Ruth Byrne has provided a thorough and important analysis of the relationship between rationality and imagination (Byrne Reference Byrne2005). This is a very useful extension of more typical analyses of the nature of rational thought which often concentrate on the ability to give the right answer, and mostly neglect the ability to go beyond the problem parameters that is an important aspect of our intuitive understanding of imagination. As Byrne points out, people spend a great deal of cognitive energy in reflecting on what could have been in different situations and on the resulting possibilities. This, in itself, is not particularly new; the novel aspect of her analysis is the use of a framework that explicitly situates imagination within the more constrained focus of rational thought. In other words, the possibilities that people imagine are not simply generated at random, but are constrained by the nature of the cognitive processes that define rationality.

Nonetheless, I would claim that the relationship between imagination and rationality that is sketched out in this book is at least partially flawed, because it does not allow for any real development of either imagination or rationality. To make this point clear, I must start by simplifying what is a complex analysis. The essence of Byrne's argument is of course derived from mental model theory. This claims that people have standard semantic representations of reality that allow them to represent not only what is, but a subset of what is possible. These representations are essentially rational, because they are defined in such a way that, when they are applied consistently and used with the standard algorithms defined by the mental model theory, they will give responses to inferences that are either textbook ones or, in some more general sense, optimal. Possibilities that are generated must remain consistent with these semantic representations. This in turn allows for a fairly tight definition of imagination that constrains the possibilities implied by imagination to those that are consistent with the rational semantic representations that underlie the interpretation of relational terms, such as if–then.

This general model proposes an essentially static relationship which accounts for the influence of rationality upon imagination, but does not allow the inverse effect. The key problem is that possibilities are processed within the constraints imposed by whatever processes define rational thinking. This is a basic limitation, particularly when considered in a developmental context. One point of view, which implicitly underlies Byrne's analysis, is that the processes that define rational thought are essentially innate, and that there is no need to suppose any underlying developmental change. If this is not the case, then there has to be a mechanism by which the nature of these processes change. And, one of the key intuitive candidates must be imagination. In other words, if whatever constitutes rational thinking in adults is derived from a developmental process, then imagination might be considered to be one factor in creating more complex forms of thinking (and possibly more rational ones).

This is, in fact, the hypothesis that was put forward by Piaget (Reference Piaget1981) in a series of imaginative experiments examining the relationship between the range of possibilities that children of different ages were able to generate in a given context and the complexity of their cognitive processes. Older and cognitively more advanced children were able to generate a larger and qualitatively more diverse set of possibilities than were younger ones. My students and I have found similar results when looking at the relationship between different levels of conditional reasoning and the kinds of possibilities that are generated by children and adults (Janveau-Brennan & Markovits Reference Janveau-Brennan and Markovits1999; Markovits & Vachon Reference Markovits and Vachon1990; Venet & Markovits Reference Venet and Markovits2001). Most interestingly, these relate not only to the quantity of such possibilities, but also to their nature. In other words, if one looks at imagination developmentally, there is clear evidence of a qualitative shift that goes from more constrained, experientially based possibilities to more general and abstract forms of possibilities. Piaget (Reference Piaget1981) argued that the relation between the cognitive processes that determine reasoning and children's imagination was bidirectional. Although many possibilities are indeed implied, and limited, by the cognitive processes used by a given subject in a way that is consistent with Byrne's model, Piaget also claimed that some were generated because children observed or deduced possibilities that were in fact inconsistent with these cognitive processes. These possibilities become sources of disequilibrium that can only be made “rational” by adjusting the characteristics of the cognitive processes. In other words, this idea assumes that imagination can sometimes go beyond available cognitive processes and result in a reconfiguration of what is considered to be rational.

Direct evidence for any such process remains anecdotal, although anyone who has listened to a child work out a complicated problem will find it quite convincing. There are, however, examples of this kind of process in the development of science that make very useful analogies. For example, Newtonian mechanics postulated that velocities are linearly additive, which is of course a very intuitively rational concept. The Michelson-Morley experiment provided empirical data that was simply inconsistent with Newtonian theory (Michelson & Morley Reference Michelson and Morley1887). In other words, the results of this experiment were not possible within what was considered to be rational at that time. It was not until Einstein's special theory of relativity that an explanation of this result was (eventually) accepted: In this theory, velocities are not linearly additive, which has replaced Newtonian rationality. Einstein derived this theory, not by empirical work, but by a series of thought experiments, that is, by imagination, that allowed him to go beyond Newtonian rationality.

Thus, what I suggest here is that Byrne's analysis fairly represents a major part of the work done by the imagination, which is to examine possibilities in a way that is consistent with what a person's “rational” processes allow as being possibly true. However, it neglects the potentially critical role of the imagination in constructing possibilities that are not rational, but that suggest the necessity of revising our definition of what is rational.