A strong conception of free will is not compatible with either predetermined or random choices because in neither case do we decide which alternative to actualize from among many that might have been selected.
Criterial causation gets around the
causa sui argument against both
mental causation and
free will by having neurons alter the physical grounds, not of present mental events, but of future mental events.
Self-causation only applies to changing the physical basis of making a
present decision that is realized in or
supervenes on that very same physical basis.
Self-causation does not apply to changing the physical basis of making a future decision. While there can obviously never be a self-caused event, criteria can be set up in advance, such that when they are met, an action automatically follows; this is an action that we will have willed to take place by virtue of having set up those particular criteria in advance. At the moment those criteria are satisfied at some unknown point in the future, leading to some action or choice, those criteria cannot be changed, but because criteria can be changed in advance, we are free to determine how we will behave within certain limits in the near future. Criterial causation therefore offers a path toward free will where a brain can determine how it will behave given particular types of future input. This can be milliseconds in the future or, in some cases, even years away.
Assuming
indeterminism, criterial outcome is an outcome that meets certain preset criteria, but what that outcome will be is not foreseeable, and had we run the sequence of events over from the same initial conditions, with the same criteria, we may have ended up with a different outcome, because of
noise in the system.
Criterial causality therefore leaves room for non-illusory choice that is a middle path between the extremes of (a)
determinism, where there is no ability to choose freely in the strong sense because there is never the possibility of an alternative action, and (b) criteria-less indeterminism, where arbitrary
choices follow from randomness rather than from criteria one sets up oneself.
Free will skeptics might counter that the setting up of any set of criteria to be met by future inputs is itself determined by preexisting sets of criteria that have been met. This is in fact correct. The key point is that criteria will be met in unpredictable ways if there is inherent variability or noise in inputs, such as can be introduced by the randomness inherent in neurotransmitter molecules crossing the synapse. Just because new criteria are set up by a nervous system in a manner dictated by the satisfaction of preexisting criteria does not mean that either the future or present criteria will be met in a predetermined manner. Moreover, because our neurons set criteria for the firing of other neurons in response to their future input, the choices realized in the satisfying of those criteria are our own choices. Ontological indeterminism and neuronal criterial causation permits a physical causal basis for a strong free will.