← other writings

Altera: Free Will

Free Will

Nookian minds are deterministic, but the Nooks probably wouldn’t know this. The minds of conscious Nook are likely so complex that they would not be able perfectly predict how they would they themselves or other Nooks would react to a situation. Furthermore, they would be unable to devise an experiment to test whether their minds were deterministic, since they would be unable to establish the same “situation” twice. By “situation,” we mean the complete state relevant to a Nook making a “choice.” In other words, the Nook’s mind would have to have identical state, and the sensory input that they are deciding on would need to be the same. Establishing identical sensory input would not be too difficult, but it would be impossible for them to establish that a Nook’s mind’s state has the same state twice in a row. The state of a Nook’s mind is complex and would likely have a long history. The Nook being experimented on would remember earlier renditions of the experiment. A Nook’s child’s mind would be very similar to the parent, but not exactly the same due to the mutation—thus experimenting on a parent-child pair would also be unfeasible.

Thus, the Nooks are in a similar situation to ourselves with regards to free will and determinism. They may suspect that their world is perfectly deterministic, but they would be unable to prove it. As a result, they may wonder, like we do, whether they have “free will.”

What do we mean by “free will?” A standard definition is: the ability to choose between different possible courses of action. This definition begs the question—what does it mean for a being to choose? At each time-step, the Nook mind is presented with sensory input and outputs an “action” that determines what the Nook will do (nothing, move, eat, or reproduce). If this process is what we mean by choosing, then the Nooks have free will despite being deterministic.

Some would say that, for the Nooks to truly make a “choice”—and, by extension, to have “free will”—a different outcome must be possible, given the same situation.

In the case of Altera, one way to accomplish this would be to add randomness into the algorithm that “runs” a Nook’s decision making process.

The simplest way we could do this is to alter our program so that each Nook’s mind’s state has a slight random perturbation applied to its state just before it makes a choice. This introduces another issue, how we will arrive at this “randomness”? Computers in our world are deterministic. The only way to introduce true randomness into a computer is using randomness from outside the computer (e.g. from mouse movements, or network traffic delays). By introducing true randomness from our universe into the computer so that we can truly randomly perturb the Nook’s mind, we are coupling our universes state to the Nooks state. If our own universe is deterministic, it is entirely possible that the Nook’s choices are still deterministic (in the sense that no other outcome was possible). Of course, the Nook’s can’t tell the difference in either case.

Lets assume that our universe is not deterministic, and that we can randomly perturb the Nook’s mind’s state before they make choices. Now, alternate outcomes would be possible, given the same situation. Do the Nooks now have free will? One may object for a few different reasons, which we will consider in turn.

One may object because this randomness is added to the Nook’s mind from the “outside.” Imagine that we are running each Nook’s mind on a separate computer, and that each computer has its own randomness which it injects into each mind. Clearly where there randomness comes from is not relevant.

Another objection is that we are adding randomness that is outside of the “control” of the Nook, so it is not possible that this randomness could give the Nook free will. Well, what do we mean by “control”? Do we mean that the Nook has the “will” to make it act one way or another? Clearly this is begging the question, for we are debating what free will means in the first place. “No, but really, I don’t know how the mechanism would work, but surely there is some way that the Nook could have control over the outcome while not being deterministic!” When considered in our universe, with its huge complexity and its unknowns, some of us may find this argument convincing. However, when considered in the context of Altera it is clearly silly. There is no mechanism within Altera that would circumvent this problem. To this one may argue that Altera is too simple a world to have free will! In fact, one may argue that, for the same reason, our assumption that consciousness could develop in Altera is also not possible. I will call this the mystical metaphysical argument, because it is arguing that free will is beyond our ability to understand. It is impossible to argue against the mystic—because they can always throw up their hands and say “it just is—it’s impossible for us to understand, its beyond reason, its outside space and time, but it just is. I believe it.”

We can not argue against the mystic, but it is worth noting that, from the outside, a being that has free will via this mystical metaphysical substance would be externally indistinguishable from a being that has the random perturbations. In both cases, they would be unpredictable for us.

Thus, if your mother has free-will, but your father does not, you would not be able to tell the difference.

In conclusion, either:

  1. our actions are deterministic
  2. our actions are not deterministic, because randomness is injected from an internal or external sources
  3. our actions are not deterministic, because of some mystical metaphysical “stuff” that we can not understand, however, this situation would be impossible to distinguish from the previous possibility.

The meaning of free-will in any case feels different from what we intuitively feel it should mean, but ultimately we must recognize that our intuition is wrong. Nooks (and ourselves) do make choices, so in this sense they have free will. These choices are deterministic, but they are so complicated that no Nook can predict them.

There is no reason why this result can not be extended to our universe. We also make choices in that no one (not even ourselves) can predict what we will do. If our world is deterministic, than clearly our choices are also deterministic. If our world is not deterministic, then but our choices may be deter