Bermúdez argues that the objects of preference are framed outcomes, so that the frame is an ineliminable part of the objects under consideration. Further, this feature of preference is often a very good thing – frames highlight some dimensions and downplay others, and without them, many decisions might seem impossibly difficult to make. We think this is a wise and important insight, and it echoes points made by Keys and Schwartz (Reference Keys and Schwartz2007), who further emphasized that what people experience after making decisions are also framed outcomes. In Keys' and Schwartz's terminology, the decision process “leaks” into the subsequent experience of the decision's results (see also Frisch, Reference Frisch1993; Kahneman & Tversky, Reference Kahneman and Tversky1984). This “leakage” can be complicated, as in many of the “complex and multi-faceted” cases on which Bermúdez focuses, but it can also be simple, as when a “95% fat free” yogurt tastes less rich than a yogurt labeled “only 5% fat” (Sanford, Fay, Stewart, & Moxey, Reference Sanford, Fay, Stewart and Moxey2002). Either way, “leakage” means that frames shape not just decisions but also experiences in substantive ways.

When framing “leaks” into experience, apparently inconsistent decisions made under different frames may actually be consistent with the experience of the decisions themselves. Most approaches to rationality seem to assume that heuristics and biases exert their effects while a decision is being contemplated, but that once the decision is made, experiencing the results of the decision will be “path independent.” The object of a decision will be experienced on its own, carrying no trace of how the decision was arrived at (but see Kahneman, Reference Kahneman1994). Gilbert and Ebert (Reference Gilbert and Ebert2002) compared this “illusion of intrinsic satisfaction” to the perceptual illusion of direct access to the world, describing the tendency to behave as if “hedonic experiences were due entirely to the enduring intrinsic properties of [decisions’] outcomes” (p. 511). But this variant of “naïve realism” is mistaken, as Bermudez and Keys and Schwartz (Reference Keys and Schwartz2007) argue.

The ideas explored across many examples by Bermúdez have broad and deep implications for thinking about the rationality of decisions beyond cases focused on how options are described or otherwise framed. Research on choice overload, for example, has shown that people may feel worse about their choice if it was made from a larger assortment of options, even if they would have made the same choice from a smaller assortment (Iyengar & Lepper, Reference Iyengar and Lepper2000; Schwartz, Reference Schwartz2016). And people in poverty may be more likely to pursue government assistance after reflecting on aspects of their identity they are proud of, even though the amount of their benefits is no different (Hall, Zhao, & Shafir, Reference Hall, Zhao and Shafir2014). In the first case, regrets about the losses from tradeoffs among a larger number of alternatives “leak” into the prospective consumption of the chosen option. In the second, affirming the self prevents the potential stigmatization of accepting government benefits from “leaking” into the experience of receiving them. In both cases, factors that might seem outside the scope of rational deliberation obviously matter in a substantive, though perhaps non-normative, way.

What Bermúdez calls “quasi-cyclical preferences” and what we call “leakage” imply that there are important limits on what formal principles of rationality can tell us, since there are surprisingly few cases where the formal principles apply in their strictest form. Even seemingly inconsequential changes to the situation may “leak” into experience, affecting one's subjective outcomes and hence the reasonableness of different choices. Formal rules of rationality may allow researchers to draw important normative conclusions based on minimal, widely accepted structural claims about rationality. However, once the importance of subjective experiences and the prevalence of “leakage” are taken into account, it becomes clear that much more needs to be known before anything approaching a satisfactory theory of rationality is in hand. What is needed is a substantive theory of rationality – a theory that considers the content and not just the structure of decisions, and that evaluates that content in light of the decision-maker's goals, experiences, and life as a whole.

A substantive theory of rationality must consider the consequences of decisions, very broadly construed. It must consider short- and long-term consequences, consequences to the self and to others, consequences that are central to the decision at hand and those that may be peripheral. It also takes seriously the notion, raised by Bermúdez as well as Schwartz (Reference Schwartz1994), that there is no such thing (in the real world) as a decision in a vacuum. Frames, prior experiences, incidental environmental cues, and more will always be present in some form, and their effects may often “leak” into experiences. It seems irrational (and silly) to pay more for 7UP in a yellow-green bottle than for the same drink in a green bottle if the standard of rationality assumes the 7UP is experientially the same regardless of its container (Gladwell, Reference Gladwell2005). But a more psychological and practical view of rationality might point out that paying more for a drink that tastes different – in this case, more like a mix of lemon and lime – may be reasonable if it better satisfies a person's tastes.

What we suggest is that the entire field of judgment and decision making has to a large degree answered questions about how well our decisions conform to formal principles of rationality instead of questions about how well our decisions serve substantive rationality (Keys & Schwartz, Reference Keys and Schwartz2007). The field is not necessarily confused about what questions are being asked, though perhaps it is somewhat confused about the questions that fundamentally matter for a full understanding of decision-making as it is actually experienced in everyday life – the questions that the rest of us want answered. The main reason for this, we suspect, is that the “real” questions are much harder. But we hope that Bermúdez's many insights will fuel a broader shift in how we as a field think about not only framing, but also rationality more broadly, spurring us address the “hard” questions as best we can.