The rise of automation in the 17th century — the production of machines that “move of their own accord” — raised interesting questions for philosophical worldviews committed to the new mechanical philosophy. On the one hand, there was the promise that much of what appeared in nature to be animated (animals, plants and perhaps even the human body) could be explained solely by adapting the same mechanical principles used to explain the behaviour of clocks, mills and fountains. On the other hand, new questions emerged as to what metaphysical unity, if any, such machines, whether natural or artificial, could be said to possess. This paper draws on suggestions in Descartes and other philosophers from the period to argue that automata were widely thought to have criteria of identity distinct from those of the parts of matter that, at any time, compose them. As such, these thinkers should not be thought to have had a reductionist agenda, as is generally thought to be the case.