Hume's Recantation of His Theory of Personal Identity

Hume Studies Volume 30, Number 2, November 2004, pp. 257-264 Hume's Recantation of His Theory of Personal Identity DAVID PEARS I am going to defend a diagnosis of Hume's recantation that I have already defended —rather unsuccessfully—in more than one publication.1 My excuse for trying again is that I shall now offer a more carefully qualified defense. My diagnosis was, and still is, that in the Appendix to the Treatise Hume came to see that he could not account for the necessary ownership of perceptions (impressions and ideas)—i.e., for the fact that this very perception could not have occurred in a different set. I will argue that Hume realized that his account of the identity of particular perceptions was defective, but that he could not see his way to putting it right. Of course, there are many ways of interpreting this diagnosis. One interpretation would be that he hadn't the faintest idea what was wrong with his account of the identity of particular perceptions and perhaps did not even realize that this was in fact the source of the weakness of his theory of personal identity. The diagnosis at the other end of the spectrum would be that he knew that he needed a theory of individual essences,, but did not know how he could formulate it. In a case like this it is a mistake to be too precise about the contents of a baffled philosopher's mind. He can see the deficiency in his system and perhaps even the general way in which it might be made good. But the detailed structure of the remedy is not visible even on his intellectual horizon—it is, as it were, hull-down. So my diagnosis will not imply that Hume could see the specific theory that he needed—for example, Saul Kripke's theory of necessity of origin—but only that he would have been content if he could have formulated and adopted a theory David Pears is Professor of Philosophy, Emeritus, at Oxford University and a Fellow of Christ Church, Oxford, OXl 6DP, UK. 258 David Pears that played a similar role.2 However, any possibility of that kind was blocked by ineradicable obstacles in his system. So in what follows I shall argue only that Hume was aware of a gap in his theory of personal identity and had a vague idea of what would fill it but nothing more than that. Any detailed account of what would fill it was beyond his ken. There is a second introductory point that I need to make before I start the defense of my diagnosis of Hume's dissatisfaction with his own theory. Don Garrett has produced a comprehensive review of the various diagnoses of Hume's predicament that have been proposed, including mine, and has put forward a diagnosis of his own.3 But, unfortunately, I am unable to discuss all these competitors in this short paper. My last introductory point is terminological. I need to distinguish tokenbased properties of perceptions from their type-based properties. For example, an impression's position in a particular sequence is a token-based property; its character—for example, its being a perception of blue—is a type-based property. The same distinction applies to ideas. Now Hume's discussion of personal identity is primarily a discussion of the unity of a person and so, given his mentalistic approach, primarily a discussion of the unity of a mind. But it is not wrong, but at most only a wrong emphasis, to read it as a discussion of the reidentification of persons after a lapse of time. For the considerations that settle the question of unity or multiplicity will also settle the question of identity or diversity. Of course, the two questions will be phrased differently to reflect the different situations in which they arise. The first question would usually be raised in a mental hospital, while the second one would more often be heard in a police station or a law court. But the material needed in order to answer them would be the same. Whichever question we consider, the distinction between token-based properties...

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.