The term 'Social capital' is increasingly being used in economic analysis. It is generally presumed that social capital is just what it says it is, a social form of capital. That is, it is a metaphorically inspired form of capital. In this thesis, I reconsider the sense in which the phenomena in question are legitimately considered as capital. The existing literature draws together an array of disparate notions to explain the statistically significant role of inter-personal relationships for economic welfare. Yet this is done through ad hoc treatment of discordant concepts: methodological individualism is used to explain embedded network phenomena; the inconsistent observation of a 'dark side' to social capital is excluded from its definition.
In response, a new way of- conceptualising social capital is proposed here. This will make use of the framework of evolutionary microeconomics, which is a method of considering the specific structure and dynamics of interactions between agents. It is found that specific network structures produce the attributes of 'social capital' more efficient information flows and resolution of collective action problems. Trust and other common behavioural rules otherwise labelled 'social capital' become the subtext to linkage creation. Stylised facts of various scenarios (such as Russia's recent evolution) and specific studies (Ghanaian entrepreneurial networks) are analysed through the changing patterns of connection this network conception of social capital entails. The implications are broader. 'Social capital' is the means of managing complexity, creating the order required for legitimate forms of capital. It is thus an essential part of economic systems.