Vromen,
J. J. (2001). Review of The New Evolutionary Microeconomics: Complexity,
Competence and Adaptive Behaviour - Jason Potts.
Jason Potts's New Evolutionary Microeconomics is at once
an ambitious and imaginative piece of work. It aims at providing a micro-economic
foundation for all schools and traditions within heterodox economics. Anyone
familiar with the `family quarrels' within heterodox economics will immediately
grasp that this is no small feat. To achieve his aim, Potts brings together
various bits and pieces of notions, ideas, theories and models, dispersed
over many different disciplines. Instead of letting his readers get lost
in the bewildering complexities and perplexities of the various strands
of literature that he draws on, Potts gently takes his readers by the hand
and navigates them through the labyrinth. Potts is able to dothis, because
he found a key to connect all the strands. This key he calls the geometry
of economic space. Instead of the orthodox Walrasian notion of an integral
space (related the mathematical notion of a field), in which all elements
are connected to one another, Potts proposes a non-integral space as a
unifying notion for heterodox economics. Adopting a non-integral space
makes it possible to account both for particular interaction and for time
irreversibility. Replacing an integral space by a non-integral space implies
that the focus shifts from elements to the connections between elements.
Accordingly, what primarily evolves in Potts's evolutionary microeconomics
are connections between elements rather than the elements themselves.
This is an analysis at an extremely high level of abstraction, and Potts is aware of it. But already here, at this high level of abstraction, doubts begin to creep in. Orthodox economists often accused heterodox economists of not getting beyond putting forward the vague and unproductive notion that `everything depends on everything'. And, indeed, many heterodox economists have strong leanings toward a holistic, non-reductionistic type of approach. But if we are to believe Potts, then the accusation is more appropriately levelled against mainstream economists themselves. For the notion of integral space that Potts assigns to mainstream economics seems to imply the belief that every element is connected with all other elements in the field.
At a lower level of abstraction, I also found Potts's depiction of hetero economics, his own preferred notion of the micro-economic agent, questionable. Connections here are not defined interpersonally, but intrapersonally, as the cognitive technology an individual agent has at his disposal for combing his resources. Potts here explicitly invokes the image of Robinson Crusoe, an image that heterodox economists often have objected to. Only later on other individuals that the individual agent may interact with appear on the scene. Potts simply assumes that the logic by which agents decide whom to interact with is the same by which agents interact with the environment. This seems to run counter to how many heterodox economists view the relation between individual agents and social structure. These economists stress that the social environment in which agents are embedded affects the ways in which agents think, argue, behave and interact with one another. This is taken to imply that the social environment needs to be taken into account right from the start. Potts pays no attention either to cultural evolution, to cultural transmission of memes (or similar units of cultural transmission) and to how this affects human behaviour. Once again this calls into doubt whether Potts really comes up with a microeconomics that is acceptable to heterodox economists of all stripes.
This does not diminish my admiration for the accessible way in which Potts introduces quite complex theories and notions, however, and especially for the comprehensive and creative way in which he links these theories and notions with one another. Graph theory, complexity theory, chaos theory, self-organisation theory and the notions of genetic algorithms and dynamic efficiency all get a succinct, but clear coverage. Without eschewing some elementary technical matters altogether, Potts succeeds remarkably well in showing in an informal way how these theories and notions can complement one another. Potts makes clear why he believes that the future lies in multi-agent computer simulations. All this is done in vivid prose. Potts's book demonstrates that complex and highly abstract notions and issues can be discussed in an engaging and entertaining way.
The only hesitation I have here is that sometimes it seems that Potts
reads too much in mathematical formalisms. Potts sometimes argues as if
each mathematical formalism has unique ontological implications. I think
this is a bit naïve. Mathematical formalisms can be given many different
interpretations. And sometimes the one mathematical formalism can be translated
into another. For example, contrary to Potts's general point that the ontological
implications of the mathematical notion of a field are altogether different
from those of graph theory, it seems that a field can be described in terms
of graph theory (as Potts himself acknowledges; see note 4, p. 81). Minor
reservations as these notwithstanding, I warmly recommend Potts's book
to anyone who is in for an intellectually adventurous attempt to synthesise
various pioneering non-mainstream strands of theorising and modelling.