Convention: An Interdisciplinary Study

Abstract

In our lives we are used to face situations in which something we all desire depends on the coordinated action of our peers. If we were to solve the practical problem of what to do in these situations from scratch, our daily life would be constantly at risk or even impossible. Luckily enough social life is full of patterns that we are able to discern in order to predict what the others will do, and that, at the same time, give us reasons to keep behaving as we have done before. Some of such regularities are what we call conventions, and to this ‘topos’ this special issue is devoted. But, what makes a social regularity into a convention? It has been 40 years since the first publication of David Lewis’ seminal essay on this topic (Lewis 1969), and his book has proved to be a source of inspiration for many practitioners in different fields. Lewis’ theory was developed as a response to Quine’s attack at the notion of analyticity in philosophy of language, and was intended also as a unifying approach to the study of ‘languages’ as formal semantic models, and ‘language’ as a historical and social phenomenon. Intuitively, it is by convention that a population uses a certain language instead of another. But to clarify this platitude, one needs a theory of what a convention is. The relevance of a theory of conventions of course spans well beyond the study of language. Language, in fact, is just one among many activities that are governed by conventions. Conventional regularities can be found at the core of morality, law, economics and in practically any of the daily activities we engage in. Nonetheless, a shared view on this issue is still missing. Notoriously, Lewis took advantage of game theory to approach the problem, and in doing so he also innovated game theory itself in ways that are nowadays being rediscovered (Cubitt and Sugden 2003). However, from the onset he also advised the reader that game theory was scaffolding, and he showed this to be true in his rejoinder to the problem few years after his first take (see Lewis 1975). In any case, the basic theoretical proposal was preserved so that conventions of a population of agents turned to be regularities in action (or alternatively in action and belief) which are arbitrary but perpetuate themselves because they serve some sort of common interest. Past conformity to such regularity gives everyone a reason to go on conforming in the present instance, so that the regularity is constantly re-produced. However, an alternative regularity would have been maintained in the same way if everybody conformed to it in the first place. This concise formulation, of course, hides many of the difficulties that need to be accounted for to fully understand the phenomenon. Nothing is said on how a specific convention originates in the first place (what is the appropriate evolutionary dynamics of a certain regularity of behavior?), on how the agents coordinate their private representations on the salient pattern to follow (what is the correct pattern between the many apt to describe the regularity? how does it happen that all agents in a population share the same representation of the situation and the regularity to follow?), of how much and what kind of knowledge of the convention each agent needs to dispose in order to conform (do conventions really need to be matter of common knowledge?), on the kind of reasoning, if any, that the agents endorse in order to decide to conform (is conformity to a convention justified by reasoning or do we conform L. Tummolini (&) Institute of Cognitive Sciences and Technologies, Via San Martino della Battaglia 44, 00185 Roma, Italy e-mail: luca.tummolini@istc.cnr.it