Habit (Gr. hexis ) is typically understood as a part of a theory of action, or a concept in ethics, but Agamben claims ( Opus Dei: An Archaeology of Duty ) that we cannot understand how the concept works in Aristotle unless we recognize that it’s fundamentally a metaphysical concept: “If being is divided into potential and act, something will in fact be needed to render possible, regulate, and operate the passage from the one to the other. This element, which defines and articulates the passage of potential from the merely generic (the potential according to which we say that the child can learn to write or play the flute) to the effective potential of the one who already knows how to write or play the flute and can therefore put it into action, is hexis , the habit . . . of potential” (93). More succinctly, “Habit . . . is the mode in which a being . . . ‘has’ in potential a technique, a knowledge, or a faculty . . . . It is . . . the point where being crosses into having” (93-4).

As a result, habit both keeps potential and act apart and maintains them in relation. But a habit is also a privation because “having the hexis of a potential means being able not to exercise it” (94). That privative quality makes hexis a problematic concept: How can something negative move into action?

Applied to ethics, how can one “render governable the essential relation that links [habit] to privation and potential-to-not” (96)? Agamben thinks that this leads Aristotelian ethics into an aporia, in which virtue is sometimes an ontological property, sometimes a quality of work and action. He suggests that Bartleby the scrivener is the “perfect cipher of the aporias of Aristotelian ethics” in his “potential to write” combined with his inability to do so.

Scholastic treatments of habit and virtue begin from the aporia of Aristotelian ethics: “The joining of officium and virtue that is already implicit in Cicero and Ambrose, and constitutes the specific work of the ethics of late scholasticism, in fact has the goal of conferring effectiveness to virtue in the governing of habit and potential” (97-8). For Thomas, habits are ordered to action, and habits are located in the will. But inclinations to act can be both good and evil, and Thomas distinguishes the good from bad potential of habit by reference to the nature of the actor: “good habit is distinguished from bad not because it has a good object but because it is in harmony with the nature of the agent” (100).

For Agamben, this leads to new aporias: “On the one hand, the end of virtue consists in its very operativity, but on the other, insofar as it is a form of habit, it refers necessarily to the nature of the subject with which it must fit. The very expression ‘operative habit’ seems in itself contradictory, insofar as it refers at once to ontology (habit) and to praxis (operativity). Virtue is that by means of which being is indeteremined into praxis and action is substantialized into being” (101).

And this brings Agamben back to his theme, the genealogy of duty, which he traces from pagan sources through Christian sacramental and liturgical theology, to the modern age: “the definition of virtue presents more than an analogy with the circularity that characterizes the effectiveness of officium . The priest has to carry out his office as priest, but he is a priest insofar as he carries out his office. And just as the subject of the liturgical act is not truly such, but is acted upon by Christ ex opere operato , so also the subject of the virtuous act is acted upon by operative habit, so that Aquinas can write that in virtue, ‘God works in us without us’” (101).

One note for future reflection. Whatever the formal similarities between Christian notions of virtue and liturgical officium and modern ethical patterns, they diverge for precisely the reason that Aquinas gives in the final quotation: For Christian theology, virtue is an effect of divine concurrence with our actions. Once this formal pattern gets moved into the context of Kantian autonomy, that concurrence dissolves, and the ethics that results has to be radically different.