​Chapter 16

To say that freedom occurs where prior determination fails remains ambiguous. A cause may be unknown, the physical state may permit multiple outcomes, or the grounds before an agent may fail to settle what he will do. Each can be called a gap, but these are three different gaps, and only the third concerns freedom.

An epistemic gap exists when a cause operates but has not been found — the model is incomplete, the instruments are coarse, or the relevant variable was never measured. Science works in the presence of such gaps and closes them steadily. A gap of this kind reveals only the state of our knowledge.

A physical gap appears when the prior state does not fix a unique outcome, as in an indeterministic reading of the formalism. Here the openness belongs to the event rather than to the observer. Neither kind is freedom. An unknown cause remains a cause, and an undetermined outcome remains merely an outcome.

Freedom is in the agential gap: prior grounds cease to be sufficient, the question is open, and the settling completes it as someone's answer.

At John's table, the three gaps come apart. The observer who cannot say why he took the money has an epistemic gap — and would have that gap even if John's theft followed rigidly from his upbringing, as the upbringing was never measured. Suppose further that the neural transition was not fixed by the prior physical state: this is a physical gap, and by itself it yields a hand that might have moved or might not, without making the movement John's. The agential gap adds authorship. The prohibition was held out as revisable and did not settle the matter; the hunger was held out and did not settle it either; the question of what to do was open in the sense that no ground in view was sufficient. Then it was settled, by John, as his answer.

The three can be separated in a smaller case. Something turns out not to fit the category under which it was filed — the friend does something no reading of him accommodates, the practice previously classified as generosity shows itself as something else, a confidently used word fails on a case it should have covered. The old rubric has been withdrawn and no replacement has appeared. In that stretch the thing is not undifferentiated, since it is still there and still demands to be taken somehow; nor is it yet cut, since the line along which it will be cut has not been drawn. Laying down a different category ends the stretch; the old one does not return, and the laying down establishes what counts here — which features of the thing will matter, where the line falls. That is the third gap, and it is not the same as failing to know which of the available categories applies.

Two nearby cases look similar but remain passive throughout. The silence between two notes is a gap in the sounding, ruled by what lies on either side of it — an absence of sound within a field where sound has been established, which makes it a hole in the fabric rather than what the fabric is cut from. And not knowing which note comes next is openness of outcome within a settled code: the note may go wrong, and wrong is fixed in advance by the expectation. Not knowing which of the available things will happen is not the same as nothing being available yet.

The epistemic gap is not a condition of the third. A free act may be entirely expected by someone who knows the person well, and expectation is not derivation. The physical gap is a condition, conceded when closure was rejected: a settling occurring in the world requires that the prior physical state fail to suffice for the transition. The concession stays negative, and it locates nothing. A synapse is repetition in force, and an act placed inside one would be a part of the mechanism it was distinguished from. What is denied is that the answer was ready in advance. The three gaps are different structures covered by the same word.

The agential gap and the settling form a single event. If the gap comes first — an emptiness opening, and the settling arriving afterward to fill it — then the filling has a question hanging over it: what brought the filling about? A void followed by an event would be two things, and the second wants an explanation the account cannot give. The gap exists only as the underived resolution.

If the gap exists in the resolving, it may seem that nothing was open before the act and therefore nothing remained to be resolved. The complaint conflates two senses of openness. The question stood open before the settling — grounds laid out, none of them sufficient, the matter unsettled for John. That is a condition of the deliberation and of the person deliberating. It is not a hole in being, waiting to be plugged. The agential gap — the joint at which the outcome was not derived from what preceded it — is constituted in the settling, not before it.

A gap of this kind carries no advance sign of itself. No feature of the situation marks the joint as open, and what shows the opening is the settling — which does not amount to a later verification, since the settling delivers no evidence. Before the choice, John cannot know he is at a gap; he knows the question is open for him, which is compatible with his answer being fully prepared by what he is. Nobody occupies a position from which the third gap is visible in advance, which is why freedom cannot be scheduled.

A settling closes the gap it constitutes and leaves a residue. John now has an answer about need and property, and the answer is available tomorrow. Freedom disappears into its own result, which can then produce further outcomes.

Lehavah is not the gap. It is the renewal of the possibility of a gap — bringing what was posited back into an open question and again refusing to let the past be sufficient. And the capacity to do that cannot be freedom, given everything established: a capacity reliably producing free acts would be a mechanism. Freedom happens in the new settling or it does not.