​Chapter 12

Physical realization holds that mental or agential events occurring in the physical world are realized in physical processes. This account accepts it: the settling occurs in John's brain and body, not in a separate immaterial sphere, and nothing here requires a substance operating from outside the order. Causal closure is the stronger claim that every physical event has a sufficient antecedent physical cause fixing its outcome. That is rejected: physical realization does not entail physical predetermination.

Kim's exclusion argument is straightforward. If a mental event causes a physical event, either the physical event had a sufficient physical cause — in which case the mental cause is redundant, and the world runs without it — or it did not, and the mental event operates as a physical cause without being a physical thing, which restores the familiar difficulties of dualism. The dilemma is sound and it is aimed at a mental cause standing alongside physical consequences.

The free act is not a second cause. It does not enter the body from outside, does not push matter, and adds no energy to the system. The settling and the corresponding neural transition are one event: the completion of a process in the organism and, as an action, the resolution of an open practical question. Kim's dilemma asks which of two causes did the work, and there are not two.

Quantum mechanics provides a precedent. The standard formalism does not assign a unique predictable outcome to an individual measurement, and indeterministic interpretations treat that openness as physical, not epistemic. Everettian and hidden-variable pictures read it otherwise, and the argument does not need to adjudicate between them. Physics does not establish strong closure. The precedent does more than show that physical indeterminacy is possible. An atom whose decay is undetermined does not hold a concealed decay-time; before the decay there is no fact about when, and when the fact arrives, nothing pushes it into place. The atom described by quantum mechanics lacks an author — the collapse settles no question and answers to nobody — and any attempt to purchase freedom with indeterminacy alone stops there.

Trace the deed backward from the hand. The fingers close because the flexors contract; the flexors contract when motor neurons fire; those neurons fire as a command descends from the cortex. Every link is ordinary physiology, as determined as anything in the body, and no freedom is required at any of them. The chain runs back to the cortical transition from which the command issued, and there the inheritance stops. Asking what pushed the transition down one pathway assumes a signal poised at a fork with the pathway still to be assigned, and no such state occurs — as no state occurs in which an atom is decaying with the time of its decay outstanding. The transition and its outcome arrive together. Before it, there was no fact about which pathway would carry the command; the completion is entire, and nothing came after it to make it so.

This differs from weak constitutive identity at the point of physical determination. If the past fixed the outcome, and the free act is a description of that outcome at another level, then freedom has been named rather than found. A second description cannot open an event the first description has closed. Davidson's anomalous monism, List's account of agential control, the softer emergentisms — each preserves rational explanation and macro-level control, and each leaves the libertarian joint out. The claim here is not that a determined transition can be redescribed as a settling. It is that the transition was not determined. Where the outcome remained open, the settling is what the completion is: an answer given, and the physiology carrying it out from there.

Penrose looks for something more specific: a physically real non-computable transition. If consciousness exceeds algorithms, then a physics that exceeds computation is where it should be found. Even if such a physics is granted, non-computability is not authorship. An unpredictable transition may settle no question and answer to nobody, and Orch OR gives no account of the step from non-computability to authorship. A quantum event becomes the settling when it resolves an open question as its author's answer.

Downward causation and constraint face the same limitation. A system's organization can shape which micro-states are available, and this describes how a settled decision is realized in a body — how a commitment propagates into muscle. It does not explain the settling. Either the constraint was fixed by the system's organization, in which case it is causally inherited and the question moves to what fixed the organization, or the constraint is laid down in the act, in which case freedom has been assumed and not explained. Constraint belongs on the realization side of the account, in the execution of what the positing settles, and it cannot do the work the positing does.

From inside physical inquiry, a free act will appear as an incomplete model, noise, a statistical deviation, or a residual awaiting causal explanation. No physical marker could establish authorship. A discovered regularity might well identify the conditions under which the settling occurs — neural signatures, statistical profiles, behavioral correlates are all compatible with this account — without showing whether those conditions sufficiently fixed this particular behavior. There must be no regularity from which the outcome follows uniquely. Regularity as such may remain.

Calling the author immaterial suggests a second kind of substance, and none is introduced here. The answer runs ahead of what has been established, and the ground for it comes later, where sediment is worked out. Matter is not a class of things: it is a form held in repetition. What repeats can be scanned, subsumed, and predicted, and this is what makes it material. A settling occurs once and repeats never, and falls outside the term "matter" for that reason. Kim's dilemma needs a mental cause beside a physical one and competing for the same work. Nothing here competes: the author is what an act is when it cannot repeat, and no second thing was added to the world for him to be.

Consider a device that enters an event in its log only when a second instance allows both to be assigned to a type. An event of any size, occurring once, leaves nothing in the log. The account it produces of the stream is complete in its own terms. Refinement changes nothing: the omission follows from what counts as a record there, rather than from how finely the device resolves its input.

The settling is physically realized without being physically predetermined. It is not a second cause and adds no new force to the world. Whether a settling can be told apart from chance or from a cause not yet found is a question about evidence, and the evidence has to come from somewhere.