Navigation – Plan du site

AccueilNuméros12-1Revue des livresComptes rendusJosé Luis Bermúdez, Frame it Agai...

Revue des livres
Comptes rendus

José Luis Bermúdez, Frame it Again: New Tools for Rational Decision-Making

Dorian Jullien
p. 143-149
Référence(s) :

José Luis Bermúdez, Frame it Again: New Tools for Rational Decision-Making, Cambridge: Cambridge University Press, 2021, 330 pages, 978-110719293-5

Texte intégral

Afficher l’image
Crédits : Cambridge University Press

1Imagine that 600 people have been contaminated by an unusual disease that will kill all of them very soon. Your job is to decide which of two treatments is going to be administered to those people. On the one hand, Treatment A will save 200 people. On the other hand, with Treatment B, there is 1/3 probability that the 600 people will be saved and 2/3 probability that no people will be saved. Decades of experimental research in psychology suggest that you are very likely to prefer Treatment A over Treatment B. The same studies also suggest that you are very likely to prefer Treatment B’, which gives 1/3 probability that nobody will die and 2/3 probability that 600 people will die, over Treatment A’, that will lead to the death of 400 people. However, A’ and B’ are the same treatments as A and B: they are just described differently—or presented in different “frames”. Since the work of Tversky and Kahneman (1981), preferring A over B and B’ over A’ is called a “framing effect”. Rational choice theory cannot represent a rational decision maker holding both a preference for A over B and a preference for B’ over A’. Indeed, on the account of Kahneman and Tversky, the influence of frames on preferences is not only pervasive in everyday life but also utterly irrational.

2José Luis Bermúdez disagrees with the last part of this account, i.e., that framing effects are necessarily instances of irrational behavior. His main goal in Frame it Again is to argue that not all framing effects are a sign of irrationality of a decision maker. Furthermore, according to Bermúdez, framing effects point to a set of psychological mechanisms that are necessary to make a rational decision when confronted with complex problems. His task is therefore to broaden rational choice theory with new tools that allow for a representation of rational framing effects. In my opinion, Bermúdez provides a successful constructive criticism of rational choice theory that could motivate fruitful work from economists, psychologists and philosophers. Furthermore, the new tools that he offers could also be useful to critically scrutinize the use of rational choice theory as a norm in behavioral public policies.

3Bermúdez lays his argument throughout ten chapters. The first one gives a broad interdisciplinary perspective on framing effects by discussing especially how they are studied in psychology, economics and communication studies. Bermúdez does not take issue with the interpretation of framing effects, such as the one presented above, as instances of irrationality. However, he does take issue with the widespread use of rather simple decision situations in these disciplines. More complex decision problems involving framing effects, he argues, would reveal some principles of rational reasoning that could enrich rational choice theory. One of the running examples throughout the book comes from Greek tragedy and is presented in the first chapter (6-8). Bermúdez discusses the decision faced by Agamemnon in Aeschylus’s eponym tragedy, who has to choose between sacrificing his daughter to the goddess Artemis, which will allow the Greeks to avenge themselves against Troy, and not sacrificing his daughter, which will prevent the Greeks to avenge themselves against Troy. Bermúdez interprets the description of Agamemnon’s reasoning as an instance of a framing effect:

There is a single option … that Agamemnon frames in two different ways – as Murdering his Daughter, on the one hand, and as Following Artemis’s Will, on the other. His alternative is Failing his Ships and People (by refusing to make the sacrifice). (7)

4Agamemnon knows that Murdering his Daughter and Following Artemis’s Will is the same option. Yet he simultaneously prefers Failing his Ships and People over Murdering his Daughter and Following Artemis’s Will over Failing his Ships and People. In a nutshell, Bermúdez’s main argument is that the capacity of seeing a given problem under different perspectives is helpful to make rational choices in complex decision problems, even if it generates conflicting preferences.

5The next two chapters propose a careful review of framing effects in psychology and economics, either observed in laboratory experiments (chapter 2) or in the field in the context of decision making on financial markets (chapter 3). These two chapters constitute a good introduction to framing effects for readers who are not acquainted with the behavioral economics literature on the topic. They contain clear explanations of the problems posed by framing effects to economic theory. They even present some subtleties that are sometimes forgotten in economics (and often ignored by philosophers) – such as pointing out that framing effects are not about a change in the agent’s risk attitudes across different domains of decision, which is not necessarily irrational, but about a change in risk attitudes for a given decision problem within the same domain (35). Bermúdez also carefully presents evidence of framing effects in the financial markets and insists that, if the monetary losses that ensue are caused by the decision frames imposed on agents, the solution to avoid such losses is often to impose different frames in order to, as it were, create the opposite framing effects. This nicely illustrates how frames can be seen as helpful tools to make rational decisions (in the sense that a specific frame can sometimes yield a more rational decision than another one). My only complaint with Bermúdez’s review of the literature is that he takes the classical explanation of most framing effects through loss aversion (58) at face value (e.g., decision makers are willing to take more risks in a frame that emphasizes losses rather than gains in order to avoid the losses), without mentioning that the empirical robustness of this concept is controversial (see, e.g., Gal and Rucker, 2018).

6Chapter 4 investigates both the psychological explanation of framing effects from Kahneman and Tversky’s (1979) prospect theory and why these effects are judged as irrational through the lenses of rational choice theory. The discussion of the explanation from prospect theory is original. Bermúdez reminds us that the first (1979) version of prospect theory introduces a concept of “editing phase” that can be interpreted as the process by which an agent constructs a frame through which he or she will then make a decision. In my opinion, this interpretation of prospect theory lacks a discussion of the second, so-called “cumulative”, version of the theory (Tversky and Kahneman, 1992). In that version, the “editing phase” is replaced by a “framing phase”, which, despite the terminology, gives much less importance to the notion of frames in prospect theory. The assessment of why framing effects are judged as irrational from the perspective of rational choice theory is also original. Bermúdez distinguishes very clearly between two arguments for the irrationality of framing effects. One argument is technical: the set-theoretical foundations of rational choice theory imposes as a requirement of rationality that different values cannot be attributed to two different descriptions of the same element (i.e., of the same alternative or of the same action). This requirement is often called the principle of “extensionality” (or invariance) in the literature on framing effects. The clarity of Bermúdez’s discussion of this point deserves to be noted, even though, in the end, he recognizes that this is not a very compelling argument from a normative point of view (80). The other argument is more traditional: the preference relation that is expressed in a framing effect is not transitive and therefore can lead the agent to his or her ruin (from the classical Dutch Book argument). The discussion of these normative issues is again quite clear, even if they could have been enriched by a broader discussion of the consequentialism of rational choice theory which partly grounds the qualification of framing effects as irrational.

7The next two chapters (5 and 6) are where Bermúdez presents his main argument. He first argues in chapter 5 for the possible (i.e., not necessary) rationality of a specific type of preference relation that do not respect the principle of extensionality and which he labels quasi-cyclical preferences: “a rational agent can prefer A to B ad B to C, even though they know full well that A and C characterize the same event under different frames” (100, my emphasis). According to Bermúdez, quasi-cyclical preferences are useful to deal with complex decision problems which need to be assessed through different frames before one makes a choice. He adds three examples of such complex decision problems to the one with Agamemnon and insists on not applying quasi-cyclical preferences to the simple framing effects discussed in chapter 2 and 3. He justifies this restriction for quasi-cyclical preferences with “anecdotal evidence” (98) that when people realize that they face the same decision problem in two different frames, they change one of their preferences to resolve their internal conflict (understood intuitively as having simultaneous and contradictory preferences). In contrast, making a choice with quasi-cyclical preferences do not imply a resolution of the internal conflict. It only implies that after careful consideration, one frame provides more convincing reasons than other frames, which leads to a choice that can later be remembered with regret or emotions due to the persistence of the internal conflict. Bermúdez goes further in chapter 6, where he argues that rationality imposes informational requirements on decision makers, which in turn justify the rationality of quasi-cyclical preferences. He calls the main informational requirement that he proposes the “Due diligence requirement”:

In setting up a decision problem, rational decision-makers need to be appropriately sensitive to as many potential consequences of the different courses of action available to them as possible. (121)

8And in order to be sensitive to different consequences of a given course of action, one has to scrutinize it under different frames. In other words, rational decision processes according to Bermúdez involve an informational comparative structure which consists in comparing frames. He insists in this chapter on the fact that different frames trigger different emotional responses, which can rationally ground different valuations of a given course of action, provided that the different frames do not generate contradictory beliefs in the decision maker’s mind (though he later weakens this provisio).

9In the next three chapters (7-9), Bermúdez tries to show how his conceptual apparatus can propose new solutions to well known problems in decision theory and game theory. In chapter 7, he argues that quasi-cyclical preferences provide a new perspective on self-control problems as discussed in decision theory. Oversimplifying a little, the problem traditionally consists in explaining how some agents resist the temptation of a small reward now in order to get a larger one later. On Bermúdez’s account, an agent can simultaneously prefer, on the one hand, the small reward now over the larger one later and, on the other hand, successfully resisting the temptation (i.e., another framing of the larger reward) over the small reward now. This quasi-cyclical preference can therefore explain how self-control problems can be solved by reframing the large reward, without suppressing the internal conflict constitutive of self-control problem altogether. In a sense, Bermúdez’s solution to self-control problem is very close to Ainslie’s (1992), who also explicitly uses the notion of framing. A discussion of the differences between the two accounts would have been most welcome, especially given that Bermúdez references Ainslie’s work at the beginning of the chapter (without discussing the latter’s solution to self-control problem in terms of framing).

10Chapters 8 and 9 switch to game theory. Bermúdez first presents Bacharach’s way of solving many games in which playing the dominant strategy leads to Pareto-inferior equilibria (chapter 8). Bacharach’s solution consists in pointing out that standard game theory forces players to adopt an “I-frame” in which they focus on individual payoffs. However, according to Bacharach, as matter of psychological fact, players sometimes adopt a “We-frame” in which they focus on joint payoffs. In chapter 9, Bermúdez tries to explain the process behind the psychological fact of players adopting a We-frame. His account rests on a notion of fairness that can underlie both the I-frame and the We-frame: I might want to maximize my own payoff because I think this is fair for me, and then realize that this notion of fairness requires that I look at the joint-payoffs. As the fairness of joint-payoffs is more salient in the We-frame, this can lead to a quasi-cyclical preference for a given joint-payoff evaluated from the We-frame over that same joint-payoff evaluated from the I-frame (211). Bermúdez demonstrates that one advantage of his theoretical account over Bacharach’s is that reasoning in the We-frame does not have to focus primarily on Pareto-superior equilibria to attain them (it is a theoretical consequence and not a hypothesis anymore that players coordinate on Pareto-superior equilibria in the We-frame).

11The tenth chapter is an application of Bermúdez’s conceptual apparatus to what he calls “discursive deadlocks”, i.e., situations in which two persons with strongly opposed opinions on a complex public issue cannot seem to have a rational discussion about it. Bermúdez’s main point here is that, while most of the time such debates take the form of a clash of values, they are often, in fact, a clash of frames about the same value. Realizing this is therefore key to get a rational discussion started in these situations because “frames are more malleable than values” (227). It is in this context that Bermúdez relaxes the provisio that it is rational to adopt different frames on a given decision problem as long as it does not lead decision makers to hold contradictory beliefs. More precisely, this provisio is restricted to beliefs about factual matters (e.g., does a one-week old fetus has a heart?) but it does not apply to other types of beliefs about, say, moral matters (e.g., is a one-week old fetus a person?). Being able to hold contradictory non-factual beliefs when seeing a problem from different frames is supposed to help comparing one’s point of view with another person’s opposite point of view, while not being able to hold contradictory factual beliefs is supposed to help changing one’s mind in a debate.

12Finally, this kind of frame-sensitive reasoning requires one to be able to reason about frames and not only from frames. The last chapter draws from a set of resources from psychology to explain how one is supposed to go about in order to reason in this way. This chapter is not only quite convincing in itself, it also helps the reader to realize how the normative ideal of a rational debate presented in the previous chapter is more empirically attainable than what he or she might have thought at first sight (I, for one, have never witness a debate in which one participant explicitly changes his or her mind about the topic at stake, at least during the debate).

13To conclude, Bermúdez provides a new set of tools for rational choice theory and shows how they can solve problems not only in decision theory and game theory, but also in deliberative democratic debates. I found most of the arguments rather convincing even if I felt that they could have gained from discussing the few points I raised above. My main complaint is about Bermúdez’s claim that his conceptual apparatus does not apply to simple experiments on framing effects such as Tversky and Kahneman’s (1981) unusual disease experiment. On the one hand, I personally feel that my own behavior in this kind of experiment involves a conflict that quasi-cyclical preferences describe rather well. On the other hand, LeBoeuf and Shafir (2003) showed that when people are made aware of their inconsistencies they indeed change their choice in one frame (thus providing more than anecdotal evidence for Bermúdez’s claim). However, people tend to change their choices in the frame that was presented to them in second round, in order to be consistent with the choices made in the frame that was presented to them in first (i.e., regardless of whether the first and second frames are the “die” or “saved” frame). This observation does not fit very well in the standard framework of decision theory. Therefore, applying Bermúdez’s conceptual apparatus to such cases would have been, in my opinion, a nice contribution to rational choice theory.

Haut de page

Bibliographie

Ainslie, George. 1992. Picoeconomics. Cambridge: Cambridge University Press.

Gal, David, and Derek D. Rucker. 2018. The Loss of Loss Aversion: Will it Loom Larger than its Gain? Journal of Consumer Psychology, 28(3): 497-516.

Kahneman, Daniel and Amos Tversky. 1979. Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2): 263-291.

Kahneman, Daniel and Amos Tversky. 1981. The Framing of Decisions and the Psychology of Choice. Science, 211(4481): 453-458.

Kahneman, Daniel and Amos Tversky. 1992. Advances in Prospect Theory: Cumulative Representation of Uncertainty. Journal of Risk and Uncertainty, 5(4): 297-323.

LeBoeuf, Robyn A. and Eldar Shafir. 2003. Deep Thoughts and Shallow Frames: On the Susceptibility to Framing Effects. Journal of Behavioral Decision Making, 16(2): 77-92.

Haut de page

Pour citer cet article

Référence papier

Dorian Jullien, « José Luis Bermúdez, Frame it Again: New Tools for Rational Decision-Making »Œconomia, 12-1 | 2022, 143-149.

Référence électronique

Dorian Jullien, « José Luis Bermúdez, Frame it Again: New Tools for Rational Decision-Making »Œconomia [En ligne], 12-1 | 2022, mis en ligne le 01 mars 2022, consulté le 10 août 2022. URL : http://journals.openedition.org/oeconomia/12064 ; DOI : https://doi.org/10.4000/oeconomia.12064

Haut de page

Auteur

Dorian Jullien

Université Paris 1 Panthéon-Sorbonne, Centre d’Economie de la Sorbonne. Dorian.Jullien@univ-Paris1.fr

Articles du même auteur

Haut de page

Droits d’auteur

CC BY-NC-ND 4.0

Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0

https://creativecommons.org/licenses/by-nc-nd/4.0/

Haut de page
  • Logo Association Œconomia
  • DOAJ - Directory of Open Access Journals
  • Revue soutenue par l’Institut des sciences humaines et sociales du CNRS
    CNRS - Institut national des sciences humaines et sociales
  • OpenEdition Journals
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search