This note explores the consequence of hidden information acquisition f
or static choice theory. We show that any choice function in the obser
vable problem can be consistent with some well-behaved choice function
in a metaproblem with unobservable costly information acquisition. Th
is illustrates how choices may not satisfy consistency conditions beca
use a decision maker's decision process (in this case, information acq
uisition) depends on her feasible set. It also illustrates the importa
nce of modeling the source of violations of consistency conditions, ra
ther than simply weakening axioms on preferences.