Brains, Persons, and Society *** ABSTRACTS    Cervelli, Persone e Società ***ABSTRACTS

Andrea C. Bottani
Università di Bergamo

Why the Psychological Criterion of Personal Identity should not be regarded as a Criterion of Identity

According to the psychological criterion of personal identity (henceforth PC), personal identity across time consists in psychological continuity: the adult I am and the boy I was are the same person because the former is psychologically continuous with the latter (Shoemaker 1970, Parfit 1984, Nozick 1981). According to standard versions of PC, a person a existing at a time tⁿ and a person b existing at a time t^k are psychologically continuous if and only if a’s total mental state at tⁿ and b’s total mental state at t^k are both similar and causally connected to a sufficient degree (this can never be the case if n = k, for no mental state can be causally connected with itself in any degree). Since two different persons existing at the same time can both be psychologically continuous with one and the same person (at least in principle), psychological continuity with a person cannot be a sufficient condition of identity across time with her. Modified versions of PC avoid the difficulty  at the cost of some ad hoc flavour. According to them, a person a existing at a time tⁿ is identical with a person b existing at a time t^k (n  k) just in case a is the only person existing at tⁿ that is psychologically continuous with b (“one candidate theory”) – or else if a is psychologically continuous with b more than any other person existing at tⁿ (“best candidate theory” – see Shoemaker and Swinburne 1984, Nozick 1981).

             PC preserves Locke’s intuition that personal identity is not the identity of a substance, either material or immaterial, but seems to be incompatible (even in the aforementioned modified versions) with a number of commonsensical assumptions concerning our own identity across time. According to PC, for example, our identity across time requires neither organic continuity nor spatiotemporal continuity, it seems to admit degrees etc. Moreover, PC seems to treat persons as general types rather than particular tokens.

In this paper, I develop a thought experiment aimed to show that PC cannot work as a criterion of numerical personal identity across time (even if it might work as a criterion of something else, perhaps of survival in Parfit’s sense). Although the thought experiment is new (at least as far as I know), the starting point is quite familiar, being one of the ontological structures (like Black’s two spheres) that seem to falsify the identity of indiscernibles. Imagine the universe is divided in two indistinguishable halves, each having exactly the same properties as the other (haecceitates excluded). For any moment tⁿ, any total mental state M that I can have at tⁿ and all moments t^k > tⁿ, if one person that exists at t^k has a total mental state sufficiently similar to M, then two persons that exist at tⁿ (each in a different half of the universe) have a total mental state sufficiently similar to M, and to the same extent. But only one of them – the one existing in my half of the universe – is psychologically continuous with the person I am at tⁿ, for only the total mental state of that person can be causally connected with my mental state at tⁿ. But now imagine that lines of causal efficacy cross from one half of the universe to the other and vice versa (the global state of each half at each moment, say, is calculated as a function of the state of the other half in a very near past moment). Since in such a universe (maybe our universe, as far as we know) lines of causal efficacy follow a zigzag path from one half to the other, lines of psychological continuity have to follow the same path. According to PC, therefore, I frenetically jump from one half of such a universe to the other during all my lifespan, changing incessantly my body and occupying a spatiotemporally scattered region. This, however, is not all. I shall argue that, in a universe where lines of causal efficacy are more complicated, PC has even more implausible consequences. The most weird of them is that personal identity is not transitive. If what PC calls “personal identity” is not transitive, it cannot be an equivalence relation. And no relation can be a relation of identity if it not an equivalence relation. I conclude that PC should not be regarded as a criterion of identity.