Thoughts on Hornets

Or What We Used to Call Hornets, But Maybe Weren’t Really

April 3, 2006

I’ve been thinking about hornets. “Why?” you may ask. Because I’m bored with the little voices and can’t find my Haldol. Anyway, I claim that hornets show that the human race collectively isn’t nearly as smart as it thinks it is. Especially about hornets.

The worrisome thing is that hornets know too much. A hornet has practically no brain, probably a few milligrams or some equally depressing amount. But consider what the dangerous little spike can do.

A hornet can fly, with precisely controlled speed and angle to the ground. It can also hover precisely. This is not easy. Controlling the speed and angle of wings, or whatever the beast controls in (as we say) real time, is not a freshman’s project in programming. Boeing couldn’t do it.

A hornet can walk over broken ground, effortlessly negotiating obstacles; it can do this hanging upside down. It is no simple thing to control six jointed legs. If you think otherwise, talk to a robotics engineer. A hornet can fly up to a tree branch, adjust its angle in the air, and transition from flight to walking. Easily.

A hornet can see. How well it can see and what it can see, I don’t know. I have never been a hornet. Hornets that hunt things can certainly see well enough to find whatever it is that they hunt. This requires integrating the output of the multitudinous ommatidia that constitute its compound eyes into a useful image. Try to figure out how it does it.

Further, it can understand what it is seeing. I’m not sure what I mean by “understand.” Probably I mean the same thing you do when you look at something and know what it is. This is a quite different problem from forming an image. It is easy to get a computer to take a picture, much harder to get it to “know” what is in the picture.

How does a hornet with virtually no brain do it?

Today the language and modes of thought of computing dominate the biological sciences. One speaks of behavior as being genetically “programmed” or “hard-wired,” and of a brain’s “processing power,” of “integrating” information in “real time.” We are perhaps not always aware that we do this. When you think in terms of a particular scheme, you can begin seeing it where it isn’t, begin projecting it onto the world.

When I think of how the control of a hornet’s legs must work (except of course that it doesn’t have to work the way I believe it must), I think in terms of sensors of angle and force, of procedures to calculate this and that. Do hornets do it this way? Maybe not. Scientists as much as other people struggle to escape their preconceptions or, more usually, don’t struggle. Many don’t seem to know that they have preconceptions.

A hornet’s aggregate behavior is not trivial. It can navigate almost infallibly. In a former rural home in Virginia, I watched them set off across a bean field of a hundred and fifty yards, apparently going to the woods on the other side. They came back. In the jungles of South America, dim under thick canopies, with dense undergrowth, I have seen nests hanging. The insects fly though the growth without getting lost.

Hornets know how to build nests—what to chew, how to find it, when to chew it, and how to paste it together to make (depending on the variety of hornet) a smooth hanging grey gourd full of elaborate cells. This begins to be an awful lot of behavior contained in virtually no brain. (Stray thought: What is the unit of behavior per neuron?)

Hornets know how to mate. Mating with a hornet is not to be undertaken casually, and I do not recommend it to the reader without professional instruction. However, hornets seem to do it. In the default computer-think of the sciences (“default,” I say automatically) the explanation might be as follows: The hornet’s pheromone receptors send a medium-priority interrupt to the central nervous system which then branches to its mating procedure. Click, click, click, like mechanized tinker toys.

I wonder. I do not know whether hornets mate while flying, as ants do, but it must be beastly difficult to copulate and fly at the same time. Think in terms of airline pilots and you will see what I mean. In terms of computing, mating is an extraordinarily tricky problem. Both bugs have to “want” to do it, recognize each other, know how to align various body parts without error, and produce the needed physiological responses at the right moment.

I know how I would try to write the program to do these things. I do not know how I would make it work. Especially in bare milligrams of brain. Something curious is going on here, methinks, something that we don’t understand.

Yet further, hornets know how to protect themselves and their nests. (I have stepped on one barefoot. I can assure you that they know how to protect themselves.) This, like so much of their behavior, is not as simple as it might seem. Stinging in itself may be a reflexive spasm, though hornets that paralyze their prey with stings have to know exactly where to sting. (How do they know?) You generally do not want to miss with a tarantula. Their overall defensive behavior is a tad more complex.

They have to decide that they are being threatened. How? I could come up with some function, probably silly or at least inadequate, of apparent size, nearness to nest, velocity, and so on. (“Function.” Back to computers.) That’s a lot of calculation in no brain. In any event, being able to simulate a process in a computer doesn’t mean that the computer is doing it the same way the hornet is. Computers today clock at several gigahertz. A hornet’s barely-brain runs on slow mushy diffusion of chemicals across wet synapses. They are doing something very different.

Further, disturbed hornets look to be angry. They give every indication of being aggressive. Now, it is possible that I am anthropomorphizing. It is also possible that I am not. A thing that appears to be angry may in fact be angry.

Unless I fall into solipsism, I have to assume that if you begin screaming and throwing things at me, you are angry. On equally good evidence, I assume that when my dog behaves playfully or affectionately, she is so. I am not sure why I have to believe that an apparently infuriated hornet isn’t.

Now, add up all hornetary behavior, including a lot we haven’t touched on—communication between hornets, caring for the young, and so on—and ask how much more complex, if at all, is the behavior of whales, who have brains you could sleep in.

Hornets? I think the little monsters know, within the limits of their world, exactly what they are doing. I am not so sanguine about humans.