In his account for why stars and planets move in different directions, Aristotle reasons that if the universe is to move eternally in a circle, it must have a center that remains at rest, and that if there is to be such a center, ‘it is necessary that earth exists: for this rests at the center’ (Cael II 3, 286a20-21). Interestingly, this proposition about the position and the immobility of earth is not a pre-established fact: instead, ‘we must assume it for now, and later there will be a demonstration about it’ (286a21-22: Νῦν μὲν οὖν ὑποκείσθω τοῦτο, ὕστερον δὲ δειχθήσεται περὶ αὐτοῦ). Aristotle relies on the same assumption later on (Cael II 8, 289b5-6: ‘let it be assumed – ὑποκείσθω – that the earth is at rest’), but does not demonstrate it until at the very end of book II (Cael II 13, 293b15-17: ‘it remains – λοιπὸν – to speak about earth…’). In a paper I have been working on lately, I argue for a dual role of the use of ὑποκείσθω as both an explanatory and an expository principle by elucidating its use in this particular example from the De Caelo.
Aristotle’s treatises are full of propositions that need to be assumed for now – and that thereby perform an immediate explanatory role in the arguments in which they are used – but that are ultimately, unlike the true first principles of a science, demonstrable. It is my contention that the reason why Aristotle posits these propositions as ‘surrogate principles’ instead of demonstrating them immediately has nothing to do with their epistemic status (they can be known through a demonstration). Rather, it has to do with his wish to preserve what he thinks is the proper order of exposition that is to be followed both within and among treatises. The example from the De Caelo mentioned above provides an especially interesting case, because here Aristotle connects the order of exposition to the hierarchical order of nature itself. Aristotle thus postpones his demonstration concerning earth because his writing needs to reflect the scala naturae that exists among the heavenly bodies, and this requires him to discuss the attributes of the universe as a whole before those of its parts, to discuss stars before planets, and to discuss earth last, because it is the least honorable heavenly body in existence.
My paper consists of two parts. First, I discuss in more detail the use of ὑποκείσθω in the De Caelo passage and provide some context for it by comparing it to two other uses in Aristotle’s natural treatises that connect it to his concern for preserving the proper order of exposition (in Physica VIII 7, 260b15-29 and De Partibus Animalium IV 10, 689a5-14). Next, I show that Aristotle’s concern with the proper order of exposition in the De Caelo is in fact a concern for tracking a natural order in this treatise (and not merely a didactical or conceptual one, as discussed by, e.g., Burnyeat 2001 and 2004, and Lennox 2010). This involves drawing an analogy between this case and (1) Aristotle’s use of the human body as a road map for his discussion of the parts of animals in De Partibus Animalium, according to which parts that are ranked ‘first’ on the human body have to be discussed first and parts that are ‘last’ have to be discussed last; and (2) his use of the hierarchical scale of different levels of ‘perfection’ among animal kinds as a guide for the order in which to discuss the different modes of reproduction in De Generatione Animalium. I also draw brief attention to Aristotle’s own account of the lowly status of Earth and the center it occupies in De Caelo II 13, 293b6-15, and thereby round off my argument that this is why he discusses earth last.
 The paper is for a volume called Reading Aristotle: Argument and Exposition in the Corpus Aristotelicum, edited by Ron Polansky and Bill Wians.
 Aristotle does not seem to have a technical name for these propositions that are often introduced by the present imperative form of the verb ὑπόκειμαι. The substantivated adjective, τὸ ὑποκείμενον, refers almost exclusively to the subject of a predication or to the underlying reality of something (i.e., to existing things in general), and there is no noun-form of the verb that parallels, for instance, the derivation of ὑπόθεσις from ὑποτίθημι.