Security is a challenge in pervasive systems for several reasons. First, the large number or peers challenges the use of centralised security services like certificate authorities and reputation services. Second, a principal may be a physical object whose identity might not convey useful information for taking security decisions. Establishing a trusted channel depends more on a principal being able to demonstrate what is does, rather than who it is. Third, device mobility leads to the formation of ad hoc networks. Devices in these networks may be interacting for the first time so devices must carry sufficient information for them to establish mutual trust with other devices. This paper proposes a security model for pervasive systems, based on the idea of message quality. The model allows a principal to establish the intent of an adversary and to make the adversary prove its trustworthiness by furnishing proof of past behavior.