Choice behavior is studied differently in humans and in animals, and different theories have arisen to explain the results. I suggest that an approach derived from animal studies is also appropriate for human choice. Daniel Kahneman and Amos Tversky's prospect theory, a popular two-part account of human choice, is a functional theory amounting, after some years of evolution, to a classification of types of deviation from “rational” reward maximization. Animal choice, on the other hand, can be explained causally as the outcome of competition between a set of possible responses with different “strengths.” The strength of each response is directly related to its historical payoff probability, and responses compete in winner-take-all fashion. An “active” response occurs and is strengthened or weakened depending on its outcome. If it is sufficiently weakened, it will be supplanted by the strongest “silent” response. This cumulative effects (CE) model has been tested in operant conditioning experiments that show, for example, that when choosing between two identical probabilistic choices in a “two-armed bandit” situation, animals will fixate on one if the payoff probabilities are high, but be indifferent if they are low, a pattern not easily deducible from any kind of optimality theory. Kahneman's distinction between “fast” and “slow” systems is indistinguishable from the distinction between active and silent responses in the CE model, which therefore offers a causal account of human as well as animal choice behavior.