Communication is a process in which senders provide information via signals and receivers respond accordingly. This process relies on two coevolving conventions: a "sender code" that determines what kind of signal is to be sent given the sender's state; and a "receiver code" that determines the appropriate responses to different signal types. By means of a simple but generic model, we show that polymorphic sender and receiver strategies emerge naturally during the evolution of communication, and that the number of alternative strategies observed at equilibrium depends on the potential for error in signal production. Our model suggests that alternative communication strategies will evolve whenever senders possess imperfect information about their own quality or state, signals are costly, and genetic mechanisms allow for a correlation between sender and receiver behavior. These findings provide an explanation for recent reports of individual differences in communication strategies, and suggest that the amount of individual variation that can be expected in communication systems depends on the type of information being conveyed. Our model also suggests a link between communication and the evolution of animal personalities, which is that individual differences in the production and interpretation of signals can result in consistent differences in behavior.