02 June 2008

Optical Illusions

A colleague of mine at work emailed a link to a few fascinating optical illusions.

I think the pleasure we get from observing optical illusions is precisely the same as the pleasure of observing magic tricks.  Upon seeing a rabbit pulled from a hat or an assistant in a box sawed in half, nobody (at least no rational person) thinks that magic is real.  Obviously, the whole fun is derived from the implicit knowledge that nature is immutable, that things are what they are, and that there is a single, external reality to be grasped by opening one’s eyes.  This makes it supremely entertaining to see a talented magician make it appear that the laws of the universe were violated.  If anyone thought for an instant that empty hats really could produce rabbits, magic shows would hold no charms. 

Thus, the aspects I find most interesting about illusions are the philosophical implications.  A common, and quite incorrect, interpretation of illusions is that they reveal an alleged fallibility of the senses, and this in turn indicates that we can never really know the universe “as it really is.”  

Upon more careful thought, though, we can see that the exact opposite is true.  The senses are in fact infallible.  And optical illusions themselves are marvelous demonstrations of this fact.

For a simple illustration, let’s take the first illusion, the checkerboard with the cylinder, as an example.  Looking at the picture, it is hard to believe that squares A and B could be the same shade of gray.  It seems impossible!  So, how would we determine if they are the same shade or not?  Well, we might print out the picture, cut out squares A and B, and put them side by side on a neutral background.  And sure enough, if we do this, it turns out that they are exactly the same shade.  How do we know?  Because we see it with our eyes.  Why are we sure that this is a sufficient criteria?  Because the senses are infallible.  

Indeed, living creatures have no other connection to the universe apart from the senses.  At the most basic level, there is no other way to obtain the raw material that grounds our knowledge.  Also, there is no basis of denying that the raw material of the senses reflects reality.  It is ostensively so.

The subjectivist (a la Descartes) would claim here that since his senses have been initially fooled by this illusion, the sense organs are not to be trusted; perhaps all reality is an illusion, simply a dream, and perceptions are nothing more than the probings of some evil demon in our brains.  This is nonsense of course, and such arbitrary postulates disqualify the subjectivist from even participating in the conversation (or any conversation, for that matter).  He should be left to mutter, “I think not,” and disappear in a puff of his own logic, as the joke goes.  The important point to notice, though, is that in order for the subjectivist to even attempt to refute the validity of the senses, he must rely on the very validity that he denies (or at the very least, he must rely upon having a louder voice than his opponent).

But the problem remains: if the senses cannot be wrong, how do we objectively account for the illusion?  An important distinction to make is the difference between sensation and perception.  Illusions do not fool the senses; the sense organs can only “report” the stimulus that characterizes their function.  However, humans (and other animals) have a much more sophisticated faculty of perception, which involves memory and the automatized responses based on past experience.  

Illusions play upon this higher level perceptual faculty.  They work by creating a match (or partial match) to something with which the mind is familiar and letting the mind “fill in the blanks.”  In most situations, this automatic response enhances the comprehension of one’s surroundings.  That is the whole point - the response is automatic precisely because it is reinforced by experience.  Each time the situation is encountered, the mind strengthens the connections that facilitate the response.  In cases of illusions, however, the brain’s automatic response yields misleading information.  The mind’s attempt to complete the picture fails because the new situation is in some important respect different than similar ones in the past.  

Notice, however, that in every case, the way - the only way - to discover and resolve illusions is by collecting more sense data.  Furthermore, the way in which we humans (and possibly some of the higher animals) learn to do this is itself an illustration of the validity of the senses.  The general approach, whether we identify it as such or not, is to isolate the entity that we are examining from automatized responses.  For instance, if we have doubts about an object we are looking at, we might blink, move around to look from another angle, pick up the object to turn it around, shine a light on it, smell it, etc.  In the case of the checkerboard illusion, we cut out the shapes and placed them side by side on a neutral background.  In the “shape distortion” illusion, we might use a ruler or straightedge to determine if the lines are parallel, and so forth. 

All illusions can be examined in this light.  Let’s look at a few popular examples.  Far from demonstrating that sense organs distort reality, these examples demonstrate the utter fidelity of the senses.  

*  A mirage is an illusion that occurs in certain hot atmospheric conditions such as may be found in a desert.  If one gazes across a flat plane at the horizon, one may see what looks to be a shimmering sheet of water in the distance.  Of course, there is not actually water there - it is simply light reflected from the sky because of a very shallow angle of incidence.  Is this a failure of one’s sense of sight?  Of course not.  The light is actually reflecting from the distant ground and arriving at the retina.  The light is there, and is detected correctly by one’s eyesight.

*  A perfectly straight stick immersed in the water appears to be bent.  Is this a distortion created by the senses?  No.  Eyeballs (along with the associated nerve systems) are not “straight stick” detectors; they are light detectors.  The light is actually bent by the differences in refraction indices between air and water.  It is ludicrous to expect the eyes themselves to somehow correct for the bend.  The information arriving at the eyes is real.

*  If one stares at a red triangle for several seconds, then looks quickly at a blank white wall, one will see a faint green triangle that gradually diminishes.  Has the sense of sight created a ghost?  No.  Like all physical processes, the retina’s response to light has certain finite response times.  In the case of a decaying response to a step removal of an intense color, the time constant is apparently several seconds. 

Obviously, these examples could go on and on.  The point is that the senses are not instruments of omniscience.  They are components of living creatures that interact with matter in a manner that permits the organism to detect its surroundings.  They do not manipulate reality; they detect reality, and they do so in a particular way.  It is utter nonsense to disqualify a faculty precisely because it has a particular identity, yet that is exactly what the subjectivists do.

Here’s an interesting thought that occurred to me.  There is no way to test it, unfortunately, but I submit that if you held a picture of the checkerboard illusion in front of a newborn baby, to him squares A and B would appear to be the same shade.  He would not be “fooled” by the optical illusion.  Why?  Because he has had no experience with objects yet.  He does not know what a cylinder is.  He does not know it can cast a shadow on a checkerboard.  He does not see the checkerboard as a foreshortened array of squares.  All he sees are splotches of color.  His brain cannot possibly “correct” for the shading of square B. 

I’ll conclude with one final observation.  I’ve been referring to the shaded shapes in the checkerboard illusion as “squares” A and B.  Where are the squares?  I didn’t say parallelograms, though of course they are drawn as parallelograms on the page, not squares.  Obviously, this is another instance of the amazing power of our perceptual faculties automatically operating upon sense data.  Our mind recognizes the fact that a checkerboard of alternating dark and light squares held at an angle would look much like the drawing.   In no sense is this a “distortion” of reality.  On the contrary, this capability has enormous cognitive benefits.


Almost immediately after I posted this, LB informed me that Gus and Rational Jenn had recently posted on optical illusions as well.  


C. August said...

Great post! Yes, other bloggers may have mentioned these illusions, but your post examined them in a way that never occurred to me before and conveyed some unique and valuable insights.

...incorrect, interpretation of illusions is that they reveal an alleged fallibility of the senses...
Upon more careful thought, though, we can see that the exact opposite is true. The senses are in fact infallible. And optical illusions themselves are marvelous demonstrations of this fact.

Just awesome, and it's so obvious now after the arguments you present.

I wondered myself whether placing the checkerboard&cylinder image in front of a newborn would show that they could see the squares as the same shade (assuming this experiment were even possible). So I decided to open the image in photoshop and play with it a bit.

The squares A & B do in fact have the same color value -- #787878 (no surprise). Then I wondered if taking out the cylinder and rotating the image would make it easier for us adults to see. No luck, for me at least. (I'd post the image here, but blogger won't let me link to it. I'll email it to you if you're interested.)

The gradient shading in the adjacent squares makes it impossible for me to perceive them "as they are", even when the idea of "There is a big grain silo casting a shadow on the ground" is gone.

Thus, I don't think even the mythical newborn with no visual experience (who could still answer our experimental questions) would be able to perceive that A & B are the same color without more investigation. Something having to do with the gradients in the squares surrounding A & B trumps the idea of the cylinder casting a shadow.

Still, this doesn't disprove anything you said. Look at how much extra sense perception and investigation I had to collect in order to get here.

As I said, great article. This is one of those invaluable nuggets of wisdom that I know I will be able to use one day. Thanks!

Stephen Bourque said...

Ahh! Of course! Now that you mention it, it is obviously the graded shadows that artificially lighten square B, not the presence of the cylinder. Very good point.

M. Minroad said...

Excellent philosophical concretization. I would have to say, however, that the mythical newborn would probably judge A & B the same as an adult. Just as the perception of the color green isn't learned, but hard wired, I believe the receptors and parts of the eye and brain that deal with shading and contrast are also set genetically. However, this doesn't detract from the main philosophical point.

Nice post.

Stephen Bourque said...

Thank you for the comment, M. Minroad.

Yes, I agree with you that I am mistaken on that point, and all I can say in my defense is that it was late at night when I added it and I was really tired! It would not have passed a more careful edit.

The funny thing is that it requires only the simplest test to show that the illusion has nothing to do with the presence of the cylinder and everything to do with the gradation of shades. One simply has to hold one's hand over the cylinder to see that the illusion still holds. So, the child's learning about shadows is irrelevant.

Optical Illusion said...

Hello, I'am George. Visit my website, if you want to learn Magic Tricks Revealed. All tricks are video explained, so you can learn very easy. Thank you and have a great day.

Klaus Nordby said...

Good read, thanks! I probably agree about 97% with you.

I have a great interest in optical illusions, and have in fact made a whole webpage with many Flash-animated examples. I even have a few twists of my own I've not seen in other places. Peruse it here, if you care:


Stephen Bourque said...

Thanks for the comment, Klaus. To be frank, I think even I agree with about 97% of me, too!

What I mean by that joke is that this is an old post. I think it is correct in its essentials--namely, that perception is the infallible means by which we obtain all the raw material for conceptual operations. (I've already noted that the penultimate paragraph is dead wrong.) In fact, having re-read it, I'm surprised at how much I grasped in lieu of the thinking I have done over the last couple of years.

Since I wrote this post, I have had the benefit of following discussions on Harry Binswanger's private email list (with which you are obviously familiar), and am also working my way through his forthcoming book, for which I am indebted to him for his brilliant identifications pertaining to perception. Also, I am greatly intrigued by the work of J.J. Gibson and am just starting his book, The Ecological Approach to Visual Perception. (I see from your web site that you are an admirer of Gibson.)

So, I have been fine tuning my thinking on the matter. My understanding of the essence of perception as the basis of knowledge, though, as described by Ayn Rand in Introduction to Objectivist Epistemology, has not changed--and indeed cannot change, given the facts of reality.

Thanks for the link to your web site!

Klaus Nordby said...

Cool your're taking Harry's class AND reading Gibson! Meeee toooo. :-)

As you see in Harry's MS, he's using the term "inerrant" to refer to the senses, not "infallible" (as your old self uses) -- which I think is a splendid term. I have used "veridical" but that's not ideal, so I'm embracing perceptual inerrancy now.

Hey, feel free to email me privately if you'd like to discuss Gibson and perception! My email is on my homepage.

Froilan Vincent said...

I really love this post, Stephen! Great blog!