Abstract
Several thought experiments attributed to Galileo and Newton have become central to argument on the ontological natures of space and time, particularly the 'absolutist/relationalist' debate. I argue, however, that in the light of the restrictions that each of these thinkers imposes on their methods of reasoning, they would not be willing to draw metaphysical conclusions from the consideration of such imagined situations. To make out this argument I draw a relatively uncontroversial theme from the vast literature written on the methods of Galileo and Newton, by focussing on what they claimed themselves: that they would justify all of their doctrines - metaphysical and physical - only by their success in accounting for the empirical evidence. Although it can be argued that both Newton and Galileo made several claims that, on close analysis, appear beyond such justification, they should not be putting forward arguments for their views that clearly violate their self-imposed restrictions. In particular, I argue that a wide class of Thought Experiments cannot be used to draw conclusions about the ontological nature of space and time. After a brief characterisation of Thought Experiments, I conclude that Galileo and Newton could not admit an argument of that form as having a bearing on the 'absolutist/relationalist' debate, on pain of contravening these restrictions. However, since Thought Experiments known as 'Galileo's Ship' and 'Newton's Bucket' are indeed used extensively within the absolute/relational controversy, a different interpretation of the original versions is needed. When these examples are examined in their original contexts it is found that there are far more natural readings available, which do not bear on the absolute/relational controversy, nor any other ontological issues for space and time In a little more detail: I argue that Galileo's Ship was proposed in reply to two specific objections made against the Copernican system, rather than intended to illustrate any relativity principle for physical processes. Its use to support the relational position is clearly not what Galileo had in mind, not least because he adopted an absolutist position even when considering very similar situations. I read Newton's Bucket as an integral part of the Scholium on the Definitions - in which it appears as part of a well-structured line of argument warning of the subtleties faced when understanding Newton's own definitions of absolute space, time, and motion: all with his mechanics and metaphysics already assumed. Newton put forward two separate experiments, which were parts of the same line of argument but meant to make very different purposes: The Bucket: a practical experiment which Newton claims to have performed, and puts forward as a demonstration to empirically exhibit some effects of absolute motion that relational motion does not, if all of the mechanics in the rest of the Principia is accepted. The Globes: a thought experiment which was meant to show how the mechanics in the Principia would serve to distinguish the subtle differences between relative and absolute motion, even in the most testing circumstances: a world without any reference bodies with which to compare the motions. Again, the mechanics (including the metaphysical claims) of the Principia must be assumed for the globes experiment to have any force. The readings of each of these experiments are compared with the 'received view' of the Scholium, as well as some more recent reinterpretations. Accordingly, I argue that the most famous of these thought experiments have been radically misinterpreted - and that Galileo and Newton used the originals to argue for quite different conclusions. This is supported by looking at those contexts in which Galileo and Newton did argue about metaphysical issues connected with space and time, and noting the contrasts with those in which the 'ship' and 'bucket' appeared.