One box or two? A simple choice that splits philosophers down the middle.
A near-perfect predictor has filled a closed box with a million dollars only if it foresaw you taking that box alone. A second, open box holds a thousand. The money is already there, so why not take both? Yet those who take one box walk away rich. Philosophers still can't agree.
Newcomb's problem was devised by William Newcomb of the University of California's Lawrence Livermore Laboratory. The philosopher Robert Nozick published the first analysis in 1969, and Martin Gardner brought it to a wide audience in Scientific American in 1974. It has been a central puzzle of decision theory ever since.
The set-up is simple. Box A is transparent and always holds $1,000. Box B is opaque. Before you arrive, a predictor with an almost flawless track record has put $1,000,000 in box B if it predicted you would take only B, and nothing if it predicted you would take both. You may take box B alone, or both boxes.
Two sound-looking arguments collide. The expected-utility argument says the predictor is almost always right, so people who take one box almost always find a million, while people who take both almost always get only $1,000. So take one box. The dominance argument says the contents are already fixed; whatever is in B, taking both boxes gives you $1,000 more. So take both. Nozick noted that almost everyone finds the answer obvious, but people split roughly evenly, each side thinking the other is being silly. In a 2020 survey of professional philosophers, 39 per cent favoured two boxes and 31.2 per cent one.
Much hinges on what makes the predictor so accurate. If your choice somehow shapes the prediction, one-boxing looks right; if the prediction is a fixed fact that your choice cannot touch, two-boxing looks right. Some theorists argue the paradox arises because the story leaves key details unspecified, so different readers are really solving different games. The puzzle has been used to challenge causal decision theory, and it touches deep questions about free will and whether rational choice can be predicted.
Source: Wikipedia — Newcomb's problem · Text summarised from Wikipedia (CC BY-SA 4.0)
