The Free Will of Ebenezer Scrooge

[Philosophy Now]

Richard Kamber considers the possibility of changing destiny.

https://philosophynow.org/issues/111/Th ... er_Scrooge

[waifnstray]

A champion of the consequence argument might protest that this is because the experiment imposes a self-defeating constraint on the computer – namely, that the computer must tell you what it predicts you are going do before you do it.

To dig a little deeper - a question to consider is this: does the computer model itself or not? Does its simulation of the universe include a full simulation of itself? I think we must assume that it does not, otherwise we will most likely find ourselves in recursive modelling and a requirement for infinite data.

In this case, it is possible to imagine that the computer's predictions of the universe outside itself, treated as a closed system, are always accurate. When the computer interacts with the outside universe, however, it breaks the closed nature of the system it has modeled. At the point it decides to do this, it must re-run its simulation based on the perturbation of the modeled system that it is about to make. Having concluded that you are about to move your hand down, the computer can accurately predict your response to it telling you this - whether you will move your hand up, or down, or do neither and go to the pub. At this point, what it is doing in performing this interaction is not really telling you its prediction, but telling you what its prediction would have been had it not informed you of its prediction. It also has a new prediction for what you will actually do, after it delivers its prediction, which may be different. If you oblige it to give you this second prediction, it can do so, but having again violated the closed system it has modeled, it must internally run a third prediction of how you will respond to that new stimulus. So yes, this represents a self-defeating constraint. But the underlying reason for the constraint also applies to the "us" in the sentence:

Determinism makes it possible in principle for us to infer the outcomes of causal chains, and so prevent some of those outcomes from happening in the same way as we can with the supercomputer.

We are not, in fact, challenging determinism by changing the outcomes we predict; we are simply demonstrating that our models of the universe were incomplete, because they did not contain themselves.

To put it in another light: if we have a second supercomputer watching the first one and the experimenter interacting, and modeling them both, it would know exactly what the outcome would be.

It is this inability to model one's own predictive system, and the fact that we are therefore always forced to run our internal simulations on an incomplete model of the universe, that creates an illusion of free will.