Chapter 4 studies three theories of free will of the late thirteenth century: two that are midway between intellectualism and voluntarism, and one that is strictly intellectualist. Giles of Rome starts from the intellectualist assumption that the cognized object causes the will’s act, but makes a voluntarist concession in assuming that the will controls which aspects of the object end up moving the will. John of Morrovalle (also known as John of Murro) starts from the idea that the will moves itself, but makes the intellectualist concession that the cognized object causes a disposition in the will that predisposes the will in making a choice. Godfrey of Fontaines rejects both solutions and argues that the cognized object alone causes the will’s act. While Giles and Morrovalle grant the will some direct control of its choices, Godfrey considers control necessarily mediated by practical deliberation. In the last analysis, Godfrey leaves the question of how we control our deliberation unanswered.

Chapters 8–10 constitute Part III, entitled “Angelic Sin,” which raises the issue of how rational agents can do evil under ideal psychological conditions. Chapter 8 is about intellectualist accounts of angelic sin. Since according to these accounts, the will acts as the intellect judges best, evil acts presuppose some cognitive deficiency: either an outright error, or some occurrent nonconsideration that keeps the intellect from making the correct judgment. Thus one difficulty faced by intellectualist thinkers is how the cognitive deficiency can come about – especially since most thinkers here discussed assume that angels are infallible prior to making an evil choice. Another difficulty concerns control of the act. It is assumed that while the angels’ good or evil choice was up to them, the content of their knowledge was not up to them. Aquinas’s solution is that knowledge does not predetermine the use of that knowledge, which is up to the will. By contrast, Godfrey of Fontaines argues that the choice of the angels is caused by the cognized object; he fails to explain, however, how his theory avoids cognitive determinism.