While various attempts have been made to link nature and society more closely together within environmental sociology, it now appears as though there is a general acceptance of rather traditional divisions between these two domains. Yet ecology specifies that natural and social entities are bound together in complex interrelations. Why then does sociology insist on sifting out the social from the natural? The paper takes this question as its starting point and seeks to identify what environmental sociology might gain and lose from a shift towards ecological thinking. It does so by examining the case of actor-network theory, an approach that, in significant respects, closely approximates a kind of ‘ecological sociology’. Actor-network theory is ‘co-constructionist’: it seeks to identify how relations and entities come into being together. Critics have focused on the problems of co-constructionism: they have argued that human actors generally possess powers of reflection (through language) and that these powers of reflection provide motive forces for action. Thus some form of social analysis is still necessary. Any ecological sociology will thus need to bring these two perspectives together so that humans and non-humans can be considered within the same frame of reference but so the distinctions that generally hold between the two can also be assessed.