ABSTRACT
A new paradigm Whereas Nash-style game theory has remained firmly rooted in a (sophisticated) version of the neoclassical paradigm, complexity economics does indeed appear to have inaugurated a new, genuinely post-neoclassical paradigm. As we saw in the previous chapter, one of the main reasons for this radical novelty is that, notably under the influence of scholars like Herbert Simon and also thanks to significant improvements in computer technology, complexity economics has been able to inaugurate a truly “bottom-up” way of modeling. This has become known as agent-based modeling and is being used in “generative social science” to understand the emergence of economic phenomena out of the real-time interactions1 between boundedly rational individuals. Replacing M.Eq as “system at rest” with M.Eq as emergence, and thus decoupling M.Inst(s) from the Nash requirements and accepting a truly autonomous, adaptive rationality – these two new features mean that the frontier of mainstream economics has now definitely left p = NC and has established itself in p = post-NC. Individuals endowed with “recognizably human” bounded rationality practice two things simultaneously in their interactions: they apply their substantive rationality in order to satisfice within the environment in which they are immersed, and they apply their procedural rationality in order to learn about that environment. This means that within an objective – but, to the agents, unknown – environment, economic phenomena emerge out of the interaction between agents who hold a subjective “model” of their environment: each objective environment and the phenomena that emerge in it are a function of agents’ subjective model-beliefs about that environment. Whether the false subjective belief-models will all in the end, through inductive learning, converge on the true objective reality (which is itself a constantly emerging function of all these subjective belief-models) is a deep and unsolved question. But we saw that the legitimacy of most equilibrium notions used in economics hangs on what answer we give to this question. A crucial feature of the new, post-neoclassical paradigm is that the M.Inst axiom is no longer functionally subordinated to the M.Eq axiom. On the contrary, what equilibria will emerge from interactions cannot be predicted from any
requirement of Nash equilibrium. Emergence means order (the weakly functionalist version of equilibrium) but it does not mean a situation where every agent has reached his global optimum given the global optima of all other agents. Satisficing interactions mean that there is room for an individual’s desire to change and for an individual’s exploration of the social network. However that gradual process of exploration, learning, and revision takes place, one thing is certain: post-neoclassical economics is genuinely individualistic, in the sense that the basic building block of the analysis is now individual instrumental rationality and no longer the search for (strong) equilibrium. This implies that individuals’ procedural as well as substantive rationality have to be studied for their own sake, in order to be able to construct the “right” agent-based models of emergence. Important issues arise. How exactly is rationality “bounded?” What is the meaning of “satisficing?” How does our “recognizably human” cognitivebehavioural structure impact on the way we reach our decisions? As we witnessed when we discussed Simon’s automaton-based approach, these questions require that the Upper-Right quadrant be opened up and integrated into economics. The post-neoclassical paradigm seeks to be a wholly Right-Hand approach, and in fact aims to base its Lower-Right analyses on Upper-Right knowledge so as to build a clear basis for its agent-based strategy. Not surprisingly, therefore, two closely related developments have coincided with the emergence of the new paradigm: on the one hand, theoretical work on economic psychology, behavioral economics, and so-called “neuroeconomics”; on the other hand, empirical work in experimental economics. These areas of economic research have literally exploded since the 1980s. As with the other domains of innovation, we will not be able to offer here an exhaustive account of the whole diversity of approaches and results. The aim, rather, is to understand how post-neoclassical economists have tried to construct their specific brand of Right-Hand reductionism – which they rightly view as a “broadening” compared to the previous focus on exclusively Lower-Right issues, but which is still a strong reductionism in view of the AQAL approach we have adopted here as our background.
