At first glance, the book might seem like just another contemporarily-popular social network analysis dressed up in complexity language. The book presents the claim that chemistry concept autocatalysis is the explanatory model for the emergence and growth of organizations. The argument is that autocatalysis (the catalysis of a reaction by one of its products) is like the process of individuals acquiring skills which thereby transform products and organizations: “Skills, like chemical reactions, are rules that transform products into other products” (pp. 70-1). The process is reciprocal and ongoing as actors create relations in the short-term, and relations create actors in the longer-term.
One response of a critical reader would be asking the degree to which autocatalysis has explanatory power over the formation and persistence of organizations. In the absence of the consideration of other models, or the extent to which autocatalysis does not fit, it is hard to assess where this model falls on the anecdotal-to-accurate spectrum. This is a potential problem with all attempts, however valiant, to transplant the models and structures from one field to another. Going beyond interesting associations to correlations and even causal links is challenging.
Also not uncommonly, the authors postulate that the interesting, novel, and value-contributing aspects of a system (in this context, an organization) occur in the interstices, edges, and anomalies of the system. In actuality, this might be just one possibility (and not the principal element according to thinkers like Simondon for whom novelty most directly emerges from the central interaction of the components, features, and functionality). Worse, seeking the interstice forces the focus onto identifying borders, edges, and interstices, defining the phases of inherently [non-definable] dynamical systems. Also with a Simondonian eye, this is to miss the nature and contribution of dynamic processes at the higher level - this is trying to corral them into identifiable morphologies instead of apprehending their functionality.