Aristotle on Knowledge and Learning: The Posterior Analytics by David Bronstein (review)

David Bronstein’s book tackles Aristotle’s account, as presented in the Posterior Analytics, of knowledge (epistēmē), or rather a privileged form of it, ‘scientific knowledge’ or ‘understanding.’ We know in this way by grasping arguments of a certain kind, demonstrations, for which reason Aristotle devotes much of his attention in the Posterior Analytics to demonstrative argument. The subject is as important as anything in Aristotle, and it presents challenges as difficult as any confronting his interpreters elsewhere, which Bronstein’s book tackles skillfully and to illuminating effect. The book is especially noteworthy for the high degree of systematic unity it attributes to the Posterior Analytics. The solutions he proposes for two prominent challenges to a “unitarian” interpretation merit special attention.

(1) The accounts of demonstration in books I and II are not obviously of a piece. Book I concentrates on demonstrations that deduce (and explain) conclusions about the minor term (the subject of the conclusion) from its essence (formulated in a definition). Book II is occupied, among other things, with questions about whether and how it is possible to syllogize or demonstrate a definition, and the uses and limitations of the Platonic method of division as a way of discovering definitions. It recognizes a form of demonstration that displays (without deducing) a definition, for example, of an eclipse (i.e. demonstrations that are, in a certain way, of an essence). It is far from clear how these demonstrations are related to those studied in book I.

(2) The object of the second book’s last chapter (19), whose fifty-two lines are among the most pregnant, oracular, and obscure in the whole corpus, is to explain how we come to know first principles, which furnish demonstrations with their premises, and what the state that knows them is. The fit between the chapter and what precedes it has long struck commentators as imperfect. The problem is that the discussion of whether and how there can be demonstrations of some definitions and how others may be established by division earlier in book II would seem to be about how first principles, many of which will be definitions, come to be known. How, then, is II 19 related to this earlier discussion?

(1) Bronstein distinguishes two varieties of demonstration: Model₁ and Model₂, which are respectively the concerns of books I and II. Each depends on one of a pair of types of per se proposition distinguished in I 4. In per se propositions of type 1, the predicate is part of the essence of the subject to which it belongs. In per se propositions of type 2, the subject is part of the essence of the predicate that belongs to it. Model₁ demonstrations take type 1 propositions as minor premises. The conclusion of a Model₂ demonstration is a type 2 per se proposition, and, like it, the major premise has a subject that is part of the predicate’s essence. Bronstein’s bold suggestion is “that every Model₂ demonstration is explanatorily grounded in a Model₁ demonstration”—the minor premise of the Model₂ demonstration will be deduced by a Model₁ demonstration from premises including per se propositions about the essence of the subject or minor term (49). If he is right, then Aristotle’s theory of demonstration in the Posterior Analytics would be more of a unity than it appears to be.

(2) On Bronstein’s view, II 19 takes up a question left unanswered by the discussion of definitions earlier in book II, with which it therefore dovetails perfectly. “It aims to provide a much smaller part of the answer [to the question about how the principles come to be known] than commentators usually think” (225); the question about “how the principles come to be known” (99b17–18) is, he reckons, about the origin of our knowledge of principles, which is the faculty of perception, from the exercise...