Westcliff, Cambridge, Oxford and Berkeley, 1929–2003

Bernard Williams

Bernard Williams told the same experiment twice. Told one way, you go with your memories. Told another, you should fear what will happen to your body tomorrow, whatever memories you then have.

At night, a man in a white shirt flings up his arms before a row of soldiers aiming rifles at him, a lantern on the ground lighting him, with the bodies of the already shot in front
Francisco Goya, The Third of May 1808, 1814. Museo del Prado, Madrid. A man facing what is about to be done to his body. Williams asked whether any change to his memories would make that fear misplaced.Public domain, via Wikimedia Commons.

The view

Bernard Williams wrote two papers that shaped the debate Parfit entered. In "Personal Identity and Individuation" (1957) he imagined a man, Charles, who claims to remember the life of Guy Fawkes. If memory made identity, Charles would be Fawkes. But a second man, Robert, might make the same claims with equal right, and two men cannot both be identical with one. What can be true of two cannot be what identity consists in. This reduplication argument pointed to bodily continuity as at least necessary.

In "The Self and the Future" (1970) he went further. Two people, A and B, are told that a procedure will swap the psychological states of their bodies. Before it, each is asked to choose which resulting person will receive a large sum of money and which will be tortured. Most of us think that A should want the money to go to the B-body person, who will have A's memories and character. Told that way, the case favours the psychological view.

The argument

Now tell it differently. Someone says: "Tomorrow you will be tortured. Before that, you will lose your memories, and be given someone else's." You are not reassured; fear of the torture remains, and seems entirely reasonable. Yet this is the same case, described from A's body's point of view. Williams concluded that our intuitions follow the description, not the facts, and that fear of one's own future pain, which survives any redescription, points to the body.

The same experiment, told twice. Most people protect different bodies in the two tellings.

Where he meets Parfit

Both treat thought experiments as tests of our concept of a person, and both found that the concept gives way under pressure. Parfit accepted the point behind the reduplication argument: a relation that one person could bear to two later people cannot be identity. That is why he holds that identity is R without branching, and why branching shows that identity is not what matters.

Where he parts from Parfit

Williams drew the opposite moral. Where our intuitions conflict, he thought, the conflict reveals that the psychological view cannot be trusted, and that we should fall back on bodily continuity, with the concern for our own future that goes with it. Parfit answered in Reasons and Persons with the combined spectrum: if identity is all-or-nothing, the spectrum must contain a sharp line, which is incredible; if it is not, Williams's fear has nothing determinate to fasten on. Williams was unpersuaded, and the dispute is still alive.

Williams and Parfit side by side
WilliamsParfit
What is a person?A human being, an embodied creature with a character and a life; persons are not detachable from bodies.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?Bodily continuity is at least necessary; psychological criteria fail when duplication is possible.Relation R, psychological connectedness and continuity, holding in one line without branching.
Is identity what matters?Yes. My fear of my own future pain cannot be argued away by redescribing what will happen to me.No. What matters is Relation R, which can hold without identity, as in division.
One of you, or many?One. Identity cannot be shared, which is why a criterion that allows two candidates must be wrong.Normally one, but the concept fails when R branches; then the question may be empty.

See every view compared →

Further reading