In the debates on free will in ancient philosophy, the first agent-causal philosopher was Aristotle, followed by Epicurus, and the Skeptic Carneades.

"Near the end of his instructive and thoughtful book, The Significance of Free Will, Robert Kane skeptically allows that "maybe theories of agent-causation can be resuscitated. But the burden of proof must be on anyone who would do so" (p. 195). I accept this assessment, and in the remainder of this book, I try to discharge that burden."

Persons and Causes, 2000, p.xiii)

As a property of material, information offers us a "property dualism," rather than a "substance dualism." Information is an emergent phenomenon that distinguishes all biological systems from their subvening lower-level physical and chemical aspects. Information is neither matter nor energy but needs matter for its embodiment and energy for its communication.

O'Connor briefly comments on the role of quantum indeterminism, questioning whether it is operating in neural structures. If it were, that would make chance the direct cause of human actions which would not allow moral responsibility.

I have just argued that the emergence of phenomenal consciousness is a good bet. The agency theorist is committed (given the substance monism that the Causal Unity Thesis strongly suggests) to the emergence of a very different sort of property altogether. Instead of producing certain effects in the appropriate circumstances itself, of necessity, this property enables the individual that has it in a certain range of circumstances to freely and directly bring about (or not bring about) any of a range of effects. This further commitment leaves the theory's proponent open to a special objection, not applicable to emergentist claims generally. given the unique nature of the type of property the theory postulates, it is doubtful whether it could emerge from other natural properties. It will be claimed that this property would require a very different kind of substance than material substances, as is posited by Cartesian dualism. (It is noteworthy that many philosophers who discuss the agency theory seem to simply assume that its adherents are dualists.²⁴)

This argument also does not bear well under scrutiny. Given that there is nothing inconsistent about the emergence of an "ordinary" causal property, able to causally influence the environments in which it is instantiated, it is hard to see just why there could not be a variety of emergent property whose novelty consists in enabling its possessor to directly effect changes at will (within a narrowly limited range and in appropriate circumstances). If this possibility claim is difficult to evaluate on a purely abstract level, it is plausible when considered in relation to entities such as ourselves—conscious, intelligent agents, capable of representing diverse, sophisticated plans of action and having appetitive attitudes that are efficacious in bringing about a desired alternative. As with the parallel challenge to emergent consciousness, this properties-can-reproduce-after-their-own-kind-alone objection lacks justification.

Still, taking the agency theory seriously within an emergentist framework raises a whole host of more detailed theoretical problems and issues. The most fundamental of these is determining the precise underlying properties on which an agent-causal capacity depends. Put differently, what types of features — either functionally or intrinsically characterized — constitute a physical system's being a free agent in the technical sense? Conversely, what structural transformations in the human nervous system would result in long-standing (or permanent) loss of the agent-causal capacity generally? This is an empirical matter (one answerable only by neurobiological science) and not in the province of philosophical action theory. Yet even a casual acquaintance with how neurobiologists approach their craft is enough to give an appreciation of the enormous difficulty this most basic issue poses.

Something the philosopher ought to be able to provide some general light on is how consciousness figures into the equation. It is a remarkable feature of most accounts of free will that they give no essential role to conscious awareness. One has the impression that an intelligent automata could conceivably satisfy the conditions set by these accounts—something that is very counterintuitive. That accounts of free will fail to provide an essential role for consciousness is nonetheless not surprising, given that its basic biological functions are presently quite mysterious to most theorists. (Another aspect of the puzzle is that whereas various suggestions have been put forth concerning what specific function or functions consciousness serves, it is readily imaginable that many of these functions can be carried out by automata. One could suppose that consciousness just happens to be the means by which certain functions are carried out, although these could have been carried out by other means, at least in differently structured organisms. Yet it would be highly surprising that this should be so, given its distinctive nature.) The agency theorist can conjecture that a function of biological consciousness, in its specifically human (and probably certain other mammalian) manifestations, is to subserve the very agent-causal capacity I sketched in previous chapters.

My vague proposal begins with the idea that genuine freedom of choice confers a selective advantage. The ability to make novel responses to both old and new types of scenarios in nonrandom fashion is a valuable one, and thus one that is likely to be selected should it appear within a population." It is highly plausible that this self-determining capacity strictly requires conscious awareness.²⁶ This appears to follow from the very way in which active power has been characterized as structured by motivating reasons and as allowing the free formation of executive states of intention in accordance with one of the possible courses of action represented to oneself. (I am tempted to think that one should be able to explicitly demonstrate the absurdity of supposing an agent-causal capacity as being exercised entirely unconsciously. I can't now see how this is to be done, however.)

Another issue an emergentist version of the agency theory faces is a result of the fact that if there are agent-causal events, there is no neat and simple way of dividing them from event-causal ones. It surely must be allowed that some human behavior, even conscious behavior, is directly brought about by event-causal factors. (Not all action is free action.) As we noted, this is likely to be true of behavior governed by unconscious factors and highly routinized actions. Precisely to what extent, then, is an ordinary human's behavior directly regulated by the agent himself, and to what extent is it controlled by microdeterministic processes? More generally, how do event- and agent-causal processes interact? Even when I act freely, I am usually not even trying to control directly the precise degree of muscle contraction, limb trajectory, and so forth. This makes it plausible to hold that our memory system stores action sequences that we simply activate through conscious choice. (It also explains the facility of an experienced performer in carrying out complex movements, such as a sequence of dance steps.²⁷ It may be that these choices are at times even brought about event-causally, while we simply monitor the result and retain the capacity to agent-causally redirect things as need be.

One other empirically based objection is that free will requires the emergence of a degree of indeterminism far beyond that which we have any reason to believe is operative (as a function of quantum indeterminacy) at the complex level of neural structures. My reply is that since an emergent property has, relative to its subvening properties, a unique, nonstructural nature, we have no a priori reason to think it must result in processes that exhibit precisely the same degree of indeterminism as is present in its sustaining lower level processes.²⁸ Still, we are not supposing 'something's coming from nothing,' as many have thought: the presence of any emergent, in the view I have sketched, will be determined by more fundamental features of its possessor. What it does allow is a stable set of processes that give rise, at certain critical junctures, to a somewhat different order of affairs via `top-down' controlling features.

24. For example, Ted Honderich (A Theory of Determinism, Oxford: Oxford University Press, 1988).
25. Cf. Nozick's remarks In Philosophical Explanations, p. 307f
26. Compare Roger Penrose's narrower suggestion that consciousness may be required for nonalgorithmic theoretical judgmeni, in his The Emperor's New Mind (Oxford: Oxford University Press, 1989)
27. This is discussed by the philosopher Austin Farrer, in The Freedom of the Will (London: Adam & Charles Black, 1958), and cognitive scientists Stephen Kosslyn and Olivier Koenig Wet Mind (New York: Free Press, 1992). Cf the remarks of B. Hammel ("The Cognitive Representation of Action: Automatic Integration of Perceived Action Effects," Psychological Research 59 119961, pp. 176-186):

An agent need only think of the intended action effect to trigger the appropriate movement automatically. This implies that there is a cognitive code, a kind of action concept referring to the features of the event that a particular action might produce, and a motor program (i.e., some kind of memory structure capable of controlling the action's performance) connected to this code. Whenever and by whatever means an action effect code, or an action concept comprising several effect codes, is activated (e.g., by imagination or external stimulation), the motor program is activated, too, at least to a certain degree. (p. 176)

28. Indeed, there is no evident reason to deny that the lower level processes that give rise to emergent factors could be 'quasi-deterministic' in the following sense: apart from macroscopic contexts that exhibit the right degree and kind of complexity for the appearance of emergent features, these processes conform to laws that have a nonstatistical character. If this is possible, and it is possible, as I have claimed, that agent-causal powers could be emergent features of human beings, then we might fairly claim to have shown the way to a (decidedly quasi) kind of 'compatibilism'! Freedom of choice could be an emergent (and so causal) consequence of microphysical processes that typically conform to 'deterministic' processes.

(Persons and Causes, 2000, pp.121-123)

Information philosophy corrects the reductionist, materialist, and determinist "scientific picture of the world" that is not supported by quantum physics. Free will is not an illusion. And it is completely consistent with O'Connors' idea of agent-causation.

When we reflect on 'the emerging scientific picture of the world' — that is, when we restrict our attention to those physical processes in nature that have been extensively studied and that to a large degree are coming to be well understood — the picture of free will sketched in this book looks, by comparison, deeply mysterious. Even many who are known as philosophers find it scarcely conceivable that the two pictures can reflect aspects of the very same world. As best as I can determine, the appearance of incompatibility reflects the grip of an imagination-stultifying ideology. The empirical facts may weigh in, ultimately, in favor of a thoroughgoing reductionism, in which case free will is an illusion. At present, however, there is little reason to believe that this is true.

(Persons and Causes, 2000, p125)