6
INTRODUCTION HOMO OECONOMICUS IS ALSO HOMO COGITANS HOMO COGITANS IS ALSO HOMO OECONOMICUS For quite some time now, the science of economics has been based on the hypothesis of a homo ~economicus. Yet the postulate of an optimizing rationality, which economics generally takes for granted, constantly runs up against the large variety of descriptive models and representations that economic agents in fact develop as they reason about the problems they perceive. At the same time, in the cognitive sciences and in research on artificial intelligence, an effort is made to understand how homo cogitans constructs representations of his knowledge and reasons about these representations. Should not homo ~economicus be conceived of as a homo cogitans? Recent economic theories move in this direction when they take into account the agents' mental representations. Muth and Lucas take a first step when they introduce the agents' expectations into their theories, extending to these expectations the postulate of optimizing rationality, which classical theory deems valid for all actions. A second step in the same direction is taken by Stigler with his search theory, for this author notes that calculations and information searches have a cost, and that this cost must be paid if there is to be an optimizing of rationality. Yet once more, any decision to put a half to these calculations or searches is itself said to be dictated by an optimizing rationality. Only with Simon's theory of bounded rationality is the postulate of optimizing rationality finally rejected. Although Muth and Simon have worked and published together in the same areas, their theories diverge, the former moving toward rational expectations, the latter toward bounded rationality. For Simon, the foundation of the theory of bounded rationality resides in the findings of cognitive psychology, which have shown that our cognitive capacities are limited. Thus homo Theory and Decision 27 (1989) 1-6. 1989 Kluwer Academic Publishers. Printed in the Netherlands.

Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

Embed Size (px)

Citation preview

Page 1: Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

I N T R O D U C T I O N

H O M O O E C O N O M I C U S IS A L S O

H O M O C O G I T A N S

H O M O C O G I T A N S IS A L S O

H O M O O E C O N O M I C U S

For quite some time now, the science of economics has been based on the hypothesis of a homo ~economicus. Yet the postulate of an optimizing rationality, which economics generally takes for granted, constantly runs up against the large variety of descriptive models and representations that economic agents in fact develop as they reason about the problems they perceive.

At the same time, in the cognitive sciences and in research on artificial intelligence, an effort is made to understand how homo cogitans constructs representations of his knowledge and reasons about these representations.

Should not homo ~economicus be conceived of as a homo cogitans? Recent economic theories move in this direction when they take into account the agents' mental representations. Muth and Lucas take a first step when they introduce the agents' expectations into their theories, extending to these expectations the postulate of optimizing rationality, which classical theory deems valid for all actions. A second step in the same direction is taken by Stigler with his search theory, for this author notes that calculations and information searches have a cost, and that this cost must be paid if there is to be an optimizing of rationality. Yet once more, any decision to put a half to these calculations or searches is itself said to be dictated by an optimizing rationality.

Only with Simon's theory of bounded rationality is the postulate of optimizing rationality finally rejected. Although Muth and Simon have worked and published together in the same areas, their theories diverge, the former moving toward rational expectations, the latter toward bounded rationality. For Simon, the foundation of the theory of bounded rationality resides in the findings of cognitive psychology, which have shown that our cognitive capacities are limited. Thus homo

Theory and Decision 27 (1989) 1-6. �9 1989 Kluwer Academic Publishers. Printed in the Netherlands.

Page 2: Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

2 INTRODUCTION

oeconomicus inherits some of the most important characteristics of homo cogitans.

Within the framework of the theories of rational expectations and bounded rationality, there is no longer a single solution for the agent to find, but a set of possible solutions. This degenerescence- in the mathematical sense of the term - leads to the arbitrariness of choice. Consequently, the economist has to confront a new question: what are the mechanisms that make it possible to limit this arbitrariness? One line of research amounts to interpreting the birth and maintenance of institutions (in a broad sense of the latter term) as a reaction to the necessity of limiting the arbitrariness of choices. This inversion of perspectives has another consequence for the economist, for he or she can no longer sit back in an armchair, postulating that economic agents will engage in a rational and uniform rational behavior. Instead, the economist must do empirical research into effective decision proce- dures, or in other words, into the agents' procedural rationality (Simon, 1976). 1

On the other hand, should not homo cogitans be conceived of as a homo oeconomicus? Although at the present time the reflections on this topic are relatively disparate, they seem ample enough to make possible a genuine program of research. By emphasizing the limits of the cognitive capacities, homo cogitans makes cognition a scarce resource, a resource that homo aeconomicus can teach him to manage. One can, for example, raise the issue of the economic function of symbols. Many American psychologists (Miller, Anderson, Simon, etc.) are interested in the mechanisms of chunking by means of which a new symbol is created as the combination of other symbols and then assimilated to long term associative memory. Piaget's theory of the formation of schemata describes essentially the same process. Some authors, such as Rosh, speak quite explicitly of an economy of cognition in regard to such mechanisms.

With Varela, the economic function of the symbol shifts toward the needs of intersubjective communication. He considers a symbol to be an abbreviation of a long chain of operational explanations. Given this definition, we readily understand both the arbitrary nature and economic role of symbolic descriptions. On the one hand, the symbol's arbitrariness arises from the large number of possible abbreviations.

Page 3: Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

I N T R O D U C T I O N 3

Any limitation of this arbitrariness presupposes the existence of sci- entific communities, or more generally, of communities of labor, that would be based upon, and constantly renew, a scientific or cultural tradition. On the other hand, as abbreviations, symbolic descriptions are economically necessary to the process of intersubjective communi- cation within these communities. The economy of cognition that emerges here is situated then, not at the level of the individual, but at the one of the community.

We will not take up here other essential developments having to do with the nature of the symbol and the conditions of its emergence. On the former, the reader may consult, for example, von Foerster (1977); on the latter, Varela (1979) and Smolensky (1986). The articles by Le Moigne and Dupuy in the present issue also shed light on these issues. But the essential point in the present context is to make clear the role of the symbol in an economy of cognition, thereby conceiving of homo cogitans as a homo oeconomicus.

At this point in the argument, we see that economics becomes one of the cognitive sciences discussed in the Sloan report (1978), along with cognitive psychology, linguistics, artificial intelligence, neuroscience and epistemology. All of these disciplines are far from making up some kind of unified scientific domain, but they presently display some very strong interactions, out of which are being born new scientific dis- ciplines having their own research programs, such as psycholinguistics, neuropsychology, etc. We have seen above that economics has inter- acted with such disciplines as artificial intelligence and cognitive psy- chology; a more detailed study would bring forth its relationship to the other disciplines as well.

Making this interaction between economics and the other cognitive sciences emerge was one of the goals of a first international conference CECOIA 1. 2 The initiative was carried forward in a colloquium held on December 2, 1987 in the context of a research group of AFCET ("Association Franqaise pour la Cybern6tique Economique et Tech- nique"). The articles in this special issue are drawn from among the papers presented at this gathering.

The article by Walliser provides a distinction between "instrumental rationality" concerned with consistency between available means and pursued objectives, and "cognitive rationality" concerned with con-

Page 4: Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

4 INTRODUCTION

sistency between current beliefs and detained information. The reader will find an overview of economic theories which take into account the agent's mental representations.

The articles by Dupuy and Orlran make explicit, from the econom- ist's perspective, the role of public knowledge in the emergence of conventions (Dupuy), and the role of speculative movements (Orlran). Dupuy notes that the construction of the operator CK ("Common Knowledge") is equivalent to the existence of the Other, the only agent having the following property: "the Other knows p if and only if everyone knows that the Other knows p." This starting point, borrowed from epistemic logic, leads Dupuy to very rich interpretations of various paradigmatic examples in which the infinite specularity of the CK operator ("I know that you know that I k n o w . . . " ) is transformed into a null specularity ("I know that the Other knows").

The article by Munier concerns the contributions of cognitive science to the taking into account of uncertainty, one of the major themes of economics. He notes that the axiom of independence and the hypoth- esis of expected utility, that economists classically postulate in order to deal with uncertainty, are invalidated by various experiments con- ducted by psychologists and economists. With these experiments as his basis, he formulates certain very general rules that could prove to be pertinent to the development of expert systems for helping make decisions in a context of uncertainty.

In the article by Le Moigne, we find another recognition of the "non-optimizable optimum," and thus, of the importance of delibera- tion in human affairs. Here the focus is on the way in which economic reasoning is effectively conducted. This provides a transition to the articles that attempt simulations of effective economic reasonings.

Pinson proposes a multi-expert system of risk evaluation governed by a meta-model. This meta-model includes, among other features, meta-knowledge about the choice of the evaluative model, which permits it to adapt its problem-solving strategy to the problem at hand. Mayer and Palies point out that the major econometric models do not make possible accurate conjunctural predictions, and propose a model of the economic reasoning of the expert working in a particular

Page 5: Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

I N T R O D U C T I O N 5

conjuncture. L~vine and Pomerol have created a means of represent- ing knowledge that is well adapted to those domains where knowledge is highly structured, such as the sphere of finance. Finally, Pham-Hi describes an operational expert system used by the Banque de France.

All of these various efforts at modeling effective economic reasoning constitute a kind of experimental epistemology. Although it is impos- sible to eliminate the arbitrariness of choices and of symbols, it is possible to make explicit the presuppositions that are relied upon in decision-making processes. Here we encounter a necessity that makes itself felt more and more in modern society, which increasingly asks for the reasons behind the choices that are made. It may be hoped that the revelation of these assumptions will provide occasions for reducing intolerance and for favorizing mutual understanding among individuals who are more and more specialized, as well as among communities that are more and more numerous. In its attempt to conceptualize jointly homo cogitans and homo oeconomicus, the conference of De- cember 2, 1987 was a step in this direction.

N O T E S

1 The reader will find an overview of economic theories in B. Walliser's article in this volume. 2 CECOIA I: First International Conference on Economics and Artificial Intelligence, sponsored by the IFAC, IFORS, and the IFIP, held at Aix-en-Provence, 1986.

R E F E R E N C E S

Foerster, H. v.: 1977, 'Objects: Tokens for Eigenbehavior', in Hommage a Jean Piaget, ed. B. Inhelder et al., Delachaux & Nielstel, Neuchatel.

Lucas, R. E.: 1972, 'Expectations and the Neutrality of Money', Journal of Economic Theory 4, 103-24.

Muth, J. F.: 1961, 'Rational Expectations and the Theory of Price Movements', Econometrica 29, 315-35.

Rosh, E.: 1978, 'Principles of Categorization', in Cognition and Categorisation, ed. E. Rosh and B. B. Lloyd, Lawrence Erlbaum, Hillsdale, N.J., 27-48.

Simon, H. A.: 1976, 'From Substantive to Procedural Rationality', in Method and Appraisal in Economics, ed. S. J. Latsis, Cambridge University Press, Cambridge, 129-48.

Page 6: Homo oeconomicus is also homo cogitans homo cogitans is also homo oeconomicus

6 INTRODUCTION

Smolensky, P.: 1986, 'Information Processing in Dynamical Systems: Foundations of Harmony Theory', in Parallel Distributed Processing, ed. D. E. Rumelhart & J. L. McClelland, The MIT Press, Cambridge, Mass., Vol. 1, 194-281.

Stigler, G. J.: 1961, 'The Economics of Information', Journal of Political Economy 69 (3), 213-25.

Varela, F.: 1979, Principles of Biological Autonomy, North Holland, Amsterdam.

P A U L B O U R G I N E LAFORIA--Universit~ Paris VI, GRASCE--UniversitO Aix-Marseille III, CEMAGREF--Parc de Tourvoie-92160 Antony.