This paper presents a new theory of modal reasoning, i.e. reasoning ab
out what may or may not be the case, and, what must or must not be the
case. It postulates that individuals construct models of the premises
in which they make explicit only what is true. A conclusion is possib
le Ef it holds in at least one model, whereas it is necessary if ii ho
lds in all the models. The theory makes three predictions, which are c
orroborated experimentally. First, conclusions correspond to the true,
but not the false, components of possibilities. Second, there is a ke
y interaction: it is easier to infer that a situation is possible as o
pposed to impossible, whereas it is easier to infer that a situation i
s not necessary as opposed to necessary. Third, individuals make syste
matic errors of omission and of commission. We contrast the theory wit
h theories based on formal rules.