Annual Lecture Series of the Center for the Philosophy of Science, University of Pittsburgh
November 18, 2011.
Abstract. What are the special epistemological problems that arise in the sciences that aim to know the past? And is there a problem with calling these sorts of inquiry 'sciences' at all, to the extent that they are concerned with non-present and non-repeatable states of affairs? It was in large part in view of its pastness that Karl Popper qualified evolutionary theory as a 'metaphysical research programme'. Carl Hempel was more or less alone among 20th-century philosophers of science in making the case for the possible scientificity of history, yet for him history was to be conceived first and foremost as involving the subsumption of particular events under 'general historical laws'. For Hempel, the study of the past need no more be concerned with particular events or things than theoretical physics is. Yet even if the record of past events reveals patterns and regularities, those who study them must confront and explain irreducible singularities --why this outcropping of rock looks just like this and not some other way, or why this burial mound is exactly the size it is-- that sharply divide the sciences of the past from the sciences, like physics, to which Hempel hoped to assimilate them. It is this distinguishing feature that led G. W. Leibniz to define history, whether natural or civil, as nothing other than the science of res singulares, and that since the time of Leibniz has led many thinkers to suggest that criteria for truth and knowledge in the historical sciences must be established differently than in other fields. Two notable examples of such suggestions are, first, C. S. Peirce's notion of abduction, as distinct from both induction and deduction, which, though he did not propose it mainly for this purpose, has been particularly well suited in such fields as paleontology and archeology to the task of reasoning back to a particular event from its particular traces; and, second, William Whewell's description of evolutionary theory as a 'consilience of inductions'. In this paper, I would like, first of all, to describe the early development of a methodology for the 'science of singular things' in the work of Leibniz and his contemporaries, particularly in their effort to reconstruct the earth's biotic past based on fossil evidence. Next, I will move on to consider more recent reflections, from the evolutionary biologist George Gaylord Simpson and others, on some of the inherent epistemological difficulties in the classification of extinct species, and how this sets the project of paleontological taxonomy apart from the taxonomy of living species. Then I will turn to a consideration of the theoretical engagement with the epistemological obstacles confronted in the project of knowing the past as this has developed in archeological theory over the course of the 20th century, focusing particularly on the processualist school, which attempts to apply Hempelian ideas about history as science within a concrete domain of empirical inquiry. In all of the areas of inquiry we consider --evolutionary theory, paleontology, archeology, and civil history--, I hope to show, first of all, that the same basic difficulties are confronted: a fact that strongly favors an approach to any one of these fields as a branch, along with all the others, of a general science of the past. Second, I aim to show that the unifying thread in all of these was most clearly discerned in the 17th-century notion of 'natural history', and in particular in the insight of the natural historians that the distortive effect of the passage of time in the vestiges history studies --as when the onion-like layers of a mammoth tusk erode-- might be approached not as an impediment to real knowledge of the past, but rather as the very key to its reconstruction.