The “paintness” of the checks is aided by the form of the “X-junctions” formed by 4 abutting checks. This type of junction is usually a signal that all the edges should be interpreted as changes in surface color rather than in terms of shadows or lighting.

As with many so-called illusions, this effect really demonstrates the success rather than the failure of the visual system. The visual system is not very good at being a physical light meter, but that is not its purpose. The important task is to break the image information down into meaningful components, and thereby perceive the nature of the objects in view.


*As expressed in this limerick:

A Christian Scientist from Theale

Said, “Though I know pain isn’t real,

When I sit on a pin

And it punctures my skin

I dislike what I fancy I feel.”

Adelson EH (1993) Perceptual organization and the judgment of brightness. Science 262:2042–2044

Adelson EH (2000) Lightness Perception and Lightness Illusions. In The New Cognitive Neurosciences, 2nd ed., M. Gazzaniga, ed. Cambridge, MA: MIT Press, pp. 339–351


48 thoughts on “Do we perceive reality? The checker shadow illusion.

  1. I’ve always liked that illusion. I can force myself to see that they’re the same color by just concentrating on the image, and once I do so, I can’t turn it back off until I stop looking at the image and clear my mind.

    1. I’ve always liked it for the opposite reason. It’s the only illusion I’ve come across that I can’t see correctly no matter how hard I try. Other illusions tend to vanish once I know the trick.

      1. Try “looking past” the image, so that you still see everything in it, but you’re mentally focused on something else (not necessarily anything actually visible).

        If you succeed in preventing your brain from parsing the image as a scene, you’ll be able to see the squares as the same color. As soon as you pay attention to them again, however, the scene is translated, and they revert to looking different.

        1. Grrr…still can’t do it. May have to do with the angle of my screen–colors change darkness on my monitor depending on how I have the screen angled. I tried covering up all the squares, too! Go try again…

          (*Love* The Blank Slate–read it twice, parts of more than that. Pinker is the only linguist I know of who quotes Monty Python. In full. Props for that!)

  2. I opened the first image in a graphics program and compared the two squares. In fact, they are not precisely the same shade. If you place the colors side by side, A actually looks a little lighter!

    1. The real way to test it is to use the color picker tool to sample the color in both squares. If you do that, you’ll see that they’re identical: #787878 (ignoring JPEG compression artifacts that change the color imperceptibly on some pixels).

      1. That’s what I did, and the sample from A was a different color (at least on the pixel I chose). But as you say, it is undoubtedly a compression artifact.

  3. So, perhaps our visual systems have evolved to see differences and nuances between colors and shades, as much as it has evolved to see colors and shades as they really are? Perhaps seeing differentially is always at work, and perhaps seeing differentially has greater survival value than seeing veridically. But, having read my share of Fred Dretske’s articles and books, I suspect that shades/colors seen (veridically or not) must exist (or must have existed at the time of adaptation of the visual system) in order for our visual system to adapt to them in the first place.

  4. Good lord that video was dry; a lesson in how to make an interesting visual illusion dull through a monotonous and repetitive voiceover.

    1. Indeed; anyone who uses the word “phenomenological” twice does not know how to talk to a lay audience. (I still don’t know what it means.)

      1. “Of the phenomena” (I think), i.e. without (known or considered) theory.

        For example:

        Phenomenologically, generally objects that are sunlit gets hot.

        Physically, such objects absorb visible light (and radiate back in IR) and heat up from radiant energy.

  5. I don’t think of this as an illusion. Rather, I see it as an illustration of what our vision system is doing. It is attempting to identify properties that are intrinsic to what we are looking at. To achieve that, it is using both the information on light reflected off the surface and information on the ambient light level at the surface.

    1. Agreed. The only sense in which this is an “illusion” is that the scene itself is imaginary. When confronted with an actual scene of this sort, our visual system functions correctly to determine the actual relationship between the intrinsic colors of A and B.

  6. Purves, D., R. B. Lotto, S. M. Williams, S. Nundy, and Z. Yang. 2001. Why we see things the way we do: evidence for a wholly empirical strategy of vision. Philos Trans R Soc Lond B Biol Sci 356:285-97.

    Here. There was also a more accessible article in the May-June 2002 issue of American Scientist, but that’s not as readily available. (Which says something about scientific communication, eh?) The Purves lab also has a very cool interactive website.

  7. This is, of course, nothing new to artists, who consciously use these types of optical illusions to manipulate the viewer. In terms of color, this was most notably discussed by Josef Albers. Take a look at his Interaction of Color( I was just reading it last night.

    When I took an intro. to painting course (all art history majors at my undergraduate institution were required to take studio courses), we did some of these exercises. It’s actually pretty tricky for some of them to get the colors just right. Very neat stuff, especially since people often scoff when they see Albers-type paintings, but then go “wow” when they realize what’s going on with the colors.

    1. I just got Color and Light by James Gurney, which also covered a lot of the brain’s way of seeing color and shade from an artist’s POV.

  8. A classic illusion. You can also see they are the same by looking through squinted up eyes. this is a technique I often use when drawing to try and work out different tones.

    1. I have only managed to do it by blocking all other squares. Squinting seems to make everything a patchwork (not checker board) of two shades of grey – including the green “cylinder”.

  9. I love stuff like this. I still had to be convinced. I’m not as computer savvy as #2, I did it the old fashioned way, covered everything but squares A and B. It’s amazing——until I covered every surrounding square, they looked different.

  10. I am by no means an expert, but from Jeff Hawkins’ book On Intelligence and Judea Pearl’s work on Belief Propagation in Bayesian Networks I now suspect that most of the neocortex is subject to essentially the same ‘illusions’ that the visual cortex is subject to. At higher levels of abstraction we refer to these as cognitive biases. When people raised with a biblical world view say that they see proof of God everywhere they look, they are quite likely suffering from a cognitive illusion every bit as powerful as this optical illusion.

    1. I like bayesian networks and I like Cosma Shalizi’s work on them. He hasn’t trumpeted it as such, but you can there, as elsewhere, note (actually measure) that it is better to introduce objects (faithful states) that correspond to what the networks model.

      I.e. you start out with “beliefs”, guesses, and you can robustly (in most simple cases, natch) arrive at “reality”, predicted objects. (Provided you use a sensible method.)

  11. I used Microsoft Paint to select part of the “A” square and move it to the “B” square. It’s true! They’re the same color.

  12. whether human senses perceive a real, external reality or whether that reality is somehow “constructed” socially or by our senses.

    It should be obvious, at least from a physics view, that “reality” hasn’t anything to do in that sentence. We never perceive objects directly but observe their interactions with their environment.

    For example, we don’t perceive particles, we don’t even perceive their masses and charges, but we observe how their masses and charges interact with other particles masses and charges.

    Of course all we have are constructs, models and theories, the question is if they are faithful.

    [Unless you go nuts and also question reality – but we can define and test that too, if necessary. However, there is no generally agreed upon definition.]

    Since we know that we can do that, we also know that our senses can do this, at best.

    Illusions are illuminating on the pitfalls of our evolved senses. But it is perhaps more illuminative to start out with perusing the reconstructions that astronomers do to probe data to get images in the first place, and more or less faithful images in the end.

    Then these illusions doesn’t amaze for their sense of “construction” but for how we can get to how things should look… oh, err, how neat; the post gets to that too. 😀

    1. “Of course all we have are constructs, models and theories, the question is if they are faithful.”

      Right. If you’ve ever had a lucid dream, then you know that even our *waking* experience of the world is completely a neural construction, a virtual reality, So as philosopher Thomas Metzinger puts it, while awake we are “dreaming vigorously at the world” but the dream (waking experience) is constrained by sensory input, which keeps the brain’s conscious model more or less faithful. When we’re asleep having dreams (regular or lucid) the model is not thus constrained.

  13. I believe that there is a third position about reality.

    1) External reality exists.
    2) We experience an augmented reality, i.e
    a) We cannot receive data we do not have sensors for (ultraviolet light, ultrasonic sound, the sense of ‘nimh’ (only experienced by bats)
    b) Most of our sensors respond to changes in data, not absolute values
    c) Most data is acted upon unconsciously or ignored because it does not trigger further responses
    d) Data is processed and associated with instinctual or learned responses
    e) We become aware of data that is augmented by this process of stripping away all the unimportant stuff and adding in extra emotional importance.

    How else would evolution work? Our senses are tuned to interpret reality in the fittest way.

    Interestingly I’ve read that the strength of the Müller-Lyer illusion varies between cultures. A common reality (you can measure the two lines with a ruler) but a learned response.

    So yes, there is a common reality but our interpretation of it is augmented and individual.

  14. Which, by the way, could explain how the god struck and the god free can live in the same world but ‘see’ it completely differently. Or left wing vs right wing, or even women vs men. Same reality, different emotional response, therefore different internal reality.

  15. Another lovely illusion is an auditory one:

    A pair of broken-up scales is divided between your two ears. When you hear them both (using headphones to isolate your ears) your brain combines the two sounds making two connected scales instead of two broken ones. Each ear is totally separate so the illusion is one of processing.

    I connect this to our incredible aural ability to follow threads of conversation despite other noise (even other talking) going on at the same time. An ability that it is easy to imagine a strong evolutionary pressure for!

    A friend of mine with a hearing disability doesn’t have this ability. One-on-one (with a hearing aid) she can follow perfectly, but once there is extraneous noise (especially other talking) she has a very hard time following. I’ll bet this has to do with the very small frequency response of the tiny speaker in a hearing aid–it makes distinguishing variants in timbre (which depend on high frequency partials).

  16. Incidentally, this sort of thing defeats Plantinga’s stupid ideas about evolution and perception. If Plantinga were right, our senses would be completely reliable because God made them that way; but our senses aren’t completely reliable, so Plantinga is wrong.

  17. ‘Beauty is in the eye of the beholder.’

    So we can’t trust the beholder anymore !
    I hope beauty stays.

  18. I wonder if there is something else going on here – more cultural than biological. We all know we are looking at a chess board, on which there are only two colours and they alternate; we seen these our whole lives. Squares A and B are so close together, it is obvious what the colours should be, and so we interpret the scene that way; A and B should be different, so thats what we see.
    The cylinder and shadow partially obscure the fact that this isn’t a real chessboard. One should be able to redo this with a pattern that follows the real chessboard pattern, and then see how strong the illusion is.

    1. The local “contrast correction” (for grey-scale as well as color) works for other types of images, as well. It doesn’t require knowledge of the pictured object, nor simulated shadowing, as in the present example.

      Check out the recent special issue of Scientific American (edited by the psychophysicists Martinez-Conde and Macknik) on more details about illusions like this one. The color adaptation illusions are really neat – way more complicated than I ever suspected.

  19. Well understands the way that works and if so then it would hold true for the other squares, the perception being that once you see it with the one then the rest follow. Interesting.

  20. I see this was posted a whole year ago now. Is anyone still paying any attention to it because I have some thoughts and questions? Like how would anyone here feel if I provided simple, total, unequivocal, final proof that A and B are NOT the same colour?

  21. Here’s a trick that worked for me. Squint your eyes until the picture looks like its starting to flicker (you may have to squint your eyes a little more or a little less) and you should be able to see that square A and square B are exactly the same shade of gray.You might also notice that the light looking square that’s diagonally south-west of square B (the top of the picture being north) is also the same shade of gray. I checked it in MS Paint, and I was right.

    1. Well, here’s the trick that worked for me. You do a bit of ordinary logic:
      > A checker board is by definition made of alternating black and white squares
      > This means that if you count an odd number of squares away from a black square you get a white and if you count an even number of squares you get another black.
      > A is black and B is 3 squares away, therefore B is white.

      Neehee. I hope that doesn’t mess with your mind too much.

    2. Although the last comment was posted almost 2 years ago, I’d like to give a simple proof that’d work for everybody regardless of their specialties. Here we go:

      Save the image, then open it in simple image editor like Paint. After that, delete all checks except these a and b checks. You will see that your perception will change.

Leave a Reply