Bermúdez says “Framing effects associated with quasi-cyclical preferences are likely to be found in decision problems that are sufficiently complex and multi-faceted that they cannot be subsumed under a single dimension/attribute/value” (target article, sect. 3.2, para. 8 (H2)).

The fact that animals show framing effects very similar to those of humans suggests that what appears to be irrational or suboptimal behavior may not require human complexity and that many framing effects may have biological/evolutionary bases that offer simpler accounts of the behavior.

In the simplest case one can find, framing can be found in the learning of simple associations. What is learned depends on one's frame, the discrepancy between what is expected and what is obtained (e.g., Rescorla & Wagner, Reference Rescorla, Wagner, Black and Prokasy1972). There is much learning early in training, much less learning later in training.

In a less obvious example, one can offer an animal a single pellet of food and give it one pellet, or one can offer the animal two pellets but give it only one. If now the animal is given a choice between the two offers, either of which results in one pellet, the animal will show a strong preference for the single pellet offered (loss aversion; Sturgill et al., Reference Sturgill, Bergeron, Ransdell, Colvin, Joshi and Zentall2021). On the other hand, animals have a strong preference for receiving more than expected – presenting them with one pellet but giving them two pellets, over presenting them with two pellets and giving them two pellets (gain attraction; Clayton, Brantley, & Zentall, Reference Clayton, Brantley and Zentallin press).

The delay-discounting framing effect mentioned by the author occurs when subjects prefer the smaller-sooner (SS) outcome over the larger-later (LL) outcome. Because we humans view self-control as an asset, we value choice of the LL. What we fail to recognize is that in the natural world in which humans and other animals evolved, generally, delayed rewards can rarely be guaranteed, so historically, impulsivity may be a trait that has been selected for. There is evidence, however, that one can change the frame of such a discrimination by requiring humans or other animals to make a “prior commitment” – an earlier response that makes the SS and LL appear to be closer to each other in time (Rachlin & Green, Reference Rachlin and Green1972). The reversal of preference from SS to LL can be attributed to the difference in the way the discrimination is framed, or as a natural consequence of the fact that delay-discounting functions are naturally hyperbolic, and when extended in time, the hyperbolic delay-discounting functions naturally cross over (Mazur, Reference Mazur, Commons, Mazur, Nevin and Rachlin1987).

Impulsive choice is thought to be the frame that makes choice of the SS over the LL suboptimal. It also appears to make the reverse contingency procedure (in which one must choose the smaller amount in order to receive the larger amount) difficult, even for chimpanzees (Boysen & Berntson, Reference Boysen and Berntson1995). If, however, one uses symbols to represent the quantities, impulsivity is reduced (converting what Bermúdez refers to as a hot representation into a cool representation), altering the frame, and resulting in optimal choice (Sturgill et al., Reference Sturgill, Bergeron, Ransdell, Colvin, Joshi and Zentall2021).

It has been found that humans can learn to modify the framing of a problem to reduce suboptimal choice. We might refer to this as learning to exert self-control by distracting oneself (Mischel, Ebbesen, & Raskoff-Zeiss, Reference Mischel, Ebbesen and Raskoff-Zeiss1972). But we have found that animals can do the same. In the choice procedure involving an SS versus an LL in which the SS is preferred, similar to humans, providing pigeons with an irrelevant response-independent alternative can result in pigeons that are less impulsive and that choose less suboptimally (Mueller et al., Reference Mueller, Peng and Zentallsubmitted). One might say that the irrelevant, response-independent stimulus “distracts” the pigeon from the SS reinforcer long enough to receive the LL reinforcer. Is that a framing effect or a learning effect?

Unskilled gambling, thought to be uniquely human, is another “irrational” behavior that appears to result from framing, but it too can be found in pigeons (Zentall & Stagner, Reference Zentall and Stagner2011). For example, pigeons prefer a signal for a 20% chance of getting 10 pellets of food (a jackpot) over the superior signal for a 100% chance of getting three pellets (no uncertainty). But perhaps the pigeons prefer the uncertainty of the 10 pellets. However, they also prefer the 20% occurrence of a signal for reinforcement over a signal for 50% reinforcement (now the optimal alternative is even more uncertain).

Another framing effect shown by humans is their (presumed cultural) bias to prefer outcomes that require more effort, over comparable outcomes that require less effort (e.g., for students an A grade in a difficult subject is often judged to have more value than an A grade in an easy subject). It might be proposed that this bias comes from the cultural adage that “work is its own reward.” Alternatively, humans may value hard work because it is reinforced socially, and those social reinforcers may even become internalized (become self-rewarding). On the other hand, there is considerable evidence that pigeons show a very similar bias (Clement, Feltus, Kaiser, & Zentall, Reference Clement, Feltus, Kaiser and Zentall2000; Zentall, Reference Zentall2010). If pigeons have to work hard for alternative A but not so hard for alternative B and they both result in the same reward, when given a choice between A and B, pigeons often prefer A (the one that they had to work harder for) over B.

All of this research suggests that one can see similar framing effects in animals that likely do not have the capacity for perspective taking or a theory of mind. However, in spite of these examples of suboptimal behavior, with relatively simple modifications of the animals' environment, one can alter the animal's frame to manipulate the optimality of its behavior. Furthermore, if one can easily manipulate the environment to result in changes in the animals' frame, it suggests that one should consider the possibility that similar simpler mechanisms may be involved in the modification of many human frames.