Abstract
Reductionist conceptions of the self have been popular in the British and American empiricist philosophical tradition since Hume to the present (see e.g. Russell’s (from the 1910s), Ayer’s, Grice’s, or Parfit’s views, among others). Such accounts may be regarded as striking a middle way between physicalist and Cartesian dualist views: though not accepting an immaterial substantial self, still upholding a certain independence of the self from the body or the brain. The most detailed account in this vein is likely to be Parfit’s theory of persons, according to which facts about the self are reducible to facts about an interconnected complex of mental states, while facts regarding the self’s identity through time are reducible to facts about psychological relations (psychological connectedness and/or continuity) between such mental complexes existing at different times. I will examine two arguments that seek to refute this kind of reductionist-relationalist conception of the self (and which, interestingly, although coming from rather different philosophical backgrounds, can also be considered as empirical objections, in a broader sense of the term).
One objection, suggested by Schechtman (1990, 1996), is based on the claim that the reductionist conception of the person must assume (as Parfit explicitly does) that it is possible to identify a person 'impersonally', in other words, that the contents of those mental states to which the person is reduced, do not involve reference to the person him- or herself (otherwise the identification would be circular). According to Schechtman, this requirement cannot be met, since the content of episodic memories, being autonoetic (i.e. presented to the rememberer as experienced by him- or herself in the past), cannot be identified impersonally (and hence the self cannot be impersonally identified either). I claim that Schechtman objection does not refute Parfitian reductionism. First, it seems there are genuine memories that are not autonoetic; second, and more interestingly, there are possible "quasi-memories" that are autonoetic.
The other criticism is related to Zahavi's (2010, 2017) concept of the "minimal self", and to the Wittgesteinian idea that we cannot misidentify ourselves as subjects of experiences (both present experiences, and remembered past experiences), i.e. that the use of "I" (as a subject) is "immune to error through misidentification". According to Zahavi, there is a minimal self (which is the basis of higher-level, more complex self-concepts) present in all our conscious states, which we infallibly identify, both synchronously and diachronically. If this is so, then the reductionist-relationalist theory is untenable, since the identity of my present self with a past self is based on the identity of my persisting minimal self (hence the psychological relations between present and past mental states are irrelevant). Against this view, I argue that it is conceivable to be mistaken in self-identification; hence it is not a logical necessity that I am the subject of all conscious experiences, present or past, I am aware of. For this reason, it is not logically necessary to assume the existence of an irreducible and persistent minimal self, which would ground the immunity thesis.
To sum up: in the presentation, I set forth to show that Schechtman's and Zahavi's objections do not invalidate the Parfitian reductionist-relationalist conception of the self. Even if we were to accept the assumption that present experiences and episodic memories of past experiences always include the consciousness of a self, this does not warrant the diachronic identity of the self, because of the falsity of the immunity thesis. Beyond this negative result, I will also shortly touch upon what consequences the arguments involved may have for a positive account of the self.
Interested in receiving updates about the events of the Department of Philosophy? Sign-up to its mailing list here.