The Excitable Membrane

Summary of The Principles, Ch2, part 1.

A neuroscientist friend of mine once said “trees don’t need a retina”, and it stuck with me. It was a pithy way of stating the idea that the senses are only important in so far as they provide data to act upon. If you’re literally planted in place, and your repertoire of movements includes swaying in the wind, and perhaps slowly arching toward the sun over hours or days, you don’t need a visual system that streams on the timescale of seconds. In fact, what torture would it be if you were a seeing tree? “Shit! Here comes the lumberjack!… oh well….“. Nature isn’t so cruel.

James starts chapter 2, on “The functions of the brain,” in a similar key, if a bit more colorfully. If you hack away at a tree with an axe (a ‘stimulus’), James says, its ‘response’ is just the energy from your axe swing passively propagating along the trunk and branches. Its leaves “murmur as peacefully as ever in the wind.” If you hack away at your friend’s foot though, the story is obviously very different. In this case, James tells us, the response involves a whole series of coordinated movements of defense, escape, and agony. The stimulus doesn’t just lead to a response, but a performance. Bottled up things are released, commands are routed to the appropriate muscles, choices are made, actions are decided on and executed.

Clearly our behavioral acts can vary in their complexity and their causes, and it’s reasonable to suppose we’d learn something by trying to sort them into different types. You could split them up by some random criterion like “behaviors involving the arms only” vs. “behaviors not involving the arms only”, but we want to do better than that. We want to make distinctions that cut deep, and that carve along the same lines as nature itself. For James, the useful distinctions to make between movements pertain to their relative automaticity, and their goals.

He gets us to think about this by asking us to imagine being at a train station. The conductor yells “all aboard!” and we start desperately running because we realize we might miss it. In our desperation we trip, and our arms brace us for a fall. While down, a small cinder hits our eye, and we blink and start tearing up. So, three acts in quick succession: running, bracing for a fall, and blinking. What are the most important differences between them?

For starters, James notes that all the acts are ‘appropriate’ (his original term, which I find kind of weird). In other words they serve a goal that serves the organism, and an observer of even very modest intelligence could answer the questions: why did you 1) run; 2) put your arms out when you fell; and 3) blink? It’s important to note that James isn’t committing to any specific metaphysical position here. He isn’t proposing some kind of spooky teleology where there’s a free-floating intelligence cooking up a means to satisfy every end. He’s just saying that moving to serve a goal is baked into our bodies in a way that it isn’t for plants.

In James’ example, the running, bracing for a fall, and blinking differ in the extent to which they’re volitionally controlled. We have no choice about the blinking, which is a simple reflex. Bracing for the fall is more complex, and is what James refers to as a ‘semi-reflex’. It’s instinctual, but could still modified or suppressed. Running in response to someone saying “all aboard” is “purely the result of education, and is preceded by a consciousness of the purpose to be attained and a distinct mandate of the will.”

Everything here aligns pretty squarely with intuition, so I won’t beat this to death. One last interesting point, though, is that a truly naive (say, alien or machine) observer might have a hard time differentiating between reflexes and volitional acts. Our alien/machine, on seeing a person run in response to a train call, wouldn’t necessarily feel the need to psychologically “inject” an accompanying sense of consciousness or willfulness into the subject being observed. And in the other direction, if our naive observer were told that “like.. sometimes there’s like this other thing you can’t see — conscious experience — that accompanies certain movements”, they wouldn’t have a great reason to suppose that this doesn’t hold for the blinking reflex. Our intuition that volitional vs. non-volitional is an interesting way to split up movements comes from us smuggling in a lot of our direct experience into our assessment. It would be cool to see how long it takes a neural network to classify actions as ‘volitional’ vs. not. Would they ever arrive at this spontaneously, or would they need a huge training set? I know in the NLP world classification of intent is a thing, but I don’t know if its been applied to videos of people behaving. LMK in the comments if you know of work like this because I’m curious.

For completeness, James details the two extreme positions one could take in relating action and consciousness. If we take at face value his “appropriateness” criterion that a mark of the mind is seeing a behavior suited to a goal, we can either go ‘full embrace’, or ‘full denial’. Full embracers argue that there might be little shadings of feeling that accompany even the simplest motor acts, like reflexes. They may not be reportable by us, but they’re happening somewhere, and have some kind of causal role to play in simple behavior. Full deniers go in the opposite direction, and say that feelings don’t play any causal role in behavior, even in the case of acts that feel willed and spontaneous.