moving finger said:OR (B) the RNG generates different alternatives on the second “run”, in which case the agent (still operating determinsitically) might make a choice which is different to the choice that it made on the first run. In other words, if we could re-play the agent’s moment of choice with all of the conditions exactly as they were before EXCEPT that the alternatives for consideration are different, then the agent will not necessarily make the same choice as it did before. This (the agent’s choice) again is a completely deterministic scenario and is again completely compatible with determinism (ie re-play with different starting conditions and one may obtain a different result).
No, this isn't completely deterministic , because determinism requires
a rigid chain of cause and effect going back to the year dot. One part
of the process, the selection from options may be deterministic, but
the other part, the generation of options to be selected from, isn't.
D-ism doesn't mean that cause cause effects every now and then,
it means everyhting happens with iron necessity and no exceptions.
The only difference between re-play (A) and re-play (B) is that in (A) the conditions are indeed set to the way they were the first time round, whereas in (B) the conditions (at the moment of choice of the agent) are not the same as they were before. THIS FACT ALONE (and not any supposed “free will” on the part of the agent) is the source of the agent’s ability to make different choices in each run.
But they are different becuase of indeterminism in the chain
of causes leading up to that moment, and in my naturalistic
account of FW, that indeterminism is one of the things that constitutes
FW. You seem to be assuming that FW is supernatural or nothing;
I am not making that assumption.
In fact, we do not need the RNG in the proposed mechanism to be ontically indeterministic. It need only be an RNG in the sense of a computer software RNG, which operates to generate epistemically random, but ontically deterministic, numbers. What matters in the RA mechansim (the “apparent source of free will”) is ONLY that the agent is provided with DIFFERENT ALTERNATIVES in each re-play (this will ensure that the agent will not necessarily make the same choice in each re-play, scenario B above), and NOT that these alternatives are generated by a genuinely (ontically) indeterministic process.
Pseudo-random numbers (which are really deterministic)
may be used in computers, and any indeterminism the brain
calls on might be only pseudo-random. But it does not have
to be, and if we assume it is not, we can explain realistically
why we have the sense of being able to have done otherwise.
People sometimes try to explain this as an 'illusion', but
it do not make it clear why we would have that particular illusion.
I could quite easily \"build\" such models of \"free-will\" agents using computer software, incorporating an RNG to \"generate\" apparently random alternatives for my deterministic software agent to consider, and from which to choose. Since I am generating the computer agent\'s alternatives randomly (thus ensuring that it\'s choice need not be the same each time) does that mean my computer agent now has \"free will\", where it had no \"free will\" before (prior to me introducing the RNG)? I think everyone would agree that this notion is very silly. And does it make any difference if the RNG is genuinely random (ontically indeterministic), or whether it simply appears to be random (epistemically indeterminable)? No, of course not. It does not matter what we do with the RNG, we cannot use indeterminsim to \"endow\" the Libertarian version of free will onto an otherwise deterministic machine
Naturalists think it is not impossible to artificially duplicate human
mentality, which would have to include human volition, since there
is not 'ghost' in the human machine. You are levelling down, saying huamns have
no FW and computers don't either. I am levelling up, saying humans have FW and appropriate computers could have it as well. It all depends on what
you mean by FW. The contentious issue, vis a vis determinism, is the
ability to have doen otherwise, and that is explainable naturalistically in an indeterministic universe.
I think one will find that if one models the above RA mechanism and examines it rationally and logically, looking at the possible sequences generated, then one will find that the introduction of the RNG prior to the moment of choice acts in much the same way as introducing the RNG after the moment of choice. In both cases, there is a point at which a deterministic choice is made by the agent based on alternatives available, but in both cases the final result is in fact random. This is not free will. This is simply a random-choice-making mechanism.
No it isn't the same. Intorducing randomness after choice removes 'ownership'.
The hypothetical AI wouldn't be able to explain why it did as it did.