One of the strangest aspects of Kahneman’s theory about human decision making is the idea that memory is systematically distorting our view of reality. This is the distinction between ‘experienced utility’ and ‘decision utility’. There are different versions of it, one is that in remembering experiences, we do not care for the duration, but only for the peak and the end of the flow of affectual reflections such as pain, or another is that we construct narratives of past events that are misleading, for example in terms of implicit estimates of underlying probabilities. Thus, memory has a core position in this theoretical edifice: All decisions are based on memory. Given this fact, one wonders why behavioural economics does not explore the role of memory much deeper.

According to many philosophers, psychologists and neuroscientists memory is the most essential condition for having a ‘self’. Kahneman approaches memory in a ‘dual selves’ context, which suggests that there is a self that is independent from memory. And indeed, this is the ‘Benthamite self’ that allegedly is the primary measuring point for the flow of quantifiable utilities, such as the flow of pleasures while staying in vacations. This flow is a sequence of instantaneous events that occurs in physical time. But is this just one ‘self’ that stands in an awkward relationship with our second self which is constituted by remembering this flow, and thereby distorting our decisions to maximize pleasures and minimize pains? I think that this view is fatally flawed because without memory there is no self.

There is another ‘dual selves’ approach in the history of psychology in which memory is also essential. This is George Herbert Mead’s distinction between ‘I’ and ‘me’. Mead is neglected in behavioural economics, although one could certainly argue that the discipline that he founded, social psychology, would be a natural partner for economics as a social science. In a nutshell, this distinction means that humans (different from animals) have the capacity to view themselves in the eyes of others (the ‘me’), and that the ‘I’ is not directly accessible for us in terms of propositional knowledge. That means, we experience the ‘I’ as the source of our actions, but we can only perceive it as mediated via the ‘me’. Mead offers a Darwinian explanation for this dualism in arguing that for the human species social cooperation has been fundamental for survival, such that sociality is an in-built feature of our existence. Interestingly, this theory shares many important features with Adam Smith’s ‘Theory of Moral Sentiments’ where the ‘spectator’ plays a paramount role in explaining action: The spectator is an inner instance by which we see ourselves as Meadian ‘me’.

In this approach, memory is omnipresent because it establishes a continuous feedback mechanism between ‘I’ and ‘me’. The simplest example that Mead gives is speaking: When we talk to others, we mostly do not have an anticipatory plan in the mind what to say next, but we speak spontaneously. Only once the spoken word is out, we realize what we said, and we continue based on remembering the previous, thus enabling us to create a consistent flow of speech. However, this means that even in the flow of micro-events such as utterances, we can only approach the ‘I’ as the origin of our speech as mediated via the ‘me’, namely, as a remembered ‘I’ seconds before. In this sense, the ‘I’ is a social construct, too, but it is not entirely immersed in the social, because it is the source of our creativity, such as when saying things that nobody had expected previously, including ourselves.

There are modern neuroscientific theories about the self that also highlight the role of memory, such as Damasio’s. Only memory establishes the unity of the self, and this includes both more basal media of remembering embodied in neurophysiological sensorimotor systems, and higher-level cognitive media which establish what Damasio calls the ‘autobiographical self’. I found it fascinating that Damasio use the term ‘narratives’ on both contexts, thus using the notion of a ‘non-verbal narrative’. He even uses the term ‘me’ as the “protagonist to whom certain events are happening”, and refers the non-verbal narratives to “images”, condensing this in the remarkable statement “The self comes to mind in the form of images, relentlessly telling the story of such engagements.” That means, there is a unified basis for all kinds of narratives, which is the fundamental neurophysiological structure of the self.

Considering this, I suggest that ‘narratives’ are mechanisms of remembering and hence the fundamental elements of memory, coming along in many different shapes and forms. Without narratives there is no self, and hence no agent capable for action. In the context of Mead’s approach, a narrative would be the causal feedback structure that maintains the consistency of the relationship between ‘I’ and ‘me’ through time. This mechanism is always grounded in both neurophysiological structures and external media, including our own externalized actions, as in the example of speaking. Narratives are highly idiosyncratic, because they accumulate though the lifetime of a human individual. Therefore, they are causal structures which cannot be described as universal regularity. Seeing narratives as mechanisms can resolve this potential tension between searching for causal explanations and recognizing the uniqueness of the ‘I’.

Mead’s ideas are succinctly expressed in the chapters 22-29 of his Mind, Self, and Society. The definitive edition. Chicago, London: University of Chicago Press, 2015. Damasio’s analysis is presented in his Self Comes to Mind: Constructing the Conscious Brain, New York: Pantheon, 2010, especially chapters 8 and 9.