Demonstration and Scientific Knowledge in William of Ockham: A Translation of Summa Logicae III-II: De Syllogismo Demonstrativo, and Selections from the Prologue to the Ordinatio (review)

William of Ockham never delivered his promised exposition of Posterior Analytics. In its place, we have Summa Logicae (SL) III–II, and the Prologue to his Ordinatio. Longeway's literal translations append writings on the Posterior Analytics by several medieval thinkers and are preceded by a 140-page discussion of medieval science, beginning with Robert Grosseteste and proceeding through Albert the Great, Thomas Aquinas, and John Duns Scotus, among others. Longeway also provides indices of proper names and citations, an analytic table of contents, a glossary, and tables detailing various thinkers' views on matters pertaining to demonstration. Excepting references to the SL, the tables lack citation, and here, as in other instances, the absence of an index of subjects is felt.

Longeway's translation of SL III–II makes three deviations from Gál and Brown's critical edition (St. Bonaventure, 1974). The first two lack support in the apparatus. At chapter 3, l. 21, Longeway reads 'si' for 'sed', which is not required to preserve Ockham's point that a hypothetical syllogism's major premise need not commit one to an attribute's existence. At chapter 5, l. 24, Longeway replaces 'quin' with 'quia', which is unnecessary if 'quin' is rendered 'in fact'. Third, Longeway deletes 'non' at chapter 34, l. 36, relying on Marc of Benevento's error-prone 1508 edition, an emendation required by the sense of the text. But these emendations are minor and Longeway's translation is excellent.

Longeway views Ockham as the "founder of European empiricism" (1), and notes his relevance to contemporary issues surrounding the Posterior Analytics, sketching current literature. Longeway supports this with a careful reading of the Latin commentary tradition prior to Ockham. Grosseteste, for example, holds that proof in natural science requires that we grasp real natures, but insists on a role for observation and experimentation, moving toward an Ockhamist epistemology that denies that we can deduce an attribute's inherence through analysis of its subject (32).

Albert believes that proof in natural science emerges from attributes' definitions, thus accounting for inherence through factors extrinsic to the subject. Aquinas's understanding of proof resembles Grosseteste's inasmuch as Aquinas believes demonstration emerges from a subject's definition, with the important difference that Aquinas does not share Grosseteste's insistence that the definition reference final causality—a belief emerging from Grosseteste's Platonist conviction that exemplars in the divine mind seek to reproduce themselves in matter. Instead, Longeway notes, Aquinas relies on definition in terms of efficient causality (69).

This judgment requires qualification, for Aquinas states that an attribute's inherence instances the second type of belonging per se, viewing the subject as the material cause of inherence (Commentary on the Posterior Analytics, I.10, ll. 51 ff.). Again on the issue of belonging per se, Longeway claims Aquinas ascribes the fourth type "to an efficient causal connection between subject and predicate" (2), though Aquinas states that this mode of belonging takes in all four Aristotelian causes (ibid., I.10, ll. 122 ff.). Moreover, Longeway believes Aquinas uses the second type of per se belonging in the major premise of scientific demonstrations (147), whereas Aquinas actually states that this premise proposes the fourth (Ibid., I.13, ll. 60 ff.)—which does not rule out its instancing the second type, the fourth comprising the others.

Scotus's distinction between experiential (per experientiam) and analytic (per se nota) scientific knowledge (scientia) (Ord. I, d. 3, pars 1, q. 4, n. 228) challenges the claim that Ockam founds European empiricism. Experiential scientia must occasionally rely on the principle that like causes generally produce like effects (ibid., n. 235), resting content with mere probability (ibid., n. 237). Scotus deems this the lowest degree of scientia (ibid...