The demo on this page attempts to show that the concept of pleasure and pain is objectively explainable and definable without deferring to the supernatural or spiritual. In other words, pleasure and pain are simply the names we use to describe a relatively simple process that occurs in our brains -- and thus can happen in other computational devices as well, including those that are man made.

I think it is odd that I have never seen the following explicitly stated (although many kind of dance around it): Pleasure is the reinforcement of previously taken decision paths, making them more likely to be followed in the future. Pain is the suppression of such paths. Period.

Click on a push button. A signal will be sent out through the "neurons", that may (or may not) hit one of the light bulbs, and light it up for a bit. The path followed is somewhat random, but it tends to follow the larger neurons. The neurons will be "warm" for a while after use, as indicated by their color.

If you click the pleasure (plus) button, the size (and therefore strength) of the warm neurons will be increased. If you click the pain button*, the size of warm neurons will be decreased. This way you can train the "brain" to light bulbs of your choosing in response to clicking on a particular switch. To be realistic, you should only reward or punish the brain based on its response (that is, which light bulb is lit), rather than by looking at the neural path it follows. After all, you can't train a dog by looking at the inside of its brain, only at how the dog responds externally.

* please avoid pressing the pain button more than a few times. It's cruel.

Of course, many will say that what is happening is not actual pleasure and pain. Pleasure may cause such reinforcement, but that is not what pleasure actually is. To address that, consider the following analogous statement:

Sound causes vibrations in the air, which can be detected by the ears.

You probably you see the problem with that statement. The sound doesn't cause the vibrations, the vibrations actually are the sound [1]. Until you accept this, you will be looking forever for this nebulous intermediate entity called "sound," and never find it, no matter how close you look. While our intuition may tell us that the sound is a "thing" that is separate from the vibrations, that doesn't make it true: intuition has a well earned reputation as a big fat dirty liar.

Same thing with pleasure (and pain) -- there is no intermediate entity. And while our brains can usually detect the stimulus that caused the reinforcement (whether it be physical stimulation, or something more cerebral, such as receiving a compliment), that part is not the pleasure -- in fact, the same stimulus might be unpleasant under different circumstances. Or, someone may have surreptuously "dosed us" with Percocet [2], leaving us experiencing pleasure without having any external or internal stimulus to go with it. If and only if we have decision-path reinforcement do we have pleasure.

Another complaint might be that I am ignoring the biology behind pleasure and pain. Certainly, most people's understanding of pleasure and pain is intertwined with its biological implementation. But it doesn't have to be. Trying to understand it by looking too closely at the specifics of the biological implementation is a bit like trying to analyze the Beatles' Sgt. Pepper by looking at the vinyl record under a microscope -- not only will you miss the important part, such as how Sgt. Pepper differs from other musical works (which would be more apparent at a different level of abstraction), but you will be back to square one when Sgt. Pepper comes out on CD.

Finally, people may say I am ignoring the internal, subjective "feeling" of pleasure and pain. Philosophers refer to this slippery concept as "qualia." In my opinion, qualia is easier to wrap one's head around with other, less "emotion based" phenomena, such as color perception. I take on qualia directly (albeit rather wordily) in another essay, Violet.

I am well aware that many will find such a simple explanation to be both counter to intuition, and maybe just plain disturbing. At risk of sounding grandiose, I suggest that the assertion about has much in common with another, that was also initially unintuitive, and is still disturbing to many: Darwins' assertion that natural selection explains the existance of all biology. While mechanisms behind natural selection (and the biology resulting from it) is complex in implementation, the actual concept itself is almost shockingly simple. As is this one. [1]

As an analogy, consider that most people are probably comfortable saying that a digital camera can detect red light. But the pedantic may object, saying that the camera isn't really detecting redness, but is simply detecting radiation that is predominently in the wavelengths between 650-700 nanometers. Only a human brain and visual system can detect redness, they might say. The rest of us would probably respond by saying "ok, well, you're just being difficult."

In the case of digital cameras and color, it is easier to see this as a simple issue of semantics -- i.e. what do we mean by red? The external phenomena, or the stimulation of "red sensitive" cones on our retinas, or the actual internal sensation of redness?[2] Someone who works with cameras or image processing software would probably consider the distinction meaningless...it is easier to just think of them all as being covered by the word red, and be done with it. Anything else just inhibits communication and understanding.

I suspect that as machines with sophisticated learning ability inevitably become more common, people will become comfortable using the the words pleasure and pain to refer to states of machines, in addition to states of humans and animals. When that happens, it will surely impact our views on all of the things that are directly or indirectly tied to pleasure and pain: good, bad, happiness, ethics, value, love, and so on. Not that it will change how we think of them in everyday life, of course.

My approach to this demo is not to try to explain the details of the inner workings of the "pleasure centers" of the human brain -- I think dwelling in that complexity is likely to obscure the forest for the trees. (Remember that Darwin did okay at seeing and presenting the logic of natural selection without knowing the details we know today of DNA and molecular biology) To understand the big picture, I suggest we might do better to concentrate on the higher-level logic behind it, rather than the specific biological implementation. So that is what my demo attempts to do...show an implementation of pleasure and pain in a machine, that is as simple as possible -- while still being pleasure and pain.


footnotes and links