Natural Knowledge in the Latin Middle Ages 99Aristotle’s Metaphysics, a standard work in the faculty of arts. A third ap-
proach sought the foundations of specifi c sciences not in a higher science,
but in experience. Indeed, a well- documented bevy of thinkers at Oxford
treated the principles of natural philosophy in precisely this way.^75 Al-
though schematic, this spectrum of approaches illustrates the very differ-
ent ways in which medieval thinkers who all admired the Posterior Analyt-
ics sought to ground their natural knowledge.
Indeed, attitudes toward Aristotle’s Posterior Analytics warrant a closer
look. This work clearly set out a goal of universal and necessary natural
knowledge that found a receptive audience in the early thirteenth cen-
tury: “demonstrative knowledge” certainly counted as scientifi c. But was
only knowledge of this sort scientifi c? As Eileen Serene has emphasized,
the enduring enthusiasm for the Aristotelian ideal of demonstration gives
a false impression of unanimity among late medieval thinkers. Their views
of truth, necessity, and certainty—the key components of demonstra-
tion—varied widely.^76 Hard thinking about the problems that Aristotle
had raised (and others that his obscure points generated) introduced in-
novations, doubts, and emendations.
These developments show what an oversimplifi cation it is to identify
natural- scientifi c knowledge in the later Middle Ages exclusively with Ar-
istotle’s ideal of demonstration, or with causal knowledge, or with univer-
sal propositions, or with knowledge of the “essential natures of things.”
In fact, late medieval thinkers took some remarkably un- Aristotelian posi-
tions. Kilwardby already recognized that many cognitions appropriately
called scientia were probable rather than certain, especially in natural phi-
losophy. To be sure, in his hierarchy of true and certain knowledge, meta-
physics and mathematics tied for fi rst place, as the dignity of the former’s
subject matter vied with the certainty of the latter’s mode of demonstration.
Physica came next, followed by ethica; mechanica occupied the lowest level
of certainty. Kilwardby conceded that “natural philosophy [physica] does
not always conclude from necessary [propositions]... it often concludes
by means of probable [propositions] that, in truth, are false.” For him,
physica could attain certainty not in all, but in many contingent things,
for the latter display great regularity (unlike the more erratic arts).^77
As ideals, certainty and universality did not disappear, of course, but
it is crucial to understand that not everyone thought they were required
for scientifi c knowledge. Bacon held that “the certitude of a particular
experience of singulars is more perfect and greater than that of a univer-
sal experience.” John Duns Scotus (d. 1308) held that experience could
make fi rst principles evident. But he also believed that events (including
natural ones) occur because many particular causes act together. In addi-