Finite automata raise the question of what an agent can do in a sharp form . . .
However, they are not a useful representation of an agent's introspective knowledge of what it can do . . .
To the extent that a person or machine can achieve any of different goals,...
that person or machine has free will . !+!
Our ideas on this show up most sharply ^ considering systems of interacting discrete finite automata . . .
These are as deterministic as you can get, . .
which is why I chose them to illustrate free will. . . . .
