History of Political Economy Lunch Seminar. Center for the History of Political Economy, Duke University. December 4, 2009

Transcription

1 Page1 History of Political Economy Lunch Seminar Center for the History of Political Economy, Duke University December 4, 2009 Warning: this is not a paper but a note. It is based on my two months stay with the Chope group working at the Perkins library and archive. As a work in progress, it is supposed to be improved or completely changed during the following months of my stay. More than a note, it is a written talk on what I have in mind to write, according to the policy of our lunch seminars. At the present time its main aim is to receive suggestions and criticisms from an outstanding audience like the members of this group. Hence, this version is only for internal discussion by the Chope group. Any kind of reference or quotation is not allowed. Giandomenica Becchio (University of Torino) Carl Menger and Complexity This paper is a part of an Italian national research project on complexity theory in the history of economics. 1 One of the reasons for the conduct of this kind of inquiry is that for some years there has been growing interest in the possible connections between complexity theory and economics (Rosser 2009, Colander 2009). My focus is mainly on the philosophical context in which possible links can be identified between complexity theory and the Austrian school. Before continuing, some essential points must be established: first, very briefly, I shall seek to clarify what is meant by complexity theory ; then I shall focus on connections between complexity theory and the Austrian school in a general perspective; finally I shall suggest some possible further lines of inquiry that include: Carl Menger and complexity theory (also in honor of the title of this talk), mainly based on his second edition of the Principles and on my work in the Duke s archive; and Hayek and complexity theory (where I shall seek to show the continuity between the Sensory Order (Hayek 1952) and subsequent writings on complex systems (Hayek 1967, which includes Hayek 1955; 1962; 1964; 1964a; 1978) and a couple of unpublished works conserved in the Duke archive. 2 1 Italian departments of economics involved in this project are those of the universities of Torino, Firenze, Milano, Novara. 2 I am indebted to Bruce Caldwell for telling me about the presence of an unpublished typescript Within system and about system and its connection with my present research.

2 Page2 1. The complexity of complexity theory Complexity theory is simply a mess. There are plenty of publications on it and it has gained a place in whatever discipline that comes to mind: philosophy, sociology, computational science, neurobiology, aesthetics, anthropology and, of course, economics. Because complexity theory is so complex, it is easy to find skeptics on the one hand or almost fanatical supporters on the other. The first step in taking complexity theory seriously should be to give it an exact definition. Unfortunately, a clear definition of complexity theory does not exist: according to the physician Seth Miller there are 45 definitions of complexity (Horgan, 1997; Rosser 2009), but there are many characteristics which are recognized and accepted by scholars involved in this field of research. Generally speaking, we can understand complexity theory as a (new?) theory of knowledge. In a broader sense it can be regarded as a new scientific paradigm that makes it possible to bridge the gap between the natural and social sciences; in a narrower sense, it can be regarded as a methodological tool that allows study of specific disciplines or aspects of a discipline. As regards economics, complexity theory can be used as a new paradigm within which to describe the dynamics of individuals and social groups while trying to find a new (?) and heterodox scientific approach (Colander ; Keen 2001); or it can be used as a sophisticated tool with which to explain classical arguments in a new way, without any argumentation against the mainstream point of view (Arthur, Durlauf, Lane 1997) 4 ; or it can be regarded as something midway between these two extremes, which are now becoming obsolete (Colander, Holt, Rosser, 2004). Let us take a step backward in order to consider the roots of complexity theory in a more general framework, focusing our attention on its philosophical meaning during its early stages. One of the first authors to have explicitly spoken about complexity theory, some years before the foundation of the Santa Fe Institute (1984), was the French philosopher Edgar Morin (Morin 1973; 1977; 2008). He presented complexity theory as a transdisciplinary developing a form of knowledge based on a new meta-paradigmatic dimension applicable to the social sciences. This meant that knowledge of social dynamics can be organized and understood because societies are sets of institutions, and institutions are expressions of individual knowledge. Because there is an isomorphism between the cognitive and institutional levels, it may be possible to acknowledge and explain how the knower constructs knowledge using an antireductionist approach. 5 Morin stressed the need to overcome the dichotomies, such as holism 3 Colander talked more precisely about the incompatibility between neoclassical economics and complexity theory (Colander p.136). 4 The issue of the use of complexity theory in mainstream economics is controversial. As Kreps said, at the end of the last century, economics was broadening its interests, and new links with other disciplines in an interdisciplinary perspective were quite acceptable (Kreps 1997). 5 Reductionism means that a whole object is reduced to its minimum parts in order to classify and know it.

3 Page3 versus atomism, which marked Western thought during the last century and which produced a hyper-specialized knowledge that increased impermeable borders among sectors and subsectors, above all within departments at universities. The purpose of this hyper-specialization was to reduce uncertainty in any specifically scientific field, because uncertainty was considered to be a source of anxiety. Complexity theory, by contrast, considers uncertainty as an opportunity for creativity and for the development of new perspectives to be studied in any discipline. Morin explained that there had been three levels of inquiry in the history of scientific revolutions. The first was the Newtonian mechanics based on necessity: science consisted of universal laws able to form general theories in order to make exact predictions to be proven by experiments: matter and energy existed in an absolute space and time, and they were ruled by the determinism of the law of cause and effect. Knowledge was objective in a twofold sense: it was knowledge of objects, because their characteristics can be quantified and measured; and it was universally objective knowledge, because the knowing subject was not in question, given that mind is universally structured. The second paradigm was the equilibrium in thermodynamics based on chance and the irreversibility of entropy. The third paradigm started with the Darwinian revolution in which organizations were presented as complex interactions between order and disorder. Subsequent developments of this approach led to the definition of open systems as selforganizations where collective behaviors spontaneously emerge. 6 One of the main differences between complexity theory and the other theories of knowledge is that in its endeavor to reorganize the concept of science, it goes beyond the division between subject and object, given that these are parts of the same open system. The much abused concept of a whole not reducible to the sum of its parts is the basis of the notion of an open system, which was originally a thermodynamic one. But complexity theory seeks to bridge the gap between thermodynamics and biology or life sciences : it wants to define the laws of organization of open systems, not in terms of equilibrium, but in terms of dynamics between individuals and the environment. 7 In 1945, Schrodinger showed that living organizations do not obey the second thermodynamic principle, and von Neumann thereafter pointed out the differences between living machines (self-organized systems) and artificial ones (simplyorganized systems) (von Neumann 1966). Morin described complexity theory as a step forward from cybernetics: cybernetics recognized the importance of the interaction among a large number 6 If we apply this threefold distinction to economic theory, we can consider the classical school and the marginalism of the founders as embedded in the first paradigm (the Newtonian one); the subsequent development of marginalist theory, i.e. mathematical economics and econometrics, based on the formalization of economic issues as patterned on physics (Mirowski 1989); and contemporary research on complexity in economics as founded on neuroscience and biology. But any partition is quite tranchant and there are numerous overlaps among periods in the history of a discipline, as well as in the thought of any single scientist. 7 Modern biology has passed from the romantic concept of organicism to the modern concept of organizationalism: if organism and organizations are considered complementary and isomorphic, their functions can be described as a whole in terms of the theory of self-organizations.

4 Page4 of units and the role of uncertainty caused by limited knowledge, as well as the mixture of order and disorder within a system; but it put all these phenomena into a black box. Complexity theory delves into that black box. How is it possible to get inside the black box? For example by finding another way to conceive and interpret the dynamics of interactions between individuals and groups. An example of this new way of thinking is fuzzy logic, which can be regarded as inquiry within that black box, because standard logic is not sufficient to explain open systems. In a certain sense, Godel s undecidability theorem (which he restricted to mathematical logic) is applicable a fortiori to all theoretical systems. But what is the object of complexity theory? Generally speaking, it is the relation between subject and object in the following terms: if a subject is situated in its natural eco-system, then it is possible to examine the biological traits of knowledge because cerebral forms of learning are in that environment. This approach leads to a sort of new unity of science: every science is composed of the interaction between the knower and the known as dynamic parts of the same whole in which actions, interactions and feedback are considered in terms of logical patterns (standard logic is no longer sufficient) and empirical consequences (the well-known butterfly effect). This unity of science is very different from the paradigm proposed by the logical positivism launched by the Vienna Circle, which played the role of an epistemological policeman forbidding us to look precisely where we must look today, toward the uncertain, the ambiguous, the contradictory (Morin 2008, 31). Moreover, logical positivism was based on physicalism (the belief that all science ultimately reduces to the laws of physics) (Neurath 1931), which was a sort of reductionism to which complexity theory is strongly averse. In opposition to reductionism, complexity theory uses so-called emergentism, 8 which was developed in order to counter the dualism between monism and dualism, or between objectivism and subjectivism. It was based on the following assumptions: the category of emergence is able to explain any kind of reality and it can be applied to living beings as well as to social phenomena; the rejection of ontological dualism and reductionism and the idea that beyond the whole and its parts there is something more (a quid that emerges); the acceptance of evolutionary theory as regards biology. The concept of emergentism or emergence is older than that of complexity. John Stuart Mill was the first philosopher to used the term in order to explain some properties of dynamic realities (physical and social) (Mill 1843): the more modern concept of emergentism derives from the general system theory developed by Ludwig von Bertanlaffy (Bertanlaffy 1950). From a historical point of view, the forerunners of complex system theory were nineteenth-century Darwinian organicists like Schäffle and Spencer. In 1938, Ablowitz defined emergence as a non-additive, non-predictable or deducible, hierarchical element: the essential newness of the 8 Some examples of emergences are the following: the V shaped formation of birds when they fly together: this is not planned or centrally determined but arises from each bird s behavior based on its position with respect to nearby birds; communication among the members of a jazz band; some linguistic shifts in particular contexts like social networks.

5 Page5 theory itself lies in its emphasis on unpredictability, for in no previous philosophy has this concept been so central: it is thus a kind of philosophical analogue to the Heisenberg principle of uncertainty in the behavior of electrons (Ablowitz 1938 p.12), and, being aware of the possible mystical development of the application of emergence, he added: however, like alcohol, it is stimulant only in proper doses: many who have used it have gotten drunk in the attempt to apply it to everything (p.16). After World War II, cybernetics developed models of computational and communication technologies. More recently, during the last two decades, more advanced developments in computer technology have led to simulations of mathematically modeled social dynamics in which there are distinct computational agents for every individual (so-called multi-agents systems) i.e. n agents in communication form an artificial society in which the global behavior of a system emerges from the actions and interactions of agents. The main problem with emergentism is the proper definition of its ontological status, because its anti-reductionism may imply dangerous forms of vitalism or idealism: some philosophers, Sawyer for example, solve this problem by considering emergentism as a form of non-reductionism that accepts the ontological position of materialism (Sawyer 2005, 28); this claim is possible because these philosophers do not consider reductionism to be a consequence of materialism. From a methodological point of view, it is rather difficult not to consider the concept of emergence in opposition to methodological individualism, even though it was evoked by methodological individualists in economics: for example Menger, who was influenced by Mill, and Hayek in his struggle against scientism, considered emergence and methodological individualism to be compatible: they invert the causal arrow of the structural determinists: instead of top-down causation, [and] focus on bottom-up causation, which they often refer to as emergence (Sawyer 2005, 195). Contrary to this position, Israel proposed a different approach to the epistemological status of the science of complexity (Israel 2005). He claimed that the hope of superseding reductionism by means such as emergence was fallacious, because the science of complexity proposes forms of reductionism which are even more restrictive than the classical ones. Israel was strongly averse to the notion of complexity because it lacked any rigorous definition. According to Israel, the only area in which it had a precise meaning was algorithmic complexity theory, because there it had a quantitative definition, not a qualitative one. The science of complexity is based on negation of the principle that the whole is the sum of its parts (holism) because new properties emerge in the whole. The only meaningful complexity theories are cybernetics, chaos theory and game theory. According to Israel, it was a nonsense to say that complexity is neither a state of equilibrium nor a chaos but rather a state in which the system is creative: this is a mystical intuition, and also the so-called new mathematics (non linear) was nothing but game theory. There is nothing between holism and individualism; nothing exists from a mathematical point of view, and so-called emergence is a form of reductionism. Israel s criticisms are crucial because

6 Page6 the passage from linear thinking to non-linear thinking is considered the core of complexity theory (Mainzer 1994). 2. Complexity in economics The beginning of links between economics and complexity theory is recognized as being Simon s article on the architecture of complexity (Simon 1962), in which he wrote that even if there was no formal definition of complex systems, they can be roughly conceived as made up of a large number of parts that interact in a non simple way (p.468). Simon pointed out that the leitmotiv of complexity theory, i.e. the whole is greater than the sum of its parts, should not be understood in metaphysical terms, but in a pragmatic sense, and that an in-principle reductionist may be at the same time a pragmatic holist. 9 Besides Israel s extreme position cited above, there are strong supporters of complexity theory in economics. Colander, Holt Rosser (2004) have claimed, contrary to Israel, that the neoclassical era is dead and has been replaced by the so-called complexity era, which can be regarded, not as a revolution, but as an evolution of heterodoxy in economics. They argue that models based on a priori assumptions will decrease and be replaced by empirically-driven models. It is true that the term complexity has been overused and overhyped, but Colander and Holt Rosser s vision is a view of the economy as too complicated to be treated by aggregate models (neo-walrasian neoclassical theory no longer works, and the economic profession is changing as well), hence it s time for introducing complexity theory in economics. Complexity can be explained from a general point of view by introducing the concept of bounded rationality á la Simon; from a dynamic point of view: a dynamical system is complex if it endogenously does not tend asymptotically to a fixed point, a limit circle or an explosion (cybernetics) (Rosser 1999, 170); and from a computational point of view (information theory Albin and Foley 1998; Velupillai 2009). The problem of the whole and its parts has two aspects: the relation between micro and macro (what Keynes called the fallacy of composition, which Walrasian macroeconomics solved by using the representative agent model), and the spontaneous emergence of higherorder structures as studied by the Santa Fe group. The Santa Fe analyses focus on the dispersed interaction among heterogeneous agents, no global controller, cross-cutting hierarchies with tangled interactions, adaptation and continuous learning, novelty, out-of-equilibrium state and no optimability. According to Velupillai, complexity theory is the evolution of game theory applied 9 According to Simon, there are four aspects of complexity: 1. Hierarchy of subsystems 2. Relation between the structure of a complex system and the time required for it to emerge (more quickly than non-hierarchical system) 3. Decomposition into subsystems in order to analyze them 4. Relation between complex systems and their description.

7 Page7 to institutions; it concerns ecological economy and it uses agent-based computer as alternative to analysis modeling. Rosser (2009) has described complexity as a transdisciplinary view of the world that can be applied in specific disciplines. There are three levels: a small tent complexity (Brian Arthur 2009 and Santa Fe Institute) in which the concept of complexity can be reconciled with neoclassical economics; a big tent complexity (cybernetics, catastrophe theory and chaos theory); and metacomplexity (Albin 1982). 10 But where did complexity come from? Weintraub (2002) explained how revolutions in economic theory had reflected revolutions in the history of mathematics; and Mirowski described how, during the late nineteenth century, neoclassical economics arose from the mathematics used by the physics of the mid-nineteenth century (Mirowski 1989). The feature shared by all studies on complexity is the idea of systems with multiple elements adapting or reacting to the pattern these elements create (Brian Arthur 2009, 12); the aim is to understand the endogenous changes in aggregate patterns created by agents with a set of strategies that are dynamic and fundamentally unpredictable and emergent. These strategies can be considered as sorts of positive feedbacks (that are non-predictable) or negative feedbacks (that are perfectly predictable). In economics, positive feedbacks are related to increasing returns. Complexity theory emphasizes the formation of structures, rather than their given existence, and it can be applied to game theory, money and finance, learning processes, and the evolution of networks of heterogeneous agents. 3. Complexity in the Austrian School Bearing in mind the foregoing discussion of complexity theory and its connections with economics, we may now turn to the elements of complexity theory that can be found in the Austrian tradition. The first scholar to link the Austrian School with complexity theory was Karen Vaughn (1994): she considered Menger s idea of the spontaneous origin of institutions (which arose without a common will), he inherited by the Scottish Enlightenment (according to the common interpretation of Hayek) as the most prominent manifestation of the economic growth regarded as increasing complexity in the system. Complexity is caused by a larger number of products on the market, the division of labor, and the increasing number of economic institutions due to the development of information and to the improvement of exchanges. Complexity and spontaneous order are linked in the Austrian paradigm by a process of systematic, ordered change in either the formal or informal rule structures by which people attempt to achieve their purposes (125). According to Vaughn, the unplanned and unconscious changes in institutions are common to Menger s and Hayek s theories, as well as to those of Mises and Lachman (Vaughn 1994, 150). 10 Albin tried to find a metalogic of economics: he claimed that certain concepts such as general economic equilibrium are subject to Gödel undecidability and incompleteness; hence Gödel s theorem must be applied to the broader types of social decision-making or social evaluation.

8 Page8 More recently, most studies on this specific subject have focused on Hayek s thought, of course, and for many reasons: first of all because he wrote an article on complex systems (Hayek 1967) but also because of his studies on psychology and their link with Simon s concept of bounded rationality. In this debate there are economists who see no meaningful connection between Hayek and complexity theory (Rosser 2009, Colander, 2009 Caldwell) and other economists who stress the closeness of Hayekian themes to complexity theory (Fiori 2009, Koppl 2009, Horwitz 2008). From a broader perspective, there is also a very recent debate on whether there are actually any connections between Austrian school and complexity theory. This recent debate is between Barkely Rosser and Roger Koppl (2009). According to Koppl, Austrian economics as a school of thought can be regarded as part of the broader complexity movement in economics. (Koppl 2009, 1). He has introduced what he terms BRICE economics in order to describe elements of complexity theory within economics, pointing out that the Austrian economic tradition shares BRICE with complexity theory. BRICE is the acronym for Bounded rationality, Rule following, Institutions, Cognition, and Evolution. Both Austrians and complexity economists discard the paradigm of full rationality and adopt Simon s bounded rationality (Morgenstern and Hayek, Velupillai 2005); they both consider economic agents to be rule followers (Hayek and Arthur 1994); they both seek to show the macrofoundations of microeconomics, i.e. the fundamental influence of institutions on individuals (Menger, Mises, Hayek; Colander); they both assume the role of cognitive psychology as a consequence of adopting bounded rationality, and they defend so-called verstehen psychology (Hayek, Mises, Schutz; Simon Holland Kauffmann). Finally, they both adopt models of economic evolution (Hayek, Kirzner; Epstein and Axtell 1996). In regard to the fourth aspect (C = cognitivism), Koppl has also argued for the importance of a hermeneutic approach to economic analysis for both Austrian and complexity theorists. 11 Rosser disagrees with Koppl. He admits that there are some overlaps between Austrian economics and complexity theory (such as the idea of a spontaneous order; the introduction of emergence ; and the consideration of endogenous irregularities of the system) but there are many more substantial elements and strands within Austrian economics that do not fit in with any of the multiple varieties of complexity theory, even though there are some that clearly do (Rosser 2010). Rosser s arguments against Koppl are mainly focused on the fact that the substantial vein of Austrian thought is not consistent with complexity above all because the Austrians accepted equilibrium approach (Hayek included) and because they ignored the problem of uncertainty, with the exception of Lachmann (1976). Rosser then criticizes the presumed points of contact between Austrians and complexity economics. He claims that the role of rules in Hayek s perspective is a weak argument; and institutions are a serious difficulty as well, because Menger was in strong opposition against the precursors of institutionalism (the German historical school) 11 On Austrian themes and hermeneutics see also Lavoie 1989.

9 Page9 and Hayek was not influenced by Commons during his first stay in the US in the early 1920s. Moreover, according to Rosser, uncertainty is not an Austrian theme, but it is a post-keynesian one. Rosser highlights Austrian themes (mainly connected with other members of the school) that are not compatible with complexity theory, i.e. the emphasis on marginalism and the importance of the concept of equilibrium. He writes: it must be remembered that in the Methodenstreit, it was the Austrians who upheld the nascent neoclassical orthodoxy of marginalist equilibrium theory against the proto-institutionalism and on opposition to abstract theory articulated by Schmoller (1978 [ ]) and the followers of the German Historical School. 12 Moreover, the economic thought of Eugen von Böhm-Bawerk and Friedrich von Wieser (who emphasised subjective marginal utility and the idea of economic equilibrium) and Mises apriorism are no themes to be linked with complexity theory. Finally Rosser rejects the use of any kind of hermeneutic approach in economics. In addition Rosser claims that the typical Austrian subjectivism can be compatible with complexity theory but is not its condition sine qua non. Generally speaking I support Koppl s position: like him, I focus mainly on Menger and Hayek, but I think that BRICE explanation is insufficient because it does not comprise a set of exclusively Austrian themes. Moreover, Rosser s arguments seem to have some faults. I begin with this last point. 1. Firstly, it is rather difficult to maintain that uncertainty is not an Austrian theme. Perhaps Rosser is thinking of the measurement of uncertainty as a computational problem, or has in mind Mises apriorism as a sort of ontological rather than methodological tool to avoid uncertainty. Not only real uncertainty related to lack of knowledge but also uncertainty as a matter of perception dealing with expectations was a central theme for the Austrian school (Borch 1973; Streissler 1973) from Carl Menger to Hayek. 2. The fact that Carl Menger was strongly opposed to the new generation of the German Historical School is not an argument against his ante-litteram belonging to complexity theory. On the one hand, why should German historicism, like any other historicism, belong to complexity theory? And, on the other, the relationships between Menger and the German Historical School were rather more complex than has been claimed. The notes and comments in his archive show that the influence of the old German historical school was quite strong on Menger s composition of the Principles; it was even stronger than his debt to the Scottish Enlightenment, stressed by Hayek in his portrait of Menger (Hayek 1973, 1978), and to Adam Smith, who was quoted in Menger s Principles, although mainly to criticize him. Menger s strong opposition against the German Historical School was above all against the new generation of German historicists, and it took place after the reaction of Gustav Schmoller to Menger s Untersuchungen (1883) in the late 1880s. 12 Caldwell (2004, p. 80) argues that the term Austrian economics was originally a term of derision applied by members of the German Historical School to the Austrians during the Methodenstreit.

10 Page10 3. The idea that the Austrian School accepted the model of general economic equilibrium is also rather difficult to admit. It seems to derive from Block (1940), according to whom Menger was the mathematician counterpart of Walras, and generally speaking that the Austrian school can be considered the psychological variety of the neoclassical tendency, and also that the Walrasian school was the mathematical variant of the neoclassical tendency. The reason, according to Bloch, was methodological: they both arrived at a formal theoretical economics because they employed deductive methods, even if Menger and subsequent Austrians did not use mathematical tools (it is also true that, more recently, Karl Menger (1973) has shared the same opinion, but I have argued elsewhere that his position was probably forced by his [two souls in his breast] (Becchio 2008). Rosser maintains that this applies to Böhm-Bawerk and Wieser. It is true that they were in a certain sense the more orthodox or neoclassical economists of the Austrian School, but this is not enough to consider GEE an Austrian theme, not least because the GEE model is static: on the contrary, the role of time and dynamics in Menger and Hayek, as well as in Mises and Schumpeter and the following generations of Austrians, is fundamental (Rizzo 2010). On the other hand, Koppl s BRICE raises some problems. I think it should be reduced to RICE or to ICE, because bounded rationality is not a theme exclusive to the Austrian school, and Simon cannot be ascribed to the school (even if there are many links between Simon s idea of bounded rationality and the Austrian conception of knowledge). Even the role of rules is present in Hayek but not specifically in other members of the Austrian school. Institutions, cognitivism and evolutionism were definitely themes embedded in the Austrian tradition from its origins. According to the most recent definition of the Austrian school (Rizzo 2010), its main concerns were (1) the subjective, yet socially-embedded, quality of human decision-making; (2) the individual s perception of the passage of time ( real time ); (3) the radical uncertainty of expectations; (4) the decentralization of explicit and tacit knowledge in society; (5) the dynamic market processes generated by individual action, especially entrepreneurship; (6) the function of the price system in transmitting knowledge; (7) the supplementary role of cultural norms and other cultural products ( institutions ) in conveying knowledge; and (8) the spontaneous, that is, not centrally-directed, evolution of social institutions (Rizzo 2010 quotation from the online version). Which of these definitions matches complexity theory? Definitely 3, 4, 7 and 8, which are common aspects of the Austrian school as a whole. 3.1 Why Carl Menger had something in common with complexity theory. Let us go more specifically into Menger s political economy in order to discover possible overlaps between his thought and complexity theory. What are the elements of Menger s first edition of Principles which can be considered linked to complexity theory? There is no better starting point than Hayek s description of Menger s intentions (Hayek 1973, 1978). According to Hayek, Menger wrote in the Preface of his Principles: I have endeavored

11 Page11 to reduce the complex phenomena of human economic activity to the simplest elements that can still be subjected to accurate observation, to apply to these elements the measure corresponding to their nature, and constantly adhering to this measure, to investigate the manner in which the more complex economic phenomena evolve from their elements according to definite principles (Hayek 1978, 276-7). His aim was to find a method common to all fields of empirical knowledge. The task of his age, claimed Menger, was to establish interconnections between all fields of science and to unify their most important principles. This task would be achieved only when the laws of any particular science had been discovered. This is precisely what Menger had in mind when he wrote his Principles and invited the reader to judge whether his method of investigation has led to discovery of how the phenomena of economic life are ordered according to its laws. Finally, Menger underlined that it is possible to talk about exact laws in social sciences because they are not incompatible with human free will. Hayek claimed that Menger s aim was to trace the complex phenomena of the social economy to their simplest elements which are still accessible to certain observation (Hayek 1978, 276-7), as he wrote in his Preface to his Principles. Hayek wrote: if we were to derive from our knowledge of individual behavior specific predictions about changes of the complex structures into which the individual actions combine, we should need full information about the conduct of every single individual who takes part (277), and Menger was aware that we can never obtain all this information. Rather, starting from these known elements it is possible to combine only certain stable structures: Carl Menger was quite aware of this limitation of the predictive power of the theory he developed and he felt that more could not be achieved in this field (278-9). Menger s method of investigation was based on an empirical procedure: in the social sciences we start from our acquaintance with the elements and use them to build models of possible configurations of the complex structures into which they can combine and when are not in the same manner accessible to direct observation as are other elements. According to Hayek, when we observe the actions of other people we understand the meaning of such actions in a different way from that in which we understand physical events. The subjective character of Menger s theory is based on our capacity to understand the meaning of observed actions. For Menger, observation entails the concept of Verstehen, which means understanding and implies an introspective knowledge. Some secondary literature has stressed Menger s verstehende approach. According to Max Alter (Alter 1990), Menger s philosophical background which is fundamental in order to know his methodology is based on that approach (and on Aristotelianism 13 ). The role of Verstehen (=understanding) vs Erklären (= explanation) was chosen by Menger in order to emphasize the psychological aspect of knowledge. He chose it as part of a very typical Austrian tradition linked with hermeneutics (originally the art of interpreting Holy Scriptures), as pointed out by Lavoie (1989), who claimed that the idea of understanding as a spontaneous process in Menger s thought can be read in hermeneutic terms. 13 The Aristotelianism in Menger is the universal law of cause-effect, which he used from a subjective point of view.

12 Page12 A couple of considerations are in order. First, Menger talked about complex economic phenomena, but he immediately explained that his aim was to reduce them to their simplest elements; otherwise it would be very difficult to find the laws which they obeyed. This was a point against complexity theory, because this was reductionism from a methodological point of view, i.e. the idea that a whole can be analyzed or reduced to its atomic elements. Second, for Menger it was essential to establish a link between human and natural science, because they are both subject to exact laws (and this exactness is not in contradiction with human freedom). This point can be treated in two ways as regards a possible connection with complexity theory. It is not clear what Menger had in mind when he referred to all fields of science beyond social science. Was it physics? Was it biology? Or both (Menger was unconcerned about the distinction between them and regarded them both as natural science )? Complexity theory tries to bridge the gap between natural and social science by using the same methodological approach. But it is more likely that Menger had physics in mind: in this case, his thought is entirely embedded in the method of theorizing of the early neoclassical economists, who borrowed from the physics of the nineteenth century (Mirowski 1989). But some interpreters of Menger s thought would totally disagree with this statement. First of all Hayek, who claimed that one of the most important features of Menger s methodology was its capacity to understand the meaning of human actions in a different manner from physical events. And, as Hayek complained, this point of view (typically Austrian) was forgotten by microeconomic theory and the later development of the indifference curve related to the revealed preference approach, which were designed to avoid the reliance on such introspective knowledge (Hayek 1973, 9). More recently, Vaughn (1994) has claimed that the discordance between Menger and his neoclassical contemporaries was never widely recognized. Where is complexity theory in Menger s Principles? In three points: in the passage from a desire to the satisfaction of this desire by means of a good 14 ; consequently in the passage from goods of different orders 15 and in the central role of time and uncertainty in this passage. 16 As well known, 14 Goods are described as having four properties: being a human need; being capable of being brought into a causal connection with the satisfaction of this need; this causal connection is known; command of the thing is sufficient to direct it to the satisfaction of the need. According to Hayek (1973), the causal connection that Menger found between goods and the satisfactions of wants allowed him to consider the production factors like any other goods in the distinction between goods of lower and higher level. From this general meaning of goods derived the pure logic of choice (or economic calculus) based on the relationship between scarce means and unlimited ends: it contains at least the elements of the analysis of consumer-behavior and of producer-behavior; the two essential parts of modern micro economic price theory and this was his main achievement (p. 7); his followers developed the former, and Marshall the latest. 15 Economic goods can be of first order (consumption goods) or of higher order (means of production). 16 Many interpreters have stressed the fundamental role of time and uncertainty in Menger s thought: uncertainty in Menger can be found in the way in which the output will be obtained for the chosen input (Block 1973, 65) or: in his theory of money, which was shaped under uncertainty, [and it was] basically a disequilibrium theory of money (Streissler 1973, 167).

13 Page13 according to Menger every event, without any exception, obeys the law of cause and effect: any change from one state to another is subject to the universal law of causality. This is the wellknown Aristotelian approach in Menger s philosophical framework. Because a desire and the subsequent satisfaction of that desire are a change from one state to another, there is a causal connection between the desire and the thing able to satisfy it (good): in other words, there are forces in operation within one s organism (needs) and external useful things which can satisfy those needs (goods). How do these forces operate within a human organism? And what is the relationship between them and the external environment in which goods for their satisfaction can be found? Unfortunately, there is no theory of needs (or desires) in Menger s Principles, at least not in the first edition (1871, 1981), and we have to wait until the second one. According to Alter, Menger s theory of needs (Bedürfnissen), which appeared in the second edition of his Principles (probably written after the publication of Oskar Kraus book Das Bedürfnis (1896)) ties him into the German Weltanschaung of the Geistesgeschichte, and it is in relation to biology and psychology. But he did not demonstrate that his theory was to be the bridge between biology and economics (Alter 1990, 125). But let us continue to focus on the first edition, where the passage from higher-order goods to first-order goods comes about in terms of decreasing complexity, and in which time and knowledge perform fundamental roles because the idea of causality is inseparable from the idea of time. The same holds for the reverse process from first-order goods to higher-order goods. Here certainty does not exist, and the aim of economics is to understand how people, on the basis of their (limited) knowledge, direct available quantities of goods to the greatest possible satisfaction of their needs. As Menger wrote: we can never fully understand the causal interconnections of the various occurrences in a process, unless we view it in time, and goods of higher order acquire their goods-character not with respect to needs of the immediate present, but as a result of human foresight (Menger 1981, 68). Hence, individuals are exposed to appreciable uncertainty with respect to the quality and quantity of a product. Menger wrote: human uncertainty of the whole causal process is greater the larger the number of elements that do not have goods character this uncertainty is of the greatest practical significance in human economy the greater or less degree of certainty in prediction depends upon the greater or less degree of completeness of their knowledge of the elements of the causal process human uncertainty of the whole causal process is greater the larger the number of elements involved in any way in the process even understanding them, we have no control the uncertainty is of the greatest practical significance in human economy (p. 71) From a more general perspective, Menger added that further developments in human progress are related to the increasing complexity of the system and the greater number of higher-order goods: Increasing understanding of the causal connections between things and human welfare have led mankind from a state of barbarism to present stage of well-being the degree of

14 Page14 economic progress of mankind will still be commensurate to the degree of progress of human knowledge (Menger 1981, 74). As well known, after the publication of his Principles, Menger worked on methodology and published his Untersuchungen in 1883 (Menger 1963). What are the elements of his investigation on method that could be ascribed to complexity theory? Firstly, in this book he explicitly talked about verstehen, explaining that the goal of scholarly research is cognition not only as explication but also as understanding (verstehen). It is possible to understand social phenomena in two ways: historical and theoretical. Knowledge of phenomena is to be extended beyond immediate experience from certain observed facts about other facts not immediately perceived on the basis of the law of coexistence and of the succession of phenomena. And he added: cultural development depends on the degree of development of the desire of knowledge of phenomena in their full empirical reality, that is, in the totality and the whole complexity of their nature (Menger 1963, 56). According to Menger, because of this complexity, there are no strict types in empirical reality ; hence laws of phenomena must be considered as regularities among phenomena which coexist. Knowledge is scientific when actual regularities are found in the succession and coexistence of real phenomena. This is applicable to natural phenomena, the ethical world and economic phenomena, because there are no essential differences among these realms but only matters of degree. And this is the second point at which Menger s thought in 1882 and complexity theory overlap. Because our results are formally imperfect, however important they may be for human knowledge and practical life, regularities are not absolute. Hence laws of nature do not exist; there are only exact laws. Menger wrote: the laws of theoretical economics are really never laws of nature in the true meaning of the word (59). The contrast between natural and social sciences is not a contrast of method because, on the one hand, no exact natural science can exist (meteorology) and, on the other, an exact social science can exist. 17 Menger wrote: not just any one theory of human phenomena, only the totality of such theories, when they are once pursued, will reveal to us in combination with the results of the realistic orientation of theoretical research the deepest theoretical understanding attainable by the human mind of social phenomena in their full empirical reality (63) Menger defined economy as the precautionary activity of humans directed toward covering their material needs. The exact orientation of theoretical research is to apprise us of the laws by which not real life in its totality but the more complicated phenomena are developed from elementary factors (63). There is also a realistic orientation in economics: it deals with real and observable regularities in coexistence and in succession. On adding exact orientation and realistic orientation, we reach theoretical research, that is, the understanding or verstehend of 17 According to Menger, in the realm of the ethical world there exists an exact orientation of the theoretical research able to reduce human phenomena to the most original and simplest constitutive factors ( ) our task is to try to investigate the laws by which more complicated human phenomena are formed from those simplest elements thought in their isolation (62).

15 Page15 economic phenomena: both orientations of research are adequate not only for all realm of the world of phenomena, but also for all stages of the complexity of phenomena (69). Realism and exactness are not in competition because they are different from each other (it would be an absurdity to think of measuring exact orientation by standards of realism; it would be as if a mathematician corrected the principles of geometry by measuring real objects). If realism and exactness matched, human cognition would be simplified, but this is not the case. Hence, we must be aware of the confusion between historical and theoretical (=empirical or realistic orientation + exact orientation) understanding of social phenomena. Another point in Menger s Untersuchungen can be read from a complexity theory perspective. It occurs in Book 3 (The organic understanding of social phenomena), where Menger explains the analogy between social and natural phenomena: there exists a similarity between natural organisms that are complex in their details and in their unity and structures of social life formed of parts (individuals) and wholes (groups). Neither natural nor social structures are the result of calculation, but rather of a natural process as unintended products of historical development. Menger added that only some social phenomena exhibit an analogy with natural organisms: some social phenomena are planned. Parts and whole are linked by mutual causation; there is an organic origin of some social phenomena (the organic view cannot be a universal means of consideration. Even the understanding of natural organisms can be sought in two orientations of research, one empirical-realistic (collective), the other exact (atomistic). The exact understanding of the origin of those social structures is the unintended result of social development: the problem which science has to solve consists in the explanation of a social phenomenon of a homogeneous way of acting on the part of the members of a community (152). The general nature of the process to which those social phenomena owe the origin which are not the result of socially teleological factors, but are the unintended result of social movement (158). This way of describing social phenomena and institutions is based on a strict connection between biology and economics; and in this form it is quite similar, on the one hand, to what German social thinkers of that time had in mind, and, on the other, to what Marshall later wrote in Book 4 of his Principles. The idea of a Menger embedded in the German tradition may seem odd. We are accustomed to thinking of him as an heir to the Scottish Enlightenment and in strong opposition to the German Historical School (according to Hayek s interpretation). The Scottish heritage in Menger is apparent in a certain sense, especially if we consider Ferguson and Hutchinson, whom Menger often quoted. But matters become more complicated if our reference is Adam Smith. Menger often quoted him in his Principles, but mainly to criticize his theory of value. Examination of Menger s archive at Duke reveals that references to Scottish authors are very few compared with German thinkers. Hayek himself (1973: 1978) recalls that Menger started to take economics seriously after reading Rau s textbook, which he extensively annotated. It should also be pointed out that Menger dedicated his Principles to Roscher and that he, unlike Jevons and Walras, did not claim to be a revolutionary thinker. He was aware of being embedded in the German tradition: in his sense of realism he is closer to Marshall than to the