Can manipulations of cognitive load be used to test evolutionary hypotheses?