In his discussion of the well-known four causes, Aristotle makes a puzzling claim to the effect that 'the hypotheses are material causes of the conclusion' (Physics 2.3, Metaphysics 5.2). This is usually taken to mean that the premises of any valid deductive argument are material causes of the conclusion. By contrast, Malink argues that Aristotle's claim applies not to deductive arguments in general, but only to scientific demonstrations, i.e., to explanatory deductions from unproved first principles. For Aristotle, the theorems of a given science can be viewed as compounds consisting of the first principles from which they are demonstrated. Accordingly, the first principles of a science can be regarded as elements from which theorems are obtained by composition (synthesis), much like a syllable is obtained by composing letters. First principles are thus material causes of the theorems demonstrated from them.