Lecture 12: Testing the personality theory with duplication cases

What this lecture covers

This lecture puts the body theory and personality theory of personal identity to the test using vivid thought experiments. Kagan first presents two versions of a mind-swap scenario from philosopher Bernard Williams, both structurally identical, that pull intuitions in opposite directions depending on how the story is told. He then turns to duplication cases involving a fictional "Napoleon" personality appearing in two places at once, arguing that this exposes a fatal problem for the personality theory unless it is patched with an awkward "no branching" rule.

After watching, you should be able to describe Williams' two versions of the mind-swap case, explain why duplication is a problem for the personality theory, and state what the "no branching clause" is and why Kagan finds it hard to accept.

Key ideas

• Bernard Williams' mind-swap thought experiments: a mad scientist can transfer personalities between bodies; asking which resulting body you would want tortured tests which theory of identity you actually hold.
• Case one favors personality: when framed as "your mind moves to a new body," most people want the body with their old personality spared, suggesting personality is what matters.
• Case two favors the body: the same setup, framed as "you get amnesia, then get deluded into thinking you're someone else, then get tortured," makes people want their own original body spared, suggesting the body is what matters.
• Same underlying case, opposite intuitions: since both stories describe the identical transfer, conflicting intuitions show that appealing to these cases alone cannot settle which theory is correct.
• The single-Napoleon objection fails: someone who merely believes they are Napoleon, without Napoleon's full memories, is clearly not Napoleon, so this is not a real problem for the personality theory.
• The full-memory Napoleon case is more troubling but survivable: if someone had all of Napoleon's genuine memories and no other identity, Kagan finds it plausible to say that really is Napoleon.
• The duplication (branching) case is fatal: if two people, in Michigan and New York, both wake up with Napoleon's complete personality, no principled reason favors calling either one 'the real Napoleon,' and saying both are literally one person in two places violates basic assumptions about identity.
• The no branching clause: a proposed fix says you are the same person as your past self only if no one else shares your personality at the same time; Kagan objects that this makes personal identity depend on facts entirely external to you, such as what's happening in Michigan.

Walkthrough

A mad scientist's experiment to determine personal identity (0:00)

Kagan sets up the method: use science-fiction thought experiments to pull the body and personality apart, then ask which resulting person you would want spared from torture, as a guide to which theory of identity you actually believe. He presents Bernard Williams' first case, where a mad scientist swaps his and Linda's memories and personalities between their two bodies; Kagan's intuition is to spare the body with his own personality, suggesting personality is what matters.

The experiment continued: mixed personalities and bodies (7:12)

Kagan presents Williams' second version of the very same scenario, this time framed as being given amnesia, then deluded into believing you are someone else, and then tortured. Told this way, his intuition flips: he wants his own original body spared, suggesting the body is what matters. Since both stories describe an identical setup, he concludes that intuitions about these cases pull in opposite directions and cannot by themselves settle which theory is correct.

Duplication as an objection to the personality theory (19:38)

Kagan turns to a different style of objection using a "Napoleon" case: a man in an asylum who merely believes he is Napoleon is obviously not Napoleon, since he lacks Napoleon's actual personality. But if someone had Napoleon's complete memories and desires, with no trace of any other identity, Kagan is willing to consider that this really would be Napoleon. He then escalates the case: what if two separate people, one in Michigan and one in New York, both woke up with Napoleon's complete personality?

No branching clause as an acceptable fix? (40:26)

Working through the options for the duplication case, Kagan rules out saying only one of the two duplicates is really Napoleon (arbitrary), and rules out saying both are literally the same single person in two places (violates basic assumptions that a person can't be in two locations at once). The least bad option, he argues, is to say neither duplicate is Napoleon, which means the simple personality theory must be rejected. He then proposes a revision: add a "no branching clause," so sameness of personality only establishes identity when there is no competing duplicate. He objects that this makes whether you are the same person as your past self depend on facts entirely outside yourself, such as an unrelated event happening in Michigan, which he finds very hard to accept.

Before you watch

• Watch the previous lecture introducing the body theory and personality theory of personal identity, since this lecture assumes familiarity with both.
• Recall the brain-centered version of the body theory from the prior lecture, since a student's question in this lecture revisits it directly.

Check your understanding

1. Describe Bernard Williams' two versions of the mind-swap thought experiment, and explain why they are really the same underlying case.
2. Why does Kagan think the simple 'a man in Michigan believes he's Napoleon' case is not a real problem for the personality theory?
3. Explain why the duplication case, with Napoleon's personality appearing in both Michigan and New York, causes trouble for the personality theory.
4. What is the 'no branching clause,' and how does it try to rescue the personality theory?
5. Why does Kagan think the no branching clause is hard to accept, even though it avoids the duplication problem?