To date OR has no means of modelling, and therefore predicting the behaviour of knowledge in a system. Such knowledge bearing systems are ubiquitous, and include social networking structures (of increasing importance in politics and in marketing) and more conventional organisational structures (such as communities of practice). Taking into account the critical nature of knowledge production and dissemination as strategic issues for firms, this is a serious gap in our capability. This paper makes a contribution to our understanding of knowledge as an emergent property of a system. It offers a generalised theoretical model in a neat analytical algebraic form, derived from clearly stated axioms, in turn deriving from empirical work in the literature of knowledge management. Until recently, the prediction of knowledge dynamics has suffered from the lack of a credible knowledge measure. Without such a measure the dynamics of knowledge as a property of a network of communicating parties cannot be modelled. This has serious consequences for our understanding of knowledge dynamics because it does not allow us to understand the interplay between the important components of a knowledge network, such as the connective topology of the network, the characteristics of knowers, and above all the behavioural characteristics of knowledge itself. This paper shows how a simple, tractable and generalised analytical theory emerges from simple assumptions about the behaviour of knowledge as a connective property of a set of relevant concepts and about the knowledge capture, retention and onward transmission properties of the components of the interconnected system.