This is an excellent essay. I’m trying to figure out where I fall off the bus on Parfit’s theory of identity, despite sharing many of his conclusions. I’m going to lay out my misgivings, and those who understand Parfit better than me, please set me straight.
I think reductionism that is not tied to physicalism implicitly invokes a “further fact” in the form of an abstraction of identity that could follow a hypothetical non-physical continuation of the self.
If there is no further fact beyond the psychological/physiological states, then a notion of identity that refers to those states must be constrained by their actual physical properties.
I think all of the thought experiments Parfit uses to show that reductionism allows for branching and other ambiguities in identity are physically impossible in principle. I don’t want this to become a rant about metaphysical possibility, but to summarize: I don’t think there can be any physically consistent universe in which there could be something like a person, and that person could be fissioned, teleported, or blended in a continuous spectrum with another person. The most plausible case would be split-brain fission; however, I suspect this is actually impossible, and if it could happen, we would develop a convention like we have for mother and daughter cells. For the spectrum, I don’t think there’s a way to even make it conceptually coherent while respecting physicalism at all.
If all these thought experiments are impossible, then, rather than undermining the concept, reductionism leads to a clear, unambiguous theory of identity.
Also, to clarify, I’m totally comfortable with identity being ambiguous in the conventional sense. “Am I the same person as I was yesterday?” is partly a matter of taste.
Why would any conceivable arrangement of particles be "physically impossible in principle"? (It may be difficult for us to *reliably cause* the transitions in question; but - from what I understand - quantum mechanics would seem to imply that a transition from one described state to the other is perfectly *possible*, however unlikely.) The various world-states or arrangements of particles that Parfit describes seem perfectly coherent to me.
The specifics of why I think each case is impossible depends on its particulars and i can go into them if you like but first i should point out there are tons of arrangements that are forbidden by physical laws. A super simple example is there is no possible way you can pack three electrons in an orbital. Probably a more relevant one to this would be the no-cloning theorem.
Even if we were to allow the entire universe to tunnel into different arrangements while breaking locality and thermodynamic constraints, the identity question is about continuity of the physical process. As I understand it, the process is the person and the rearrangements would necessarily introduce discontinuities.
As far as the spectrum, i'm not trying to be dismissive, but i don't understand what it would mean for an arrangement of atoms to form a brain that produces a phenomenological mix of two people at once.
Just to add a little more about whats actually allowed by quantum mechanics.
Quantum mechanics requires the wave function to follow unitary evolution which respects relativistic locality for physical systems. This means that even if a final state is possible it might not be reachable from another.
A quick clarification: on Parfit's preferred version of the psychological view, you don't really need any kind of physical continuity between the two stages. All you need is for person-stage P1 to exist at t1 with a certain psychology, and then a person-stage P2 to exist at t2 with a sufficiently similar psychology, and then P3 at t3 sufficiently similar psychologically to P2, and so on, to qualify as securing "psychological continuity" between P1 and Pn.
Edited to add: I don't even think the times need to be continuous. If quantum fluctuations created a near-perfect Napolean duplicate in my living room tomorrow, he could still count as psychologically continuous with some earlier Napolean-stage.
Thank's for clarifying that does help my thinking.
Setting aside physical possibility for a second, I would say that if P1, P2, and P3 differ at all and are not connected by a causal chain, then they are different people. Psychological or even physical similarity are, in my view, red herrings for identity. People change radically over time, and what unifies their identity is the continuous path between them. This continuous chain is also what guarantees psychological continuity because it is the mechanism by which prior states imprint themselves on later states. Memories are nothing but effects of prior causes.
This includes periods of unconsciousness such as sleep, anesthesia, and amnesia where awareness is absent but the underlying physical process still links the stages together.
Returning to physical possibility, any physically possible sequence of (P, t) pairs cannot include jumps, branches, or merges. To use your Napoleon example, suppose Napoleon from the year 1800 suddenly appears in your living room in 2025. For me to regard this as the same Napoleon, his 2025 state would have to be exactly what you would get by taking his 1800 state and instantaneously skipping forward 225 years in the same lawful physical process. In phase-space terms, this would require a Lipschitz-continuous trajectory, in either the classical description of positions and momenta or the quantum description of the system’s state vector, where all relevant variables satisfy strict continuity criteria. That is obviously impossible even in principle, since both quantum and classical physics forbid such a jump by violating essentially every constraint on the lawful evolution of fields. Without this causal continuity, the 2025 figure would, even if such duplication were possible, just be a perfect copy and not Napoleon himself.
Maybe I'm missing something in some non-physicalist link between psychological and physical states?
Today there is no need for a convention concerning the identity of conscious beings like ourselves. You're born, you change a little or a lot, you die. That's the whole story. If we eventually have the ability to dismantle people and reconstitute them, or copy a mind into another body, we'd need conventions to deal with such cases. I don't find it difficult at all to believe that a person's identity could be a matter of convention if there were difficult cases to deal with.
If Captain Kirk was beamed down to Omicron Persei 8, it would be easy to say he's still Captain Kirk. But if the transporter malfunctioned and two captains arrived, there would be no "correct" answer. We'd need a convention.
This is an excellent essay. I’m trying to figure out where I fall off the bus on Parfit’s theory of identity, despite sharing many of his conclusions. I’m going to lay out my misgivings, and those who understand Parfit better than me, please set me straight.
I think reductionism that is not tied to physicalism implicitly invokes a “further fact” in the form of an abstraction of identity that could follow a hypothetical non-physical continuation of the self.
If there is no further fact beyond the psychological/physiological states, then a notion of identity that refers to those states must be constrained by their actual physical properties.
I think all of the thought experiments Parfit uses to show that reductionism allows for branching and other ambiguities in identity are physically impossible in principle. I don’t want this to become a rant about metaphysical possibility, but to summarize: I don’t think there can be any physically consistent universe in which there could be something like a person, and that person could be fissioned, teleported, or blended in a continuous spectrum with another person. The most plausible case would be split-brain fission; however, I suspect this is actually impossible, and if it could happen, we would develop a convention like we have for mother and daughter cells. For the spectrum, I don’t think there’s a way to even make it conceptually coherent while respecting physicalism at all.
If all these thought experiments are impossible, then, rather than undermining the concept, reductionism leads to a clear, unambiguous theory of identity.
Also, to clarify, I’m totally comfortable with identity being ambiguous in the conventional sense. “Am I the same person as I was yesterday?” is partly a matter of taste.
Why would any conceivable arrangement of particles be "physically impossible in principle"? (It may be difficult for us to *reliably cause* the transitions in question; but - from what I understand - quantum mechanics would seem to imply that a transition from one described state to the other is perfectly *possible*, however unlikely.) The various world-states or arrangements of particles that Parfit describes seem perfectly coherent to me.
The specifics of why I think each case is impossible depends on its particulars and i can go into them if you like but first i should point out there are tons of arrangements that are forbidden by physical laws. A super simple example is there is no possible way you can pack three electrons in an orbital. Probably a more relevant one to this would be the no-cloning theorem.
Even if we were to allow the entire universe to tunnel into different arrangements while breaking locality and thermodynamic constraints, the identity question is about continuity of the physical process. As I understand it, the process is the person and the rearrangements would necessarily introduce discontinuities.
As far as the spectrum, i'm not trying to be dismissive, but i don't understand what it would mean for an arrangement of atoms to form a brain that produces a phenomenological mix of two people at once.
Just to add a little more about whats actually allowed by quantum mechanics.
Quantum mechanics requires the wave function to follow unitary evolution which respects relativistic locality for physical systems. This means that even if a final state is possible it might not be reachable from another.
A quick clarification: on Parfit's preferred version of the psychological view, you don't really need any kind of physical continuity between the two stages. All you need is for person-stage P1 to exist at t1 with a certain psychology, and then a person-stage P2 to exist at t2 with a sufficiently similar psychology, and then P3 at t3 sufficiently similar psychologically to P2, and so on, to qualify as securing "psychological continuity" between P1 and Pn.
Edited to add: I don't even think the times need to be continuous. If quantum fluctuations created a near-perfect Napolean duplicate in my living room tomorrow, he could still count as psychologically continuous with some earlier Napolean-stage.
Thank's for clarifying that does help my thinking.
Setting aside physical possibility for a second, I would say that if P1, P2, and P3 differ at all and are not connected by a causal chain, then they are different people. Psychological or even physical similarity are, in my view, red herrings for identity. People change radically over time, and what unifies their identity is the continuous path between them. This continuous chain is also what guarantees psychological continuity because it is the mechanism by which prior states imprint themselves on later states. Memories are nothing but effects of prior causes.
This includes periods of unconsciousness such as sleep, anesthesia, and amnesia where awareness is absent but the underlying physical process still links the stages together.
Returning to physical possibility, any physically possible sequence of (P, t) pairs cannot include jumps, branches, or merges. To use your Napoleon example, suppose Napoleon from the year 1800 suddenly appears in your living room in 2025. For me to regard this as the same Napoleon, his 2025 state would have to be exactly what you would get by taking his 1800 state and instantaneously skipping forward 225 years in the same lawful physical process. In phase-space terms, this would require a Lipschitz-continuous trajectory, in either the classical description of positions and momenta or the quantum description of the system’s state vector, where all relevant variables satisfy strict continuity criteria. That is obviously impossible even in principle, since both quantum and classical physics forbid such a jump by violating essentially every constraint on the lawful evolution of fields. Without this causal continuity, the 2025 figure would, even if such duplication were possible, just be a perfect copy and not Napoleon himself.
Maybe I'm missing something in some non-physicalist link between psychological and physical states?
Today there is no need for a convention concerning the identity of conscious beings like ourselves. You're born, you change a little or a lot, you die. That's the whole story. If we eventually have the ability to dismantle people and reconstitute them, or copy a mind into another body, we'd need conventions to deal with such cases. I don't find it difficult at all to believe that a person's identity could be a matter of convention if there were difficult cases to deal with.
If Captain Kirk was beamed down to Omicron Persei 8, it would be easy to say he's still Captain Kirk. But if the transporter malfunctioned and two captains arrived, there would be no "correct" answer. We'd need a convention.