Heard the one about the polimorphic action?

At a time when most discourse about human agency is becoming warm and fuzzy, it is perhaps surprising to find a book where the authors use two-by-two grids and phrases such as "instances of each of the four logically possible cases of action cascade" and "filling up the logical space we have created" to describe their ideas. Yet, at the core of this book is a very anti-reductionist agenda, which argues that there are many human actions that cannot be reduced to a set of context-free rules; actions that can only be successfully performed as the result of socialisation into a particular form-of-life.

This work is a further refinement of ideas that Harry Collins first presented in his book Artificial Experts (1990). Drawing on Wittgenstein's ideas and his own studies of the practices of scientists, Collins addressed the question of what separated distinctly human actions from those that could be performed by machines. The resulting discussion in that book about "regular action" and "behaviour-specific action", however, did not feel quite right, and this new book seems to provide a better way of addressing this distinction.

According to Collins and Martin Kusch, there are two main shapes of actions. Polimorphic actions depend on an understanding of society for their execution, while mimeomorphic actions can be achieved without this understanding. Joke telling is polimorphic, as there are no rules stating when it is appropriate to tell a joke (sometimes telling a joke at a funeral can be the most appropriate thing to do), whereas swinging a golf club is not dependent on society in this way and is a mimeomorphic action. A particular polimorphic action can be achieved in a number of ways, and an appreciation of the social situation is required for using the most appropriate one; in contrast we are indifferent to various alternative ways of implementing a mimeomorphic action.

This emphasis on the form-of-life underlying a polimorphic action enables distinctions to be made in areas that are traditionally often confused. For example, riding a bicycle can now be viewed as either mimeomorphic bike balancing (an embodied, not easily explainable action that can, however, be reduced to mathematical formulae) or the polimorphic riding of a bicycle in the middle of traffic (which requires a good understanding of the norms and conventions of fellow road users and cannot simply be reduced to a set of rules). These notions are then developed to differentiate between different kinds of sub-actions (using the grids and logical spaces mentioned earlier) and thus provide a clear vocabulary for understanding many kinds of human action.

The weakest section of the book is when Collins and Kusch apply their ideas to organisational issues. They introduce the concept of action cascades (for combinations of polimorphic and mimeomorphic actions) but it is sadly wasted in a short review of different organisational forms, which picks on rather obvious examples such as McDonald's and military drill. The problem is not that their analysis of these cases is wrong; it is simply that this area is so much richer than their examples suggest. This does, however, leave plenty of scope for other researchers to try to apply these ideas in the rich situations of organisational life.

I encourage anyone with an interest in human action to read this book. It is clearly written, with many examples used to illustrate the distinctions being made. It may not provide a final answer to questions of what it means to be human, but it does provide a means of stabilising the debate (and the vocabulary) in this area and enables the discourse to develop further.

Edgar A. Whitley is senior lecturer in information systems, London School of Economics.