Caltech and Johns Hopkins, born 1966

Sean Carroll

A measurement has been made and the world has branched, but you have not yet looked at the result. Sean Carroll points out that in this short interval you know everything about the multiverse except one thing: which branch you are in.

An empty, grey-white room with bare floorboards and a series of open white doors leading one into another, with soft light falling through
Vilhelm Hammershøi, White Doors, Strandgade 25, c. 1913. Doors opening into rooms that look almost the same. Which one are you standing in?Public domain, via Wikimedia Commons.

The view

Sean Carroll is a theoretical physicist and one of the clearest public advocates of Everett's theory, which he defends in Something Deeply Hidden (2019). With the philosopher Charles Sebens he proposed an account of quantum probability in "Self-Locating Uncertainty and the Origin of Probability in Everettian Quantum Mechanics", published in 2018.

The key observation is about timing. Branching, on their account, happens very quickly, as soon as a measurement result is recorded in the environment, long before the result reaches your brain. So there is an interval in which the world has branched, there are already several copies of you, each in a different branch, and none of them yet knows which branch it is in. In that interval you know the complete state of the multiverse and are still uncertain about something: your own location in it. That is self-locating uncertainty, like waking in one of two identical hotel rooms and not knowing which.

The argument

Sebens and Carroll argue that a rational agent in this position should assign credences to being in each branch according to the squared amplitudes, the Born rule, using a principle they call epistemic separability: what you believe about your own situation should not depend on the state of parts of the universe that are causally separated from you. If the argument works, Everett's theory recovers the probabilities physicists use without adding anything to the equation.

The particle has been measured. The world has branched. You have not looked.

Between branching and looking there are already two of you, each with the same memories and the same question.

Where he meets Parfit

Carroll writes about personal identity as Parfit does: there is no deep fact about which future person is "really" you, and our talk of the same person across branchings is a convention that serves us well. Both take the view that what we should care about is not settled by asking which successor will be me.

Where he parts from Parfit

Carroll's interest is probability, not what matters. His account depends on there being a fact, after branching, about which branch each copy is in, and so on a genuine question of identity for the copies, even if not for the person before branching. Parfit's view leaves no such question to be uncertain about. Carroll also has a page on the time site, for his account of the arrow of time.

Carroll and Parfit side by side
CarrollParfit
What is a person?A physical system, a pattern in the quantum state of a brain and body, described by the same laws as everything else.Nothing over and above a brain, a body and a series of interrelated physical and mental events.
What makes you the same person over time?Continuity along a branch; after branching there are two people, each with a good claim to the shared past.Relation R, psychological connectedness and continuity, holding in one line without branching.
Is identity what matters?Not deeply. Talk of "the same person" across branches is a convenient convention, useful for planning.No. What matters is Relation R, which can hold without identity, as in division.
One of you, or many?Many after each branching. Before you look, you are uncertain which of them you are.Normally one, but the concept fails when R branches; then the question may be empty.

See every view compared →

Further reading