Psychohistory and Determinism

Foundation, by Asimov

I recently read the novel Foundation, by Isaac Asimov. It was quite enjoyable and made me think about many things, so I thought I’d write a bit about it. In the Foundation universe, a mathematician named Hari Seldon develops a field of mathematics named psychohistory. This field is to contemporary social science as modern science is to ancient science; that is, not only does it help us explain the current state of human affairs, but allows us to make accurate predictions about the future of humanity. Using this science Seldon predicts the fall of the Galactic Empire and 30 000 years of obscurity after the fall, just like the Dark Ages followed the fall of the Roman Empire. He is unable to prevent the fall of the Empire; but to reduce the length of the Dark Ages to a mere millennium, he creates a Foundation to preserve the totality of human knowledge so that civilisation does not need to be rebuilt from scratch. That is the basic premise, the rest of the plot is pretty much a direct consequence of it.

Although the rest of the plot was quite entertaining, the concept of psychohistory was what most attracted my attention, so I will be focusing on that. The science was incapable of making predictions about a single person because their actions exhibit a degree of randomness, but the more people it took into account the more accurate were its predictions (much like in quantum mechanics a single particle exhibits some randomness, but as the number of particles approaches infinity, the amount of randomness approaches zero).

Note that this implies that, unless a more advanced science can be developed that is capable of predicting the behaviour of a single person, humanity as a whole behaves deterministically, while the actions of any single person are indeterministic. Allow me to break this down a bit. The way I see it, there are three things that can determine the behaviour of any single person, all external conditions being equal: reason, emotion, and randomness. To illustrate what I mean, imagine I am happily walking from town A to town B, and on the middle of my trek, I reach a fork on the road:

The road of forking paths

In my view, I have three ways to inform my decision. For the sake of argument, assume I can only use one of them at the time, so that in Case 1, I will be completely rational, in Case 2 I will be completely emotional, and in Case 3 I will be completely random.

  1. Reason: “I remember towns A and B are on the same parallel. I’ve been walking roughly North-East this whole trip, so it is most likely that the fork that leads East will bring me closer to B than the one that continues North-East. Let’s try that one first.”
  2. Emotion: “North-East is my favourite direction. Let’s just continue walking North-East, that would make me happiest.”
  3. Random: “Umm, a decision, let me toss this perfectly random coin I always carry with me. Heads I go North-East, tails I go East. Oh, it’s tails, let’s go East!”

Let us now analyse this situation using the many-worlds interpretation of quantum mechanics. In every universe, given the same situation and with the same information, the rational me should always make the same decision. If the decision was based entirely on reason, even if the answer was actually wrong, it seems to me that I would have arrived at the same conclusion in every possible universe. A computer that is fed all of my memory, and knows how every brain process of mine works, should be able to determine which path rational me will choose. Thus reason is completely deterministic.

Similarly, emotional me’s favourite direction will be North-East in every universe with the same conditions, and thus it seems clear that I will always choose North-East in the same situation. Given that emotion is, at its core, a combination of chemical reactions and brain processes, a computer that knows all about me should be able to determine, knowing that emotion is the sole leading principle in my life, which path I will choose. Thus emotion is also completely deterministic.

Random me, however, makes his decision based on a toss of his perfectly random coin. If the toss of the coin is truly random, it is evident that no matter how powerful the computer, it should not be able to determine what his actions are. Thus randomness is not deterministic.

If this logic is correct, the only way in which the actions of a single person cannot be determined by an all-knowing computer is if at least some of this person’s actions are random, or if randomness plays some part in their decisions. Randomness is the only way to escape determinism.

But is a person truly capable of randomness? In the real world, even the toss of a coin is deterministic. A computer given the exact initial conditions of the toss, all possible data on air flow, distance from the floor, etc. will be able to exactly determine the outcome of the toss. See for example the Eudaemons, a group of physicists that used a computer to win at roulette.

If not, it would follow that each person’s actions are actually completely deterministic. Note that this does not yet rule out Free Will, since according to Compatibilism it is possible that humans have free will even if their actions are completely deterministic.

This post is getting a bit long, so I will be discussing whether human behaviour can be truly random and, in either case, whether we can have free will in a future post. Hopefully it will be tomorrow, but I don’t yet have internet in my apartment, so it may have to wait a bit longer.