A cognitive system is a set of processes responsible for intelligent behaviour. This thesis is an attempt to answer the question: how can cognitive systems be demarcated; that is, what criterion can be used to decide where to draw the boundary of the system? This question is important because it is one way of couching the hypothesis of extended cognition – is it possible for cognitive systems to transcend the boundary of the brain or body of an organism? Such a criterion can be supplied by what is called in the literature a ‘mark of the cognitive’. The main task of this thesis is to develop a general mark of the cognitive. The starting point is that a system responsible for intelligent behaviour is a coordinated coalition of processes. This account proposes a set of functional conditions for coordination. These conditions can then be used as a sufficient condition for membership of a cognitive system. In certain circumstances, they assert that a given process plays a coordination role in the system and is therefore part of the system. The controversy in the extended cognition debate surrounds positive claims of systemhood concerning ‘external’ processes so a sufficient condition will help settle some of these debates. I argue that a Coordinated Systems Approach like this will help to move the extended cognition debate forward from its current impasse. Moreover, the application of the approach to social systems and stygmergic systems - systems where current processes are coordinated partly by the trace of previous action – promises new directions for research.