Reading Assignment Jeremy Chapman
The Framework of Philosophical Perplexity 3726223
Sklar pursues the possibilities of Poincaré's parable in more depth. Specifically, he introduces the philosophical possibilities inherent in this conventionalist world. He addresses three specific points:
Sklar briefly refreshes our knowledge of the conventionalist position according to Poincaré's parable: one who argues that there are two alternative theories about the world which are mutually incompatible yet support each other equally well, with an equal "refutation potential" (for one to be refuted, the other must also be discarded). The problem is that theories are tested by observation and experiment, and no conceivable experiment could lead us to prefer one theory to another. The only option left is to simply choose one of the theories as a "true" model of the world. This is a conventional truth which is distinguished from truth that is proved by its accordance with observational evidence. The empiricists, on the other hand, were busy defending the position that the appearance of mutually exclusive theories that were equally compatible and incompatible with observational evidence was merely that, an appearance. Theories must be hypotheses developed over time to fit experimental data and observations. Since this is the case, the two theories must be one and the same, for if they are equivalent in respect to outcome of experiment, they are not different. Since there are not two different theories there is no need to choose between one or the other, and hence no need for a conventionalist position.
2. Observational Basis and Theoretical Superstructure
For theories to be formulated, they must be based on observational evidence. From the position of a relativistic spacetime theory, there are few truly observable and testable entities. In order to fully characterize the geometry of spacetime we need to have "facts" that are beyond the scope of our observational capabilities. We would need to be able to make observations about whether (idealized) rods at a distance are the same length, rather than simply whether they are the same as their neighbours. Another requirement is the ability to know time separations of objects over a distance, and therefore we require synchrony of the clocks at a distance, and also the synchrony of a clock with itself at a later time. These are all example of what Sklar calls "outrunning of the data by the theory", whereby the theory advances too quickly for observation to follow. This all goes to say that the conventionalist position is a firmly grounded one in geometric theory, and that a choice must be made between possible theories. But what are these alternative theories at our disposal?
3. Duhemian and Super-Duhemian Alternatives
Pierre Duhem has a thought experiment whereby he introduces the notion of a"crucial experiment" in science. Starting with hypothesis H, of general form "All..", and we wish to test this hypothesis. Then imagine that we proposed a result for this experiment E, and on performing the experiment, we received the result not-E. Does this mean that hypothesis H is necessarily false? However, in order to get a result from hypothesis H, we need additional theories and practices A. This includes methodology, specific laws governing measuring devices, etc. The result not-E is a result of a problem in the total theory H and A. Does this mean that Poincaréan conventionalism is just a flavor of Duhemian claims about refutability? Not necessarily, for while it is conceivable that one could construct further experiments in order to distinguish between problems with H and A, by eliminating one subsection of A at a time and repeating the experiment, in Poincaré's world, there are no possible experiments that could give one theory more robustness than the other. Therefore, Poincaré's is a "super-Duhemian" argument.
Questions