In the experimental study of real-life decision making a variety of judgmental slips and inconsistencies have been revealed (intransitivity, discrepancies between choosing and matching, between choosing and pricing, positive and negative time-discounting in the same person when choosing over time, etc.). Some researchers (the proceduralists) have questioned the very idea that there are internal states of preference, and internal subjective criteria for rank-ordering the desirability of possible outcomes, not to mention the very existence of a subjective expected utility function.  Cognitive performance is, according to them, data-driven and task-elicited on the spot, not the expression of internal predispositions. I will, instead, continue to defend an internalist thesis: Such complex internal states and cognitive dispositions do exist, though accessing them may be difficult. The structure of the task at hand and the mode of presentation (framing) of the problem selectively activate or enhance certain mental representations, generating partial "projections" of these states and functions, while selectively suppressing, or attenuating, others.  I will show that an explanation based on a selective activation/suppression of cognitive components can reconcile the reality of internal states with the puzzling experimental data that have been cumulating over many years.