Introduction

When one individual is signalling, or two individuals are interacting, they do so within a network of potential receivers (see McGregor, 1993; McGregor & Dabelsteen, 1996; Ch. 1). As the other chapters in this book show, the decisions that both signallers and receivers make about their future behaviour are thus contingent not only on each other's behaviour but also on a wider network of individuals (McGregor & Peake, 2000). This view is finding support in empirical studies showing that individuals use information that could only be extracted from network interactions (e.g. Oliveira et al., 1998; Peake et al., 2001, 2002; Ch. 2). These empirical findings also have implications for the theoretical study of signalling strategies (e.g. Nowak & Sigmund, 1998; Johnstone, 2001). For example, an individual's signalling strategy may no longer be predicted solely from the responses of an opponent. We consider that the signalling strategies of individuals will only be explored realistically by models that include the potential responses of signallers to other individuals. In this chapter, we ask whether current modelling approaches can be adapted to include networks or whether new modelling techniques need to be considered.

The aim of creating a model is to advance our conceptual understanding of a system and create empirically testable hypotheses (Wilson, 2000; Hemelrijk, 2002) by simplifying the real world using words or mathematical expressions. Most hypotheses start with a verbal model and develop into mathematical models, which more precisely specify limiting conditions and assumptions and often provide a deeper understanding of the logic underlying the hypothesis.