The Anthropomorphization of the Human Brain

I recently came across David Eagleman’s theory that dreaming is a maintenance function for the visual cortex.

Keeping our eyes closed while we sleep means that our visual cortex receives little input for a third of the day. And cortical areas that receive no input tend to lose territory due to neuroplasticity. Think about how, in blind people, the visual cortex can be reallocated to process other senses. Dreaming, Eagleman suggests, is the process of sending internally generated inputs to the visual cortex to keep it active and prevent its reallocation.

When I first read the abstract, I was skeptical. But the paper brings some interesting data that makes the theory more convincing. Anyway, the paper is not the point of today’s discussion. The point is a comment I read about it, from David Galbraith (a must-follow on Twitter). He wrote: “We dream to preserve our sight. I love how this cuts through millennia of bullshit theory like Freud’s.”

I agree. If we see the brain as a machine, the likely function of something that happens while the machine is at rest is some form of standby, calibration, or maintenance. This should be common sense to any engineer. The idea that the function is instead to produce meaning is often downstream from a mistake I like to call “the anthropomorphization of the brain,” which consists in attributing human-like qualities to a system that is not a human.

You know how you may talk about trees being cut down as if they felt pain? That’s anthropomorphization.

You know how you may tell happy birthday to your dog and expect it to understand? That’s anthropomorphization.

You know how you may look at your brain and think that its dreams must serve a consciousness-specific function, even though REM-like states are not uniquely human and appear to be so evolutionarily ancient that even fireflies seem to have them? That’s anthropomorphization, too.

You are taking a biological machine and attributing to it features specific to human consciousness.

It doesn’t matter that the human brain happens to be located inside a human body. It doesn’t matter that the human brain happens to be what produces human consciousness. Anthropomorphization of its functions may still happen, and it frequently does.

We should be wary of explaining features of the human brain by thinking about the brain as a human being, when the same features could be explained by thinking about it as a machine. The latter should usually be our starting point.

The hive mind fallacy

In my 2020 book The Control Heuristic, my treatise on human behavior, I describe the following fallacy, related to anthropomorphization.

When we observe a colony of bees, we are likely to describe it as if it had a single will: a hive mind. We use sentences such as “The colony decided to settle on that tree,” or “The colony decided to attack the bear.”

There is no hive mind, though. Bee colonies do not have a single will. Instead, individual bees make individual decisions. For example, when the time comes to choose a new place to build a nest, scout bees independently explore their surroundings and make dances expressing the location of their favorite spot. Eventually, if other bees observe convincing dances, they inspect the proposed locations, change their own behavior, and sometimes change their own dance. Soon enough, the dances of most of the colony converge on a single choice, at which point the colony moves to the chosen spot.

We have the impression of a hive mind because the individual actions of the bees converged to a single colony outcome, which we then interpret as the hive mind’s decision. Still, there is no such thing as a hive mind of bees communicating telepathically. Each bee acts locally. The colony-level behavior emerges from the interaction of many bees.

We can describe the behavior of a bee colony at the group level. However, to understand it, we need to look at the rules governing the individual behavior of its components.

The same applies to the human brain. We can describe its behavior using sentences such as “it decided” and “it wanted.” However, the brain is not a single agent with a single will. It is a collection of brain regions and systems, each processing partial information and producing partial outputs.

If it seems otherwise, it is because our brain is one organ, so it looks like “our brain made a decision.” But that is a bit like saying that “Coca-Cola decided to raise prices.” Coca-Cola is not a conscious entity. It is a collection of employees, a few of whom made a decision, which we then attribute to the company. We anthropomorphize the company.

The same applies to our brain. It is a collection of regions and systems, each processing information and pushing toward certain actions, which we then attribute to the brain as a whole. But it is not “the brain” making decisions as a whole. It is specific individual systems within the brain taking individual decisions based on their individual knowledge, understanding, and reward functions, which are a subset of those of the human host.

Of course, brain regions collaborate and share information, just like employees within a company do. And of course, some brain regions have outsized influence on the behavior of the whole organism, just like executives have outsized influence on the behavior of a company. Still. If you want to truly understand Coca-Cola’s decisions, you must not stop at “Coca-Cola made this decision.” You must ask which specific people inside Coca-Cola made the decision, based on what information, under what constraints, and according to what incentives.

Similarly, if you want to truly understand the human brain, you must go beyond “the brain made this decision.” You must ask which specific systems within the brain produced the outcome, what information they had, what information they lacked, and what local rules they followed.

“The brain decided” is useful shorthand, but shorthand nevertheless. The mistake is to forget that, and to start looking for a person where there are multiple independent units.

If you enjoyed this essay, you will love my book, The Control Heuristic.