Splitting a person in two

If your brain were divided and each half woke up with your memories in a different body, which one is you — and why can no answer satisfy?

Your brain is divided down the middle and each half transplanted into a waiting body. Both operations succeed. Two people wake up, each with your memories, character and half-finished intentions, each convinced without hesitation that they are you.

Surgically this is fantasy, though not free-floating — it is built on the real practice of severing the corpus callosum to control epilepsy. What makes it worth taking seriously is its structure: there are exactly four things one can say about the case, and every one costs something we were not prepared to lose. That is a different situation from not knowing the answer.

First, why this is not the Ship of Theseus again. That puzzle, as identity over change sets out, turns on the word "same" quietly naming several criteria — matter, form, function, unbroken path — which normally agree and can be pulled apart. Here only one criterion is in play, since everybody in this debate already accepts that what carries a person forward is psychological continuity rather than matter. The trouble is not that criteria disagree. It is that the surviving criterion duplicates.

Make the difficulty formal, because it is a formal difficulty. Identity is an equivalence relation: if A is identical to B and A is identical to C, then B is identical to C. Psychological continuity carries no such guarantee; it can run from one thing to two at once, the way a road forks.

So put the case through it. Suppose you, before the operation, are identical to Left and also to Right. Then Left is identical to Right — and they plainly are not, being in different rooms from the first second, one able to be in agony while the other sleeps. Transitivity breaks, and it breaks before any question of evidence arises.

That gives four answers, and they are exhaustive.

You are neither. Fission is death. But consider the case where only the left half is transplanted and the right destroyed. Nearly everyone calls that survival — drastic surgery, not an execution. Now save the right half too, and the same operation on the left becomes a death. Derek Parfit's objection is that doubling a success cannot produce a failure, and nothing about Left differs between the scenarios.

You are one of them. Which? The operation is symmetrical, the halves equally continuous with you. Either the answer is arbitrary, in which case identity turns on nothing, or some further fact fixes it which no examination could reveal.

You are both, because there were always two. David Lewis's answer (recalled, from a 1976 paper) takes persons to be extended in time as well as space — spacetime worms rather than things wholly present at a moment. Two worms overlapped before the operation and shared their stages, as two roads share a stretch. Transitivity is saved, since Left and Right were never identical. The price is odd counting: how many people are in a room now depends on how often any will later divide.

You survive, but identity is not what matters. Parfit's conclusion in Reasons and Persons (1984). What we care about in caring about survival is psychological connectedness and continuity with the right sort of cause — what he labels Relation R. R can hold one-to-many; identity cannot. So in fission you get everything you wanted, twice, and lose only a label. Parfit thought this liberating; many readers find it a redescription of death, and both reactions are still live.

One repair is worth examining, because its failure is the strongest argument for the fourth answer. Keep psychological continuity as the criterion but add a clause: it confers identity provided it has not branched. That restores transitivity and delivers the first answer. But look at what it implies. Whether Left is you depends on whether a second operation succeeded in another room — an event with no causal connection to Left, whose brain, body, memories and prospects are identical either way. Something we took to be a fact about a person turns out to be a fact about the world's inventory of other people. That is the "only x and y" objection.

The second answer has a serious version: a simple, indivisible self — a soul, or something like a Cartesian ego, defended in this connection by Richard Swinburne — which goes with one half and not the other, so the answer is determinate though no test could find it. Its cost is that a fact nobody can detect is a fact nobody can act on.

A word on the neurology, which is over-claimed in both directions. Commissurotomy patients, studied by Roger Sperry and Michael Gazzaniga, can be given information restricted to one hemisphere and respond in ways the other cannot report; Thomas Nagel argued in a 1971 paper that we may have no principled way to count the minds involved. But they live as one person, the dissociations appear under specific experimental conditions, and later work has questioned how divided the consciousness really is. The neurology motivates the thought experiment without establishing it.

Two refinements and a correction.

The correction first: fission is not about uncertainty. Every physical and psychological fact can be stipulated in advance, and the answers still do not converge — the mark of a concept applied outside the conditions it was built for.

Which suggests what has gone wrong. Our concept of a person formed in a world where psychological continuity never branches, so it never had to choose between two properties it happened to carry together: being one-to-one, and tracking what we care about. Fission separates them, and every answer above decides which to keep.

This is also a different puzzle from the non-identity problem, which asks who is wronged when a choice determines which person exists. That case has a clear person and an unclear harm; this one has a clear harm-free outcome and an unclear person.

Every term the collection defines is gathered in the glossary.