Choices that cause themselves?

[StillDreaming]

Upon reading Thomas's essay on free will, I was triggered by these sentences:

Unless choices can cause themselves[4], which seems paradoxical, (...)
and

[4] "He chose it because he chose it." This seems to be the best account many indeterminists can offer of why someone chose one action rather than another. Our idea of cause and effect comes from one event preceding and explaining another, and the idea that something could be a causa sui (cause of itself - as God is sometimes taken to be) was mocked by Nietzsche.


Many people claim that a choice is either pre-determined or random, but appearently some people claim there is a third possibility. Could someone elaborate on this self-causing-choice? I'm not sure what it means.

And given current ideas on the big bang and virtual particles, has the concept of self-causing stuff (particles, choices) become more acceptable?

[JossWorShipper]

Okay, I'm no physics major, but I think what you're talking about is the difference between action and reaction.

Action = the choice to do something that does not have to involve a stimulous. i.e. if it can happen it will.

Reaction = the choice to do something that happens in direct or indirect response to a stimulous, often caused by the choices of another.

There is rarely, if ever, just one reaction to a given action.

[StillDreaming]

Hi, JossWorShipper

If I interpret your post correctly, the second one (reaction) seems to correspond to a predetermined choice, since it the reaction(s) are determined by the input stimuli.

The first one (action) seems to correspond with a random choice, I think, since no (input) stimulous is involved. (Note: I consider both external inputs and the internal state at time N-1 as input stimuli for the 'computation' of the state at time N). Now, some seem to claim that the lack of input stimili does not imply randomness, but I fail to see how this decision mechanism is supposed to work.

For example, in a deterministic system, there are a number of input stimuli, like: seeing an ice cream [external stimulous] and observing hunger [internal]. Based on these parameters the decision mechanism decides to purchase the icecream.

A random decision system on the other hand, could just generate a random decision to purchase an icecream, regardless of any input stimulous. Perhaps the randomness could be somewhat constrained by input parameters, for example that the random generator is only enabled when both input stimuli of the previous example are available.

But how would this work for the third possibility (assuming it is indeed a possibility...). The output is neither a random value, nor a function of the input [+ internal state]. What other possibilities are there? The only alternative I see would be some algorithm that generates non-random outputs totally unrelated to the input+internal state, but this doesn't seem to be what the proponents of the third option are thinking of. In fact, within the context of free will this is equivalent to the random generator - except that it generates a non-random sequence...

Hm, after re-reading:
The output is either (partially) a function of the input, or it isn't. If it isn't, it's either predictable (in principle) or it isn't. In the former case the decision mechanism is an algorithm like I just described, in the latter case the decision mechanism is a random generator.
If the output is partially a function of the input, than the part that is a function of the input is deterministic, while the part that isn't can be treated the same way as above.

This seems to be an exhaustive list of all possibilities, and there is no room for the third possibility. So, is it a myth?