[Zurück]

G. Hanappi:
"The Concept of Choice - Why and how innovative behaviour is not just stochastic";
Vortrag: International Workshop "Innovation, Structural Change and Economic Development", Les Treilles, Tourtour, France (eingeladen); 16.06.2005 - 21.06.2005.

This paper starts with a remarkable observation of John von Neumann, namely that stochastic behaviour in economics assumes a radically different form from the probabilistic descriptions of non-living systems. It seems that this hypothesis is one of the major reasons that motivated von Neumann to develop a proper new formal language for the social sciences: game theory. Current mathematically oriented game theorists - often following some kind of Nash program - usually do not grasp what a radical break in economic theory production von Neumann and Morgenstern originally had in mind. Part 1 of the paper will unearth some bits and peaces of that perspective.

In part 2 the topic will be focused on one instance of strategic analysis that incorporates a probabilistic treatment of the behaviour of social agents in a classic way: mixed strategies. Of course, this concept is an elementary ingredient to von Neumann´s minimax theorem, and as such can be viewed as a prime example for a treatment of probability in the social sciences. Nevertheless even for von Neumann himself this treatment later seemed to be unsatisfactory. Therefore a simple game with mixed strategy Nash equilibrium is used in a simulation framework to disentangle some of the difficulties hidden by the strong assumptions usually made to arrive at analytical solutions. The (presumed) result of this part is that in an evolutionary setting the shares of population that represent entities playing pure strategies in an equilibrium mixed strategy evolves into internalized choice within each individual agent by the emergence of language (including internal and external memory). The simulation study enables some more detailed and sometimes surprising issues.

An old idea of Martin Shubik, namely to take a closer look at the categorization of environments as they are represented in payoff matrices, is further elaborated in part 3. If only some environments support evolutions that eventually lead to internalized knowledge, i.e. choice, then the dynamics of environments - how they enter and exit such knowledge enabling states - is an interesting question. Of course, this is a big issue that goes beyond the scope of this article. What will be presented are some links to the current discussion on how oscillations of the natural environment in the time domain (days, years) are translated into the spatial domain, and why it therefore seems reasonable to try to formalize social entities as pulsators. Another important consequence of this perspective is the necessity of a more evolutionary theory of knowledge implicitly needed. Again some possible routes that should be taken are sketched.

Finally a concluding part of the paper will discuss how the choice for self-organizing new knowledge, for innovation, comes about. In particular, innovation in science will be contrasted with innovation in production units - in the light of the arguments made in the preceding three parts of the paper.