1Austrian economic theory and game theory have a priori very little in common. The first emerged at the end of 19^th century with Carl Menger and is still alive today, even if its influence in economics has strongly declined during the 20^th century. Game theory was introduced in economic analysis after the Second World War and its analytical importance grew little by little, playing today a central role in theoretical and applied economics. However, in spite of these differences, both theories had connections in different periods and therefore, some commentators had the opportunity to stress these relationships (for instance, see Foss, 1999; Garrouste, 2001; Koppl, 2006; Kelly, 2009; Cevolani, 2011) trying thus to show that both research programs were compatible or at least were not incompatible.

2This paper goes further, trying to connect the respective works of two important economists, Friedrich von Hayek and Michael Bacharach, namely one of the main intellectual leaders of the Austrian Schools and one of the most original game theorists. These authors were very different theorists and worked on distinct issues, even if Hayek certainly anticipated some conceptual issues of game theory (see again Foss, 1999 and Cevolani, 2011). However, their contributions had something in common and this commonality suggests the main focus of this paper. Hayek and Bacharach can indeed be included among the scarce economists who tried to show that the combination of economics and psychology in economic analysis provides the necessary route to cope with the problems raised by economic coordination of individual agents especially on markets; moreover the former combination implied an elaborate representation of individual and social cognitive processes. Now this view was and is still rejected by most of the economists since Pareto and it is therefore essential to discuss its validity. Let us enter into more details.

3Both of our authors always rejected the viewpoint of the General Economic Equilibrium (GEE) research program according to which agents’ preferences are exogenous and given, belong to the “fundamentals” of microeconomic theory and do not need any psychological foundations to play their role. They never accepted that in standard microeconomics, usual rational choice theory was a self-sufficient framework which did not need any form of help from psychology. Quite the reverse, they both considered that subjective perceptions of the real world provide the first stage of decision processes and that, within this stage, psychological factors played a fundamental role. Therefore, they both proposed how perceptions, economic rationality and social coordination could be combined.

4Hayek and Bacharach are therefore authors—few in number—who do not conceive that economic analysis can be built without the help of psychology. However this point of view creates new analytical difficulties they could not ignore. The incorporation of subjectivity in economic behaviour makes the analysis of economic and social coordination more complex. To overtake these new difficulties we will see that both Hayek and Bacharach integrate a specific approach to human cognition in the social setting and resort to an evolutionary explanation of social coordination, even if they do not attribute the same meaning to the concept of evolution. This is the main message of this contribution and this is why we will start with individual perceptions, continue with cognition and knowledge and end with social coordination. The first part of this paper emphasizes the importance of agent’s perceptions in Hayek’s and Bacharach’s views of economic decision making and contrasts these views with the complete neglect of perceptions in GEE theory. The second part of the paper investigates how both authors use a specific form of psychology and define beliefs and forms of knowledge in an analytical context where complete explicit information is not the usual and the unique foundation of economic coordination. The last part of the paper argues that Hayek’s and Bacharach’s contributions suggest two different—and partially complementary—routes along which to combine the subjectivity of decision makers with social and market coordination.

1. Why Perceptions Matter in Economics

5For some commentators, Hayek is presented—until 1937 at least (Hayek, 1937)—as a GEE theorist, assuming therefore the homogeneity of individual agents, the absence of intentionality and the rejection of any type of psychological explanations of economic behaviour. Therefore, his contribution would have been in line with Schumpeter’s interpretation of Walras’s “pure economics” interpreted as a “self-contained” or a “closed” approach entirely independent from psychology (Arena, 2006). The prevalence of this point of view on Hayek’s contribution clearly provides a major obstacle if we wish to argue in favour of a permanent interest on the influence of psychology—and especially of perceptions—on economics in his approach, even before 1937. We actually disagree with this point of view on Hayek’s contribution to economic analysis and we tried at length to show elsewhere (Arena, 2003) that there was more continuity than discontinuity between the young and the mature Hayek.

6One of the first and major arguments presented in favour of the interpretation of Hayek as a GEE theorist is Hayek’s famous 1928 article on “intertemporal price equilibrium” (Hayek, [1928] 1984). This article is indeed used in support of a Walrasian-Paretian interpretation of the “young” Hayek arguing that this equilibrium is nothing more than a prefiguration of the neo-Walrasian modern concept of intertemporal GEE (for a characterization of this concept, see for instance Radner, 1991). At first sight, some passages and arguments developed in Hayek’s text seem to corroborate this interpretation and to anticipate the GEE research program of the 1950s and its neglect of a psychological explanation of the process of formation of consumer preferences or of the intertemporal inconsistencies of consumer choice as exemplified, for instance by Ainslie’s hyperbolic discounting (Ainslie, 1975). This first impression is, however, misleading.

7Hayek actually knew rather little of the technical aspects of Walras’s and Pareto’s mathematical constructions. His own use of GEE did not focus so much on the general interdependence of relative prices but rather on the interplay between the expectations of individual agents and the actual realisation of the various economic variables. As we will see later, this interplay was also considered by Bacharach when he investigated the notion of rational expectations (Bacharach, 1989). As Hicks after him, Hayek argued that expectations should be consistently introduced and that they played a central role in economic theory. He was however clearly skeptical about Walras’s theory of tâtonnement to explain the formation and the stability of prices in a world of certainty ruled by an auctioneer or any other type of central price mechanism. This is why when considering what individuals must know for the realisation of an intertemporal equilibrium, he noted “that this will never be so in reality” (Hayek, [1928] 1984, 76). An echo of this focus on individual expectations is also given in Prices and production where he emphasized the “very pressing question of ... what determines the expectations of entrepreneurs and particularly of how such expectations will be affected by any given change of present prices” (Hayek, 1931, 155). This is also why Hayek never believed in a pure theoretical explanation of price stability as the Walrasian one and always emphasized the need to combine economic analysis with historical and empirical considerations (Arena, 1999; 2003) referring to a “causation” based on “a chain of historical sequences” (Hayek, 1941).

8It would therefore be completely misleading to interpret Hayek as a forerunner of Debreu. Our view is confirmed by the fundamental distinction Hayek draws between a one-agent economy and a decentralised market economy. The former serves essentially a pedagogic purpose and permits us to understand deviations from the equilibrium over time at the level of the economy as a whole (Hayek, [1928] 1984, 77). By contrast, the latter concept allows the introduction of a new and fundamental question, namely the analysis of the intertemporal compatibility between individual plans decided upon by heterogeneous agents:

The obvious precondition for an exchange to take place is that, on this, as on all occasions, those engaged in exchange set relatively different valuations upon the goods to be exchanged. That this precondition can be fulfilled follows from the fact that the temporal ranking of subjective evaluations alluded to above relates wholly to the individual, and so different persons can arrive at two completely opposed sets of valuations. (Hayek, [1928] 1984, 78)

9This multi-agent economy therefore excludes a purely Walrasian-Paretian context of usual rational choice associated to homogeneous agents. According to Hayek’s point of view, economic agents are heterogeneous and are guided by “subjective evaluations” related “wholly to the individual”. Hayek therefore stresses the importance of expectations and individual perceptions and replaces Walrasian-Paretian individualism by a first form of subjectivism. As we will show this view is shared by Bacharach. As in Hayek’s approach, Bacharach’s incorporation of perceptions and psychology in economics provides an explanation of agents’ beliefs and subjective knowledge. This explains why he rejected the assumption of perfect foresight (in Bacharach, 1989). A multi-agent economy also entails the presence of strategic uncertainty that, in turn, can lead to the occurrence of disequilibria. Thus, Hayek also introduces the importance of strategies which never appear in a Walrasian-Paretian world and will be taken into account by game theory.

10Actually, contrary to what has too often be argued, before 1937, the majority of Hayek’s contributions refer to the problems of incomplete information and of the heterogeneity of individual agents. But Hayek also introduced the concept and the question of knowledge as soon as 1928. For instance, he noted that if there are “periodically recurring changes in the conditions of production” (Hayek, [1928] 1984, 85), agents can acquire the knowledge required to ensure the persistence of equilibrium. This point of view is again very different from a Walrasian-Paretian context where conditions of production are given and exogeneous, in compliance with the notion of closed or self-contained approach. Hayek distinguishes between three cases of such changes: (i) changes “which recur with precise periodicity” (Ibid.), (ii) changes “which are of uniform tendency in both direction and extent” (Ibid., 85) and (iii) changes “whose unique occurrence can be confidently expected for a definite point in time, as the result of developments which are currently observable or of known human decisions” (Ibid., 85). The first two types of change show agents can acquire the knowledge required to ensure the persistence of dynamic equilibrium and therefore that they are able to learn from the context. The third type of change corresponds to the specific case of empirical and tacit knowledge acquired through the observation of regularities. Here again, we are very far from a general economic equilibrium perspective where knowledge is totally explicit or codified and reduced to information (see Arena and Festré, 2006, introduction). We will see later why tacit knowledge plays a central role in Hayek’s approach.

11The debate on socialist planning reinforces the conclusion that Hayek was already sensitive to the problems of individual knowledge and heterogeneity before 1937. In questioning the possibility of using a general equilibrium framework as a guide to rational decision-making in a socialist planned economy, Hayek casted doubts on the ability of central decision maker to have explicit and codified knowledge of the parameters of the required calculations. He argued that some of the information about a blue-print of productive techniques in the economy was available only in the form of tacit knowledge related to “circumstances” (Hayek, [1935] 1948, 155). In 1940, Hayek provided a convincing example of this problem, showing that it was hardly possible to possess codified and explicit knowledge of real markets and activities. He pointed out that it was difficult for a central planner to define a sufficiently standardized commodity and a sufficiently precise list of supplies and demands to define any given market and therefore to investigate changes in the mechanisms of supply and demand in real time (Hayek, [1940] 1948, 188-189). This view is shared to some extent by Bacharach in his criticism of the GEE view of standardized goods. For Bacharach, “goods” should indeed be replaced by “commodities”, which allows the possibility of a difference between subjective and social representations of the objects of exchange.

12Hayek’s view on socialist central planning are corroborated by comments Hayek made in the 1930s on expectations, refuting the idea that entrepreneurs’ errors could be regarded as a sufficient explanation of crises, unless they could be regarded as “justified errors” (Hayek, [1939] 1963, 443). While “avoidable errors” are caused by external shocks, “justified errors” arise from “guides or symptoms” that prove to be systematically misleading, as is the case with disequilibrium monetary prices.

13It does not, therefore, come as a surprise that in the introduction to Economics and Knowledge, Hayek commented on the continuity between that work and those that underlay his theories of capital and of business cycles:

It has become more and more obvious that in the treatment of the more “dynamic” questions of money and industrial fluctuations the assumptions to be made about foresight and “anticipations” play an equally central role, and that in particular the concepts which were taken over into these fields from pure equilibrium analysis, like those of an equilibrium rate of interest, would be properly defined only in terms of assumptions concerning foresight. The situation seems here to be that, before we can explain why people commit mistakes, we must first explain why they should even be right. (Hayek, [1937] 1990, 29)

14We can therefore conclude that even before 1937, Hayek defended a conception of individual agents which clearly differed from the one which prevailed in the GEE research program. He indeed assumed heterogeneous agents, incomplete information and a first form of subjectivism. Moreover, even if he did not reject the concept of equilibrium over time, his approach did not however appeal to the theory of tâtonnement but to the compatibility of subjective perceptions and expectations in a world where prices were permanently changing.

15Hayek’s post 1937 contributions on economics and psychology confirm the above interpretation (see especially Hayek, 1937, 1940, 1945 and 1960). We shall discuss these contributions in the following sections. However, already at this stage it is important to note that in his post 1937 contributions Hayek reinforced the role assigned to cognitive psychology in his view of market coordination through the convergence of subjectivist perceptions among individual agents, which is a theme common to Bacharach’s conception of coordination. Hayek showed how tacit knowledge and beliefs strongly contributed to help to the creation of the “knowledge of society” to which individuals have no access, and of a market social order in which the dispersion of individual knowledge and beliefs prevail. These views remind of Bacharach’s distinction between individual frames and “universal frames”, agents being only able to take up the former but never the complete set of the latter (Bacharach, 1993; 2001; 2006).

16Contrary to Hayek, Bacharach’s contribution does not stem from a critique of the GEE. However, according to him explaining market exchanges also requires to emphasize (i) that people are motivated by desires for goods and more specifically by “notional desires” and (ii) that these “notional desires” are the actual triggers of market exchange. In this way, Bacharach (1990, 346-347) opens the door to the introduction of psychology in consumer theory. Bacharach (Ibid., 387) indeed clearly refers to “the psychological fact of the notionality of our desires.” Fundamentally, in Bacharach’s approach all of the types of goods are “experienced goods” that he identifies as “commodities” (Bacharach, 1990, 387). In Bacharach’s vision, as well as in Hayek’s approach, consumers are driven by the perceptions they have about the way goods may satisfy their desires. According to Bacharach, these are “notional desires”. Consumers experience satisfaction only when consuming goods. Accordingly, before purchasing, consumers are motivated by the perceptions they have about the eventual satisfaction drawn by consumption; and those perceptions are a matter of framing.

17Bacharach’s conception of framing is very different from the one of Kahneman and Tversky (1979; 1986) that is commonly used in economics and in the “framing effect literature”. Bacharach retains from Kahneman and Tversky’s “prospect theory” (1979) the idea that individuals’ beliefs and preferences depend on their subjective descriptions (representations) of the world (Bacharach, 1986, 183). However, Bacharach distances himself from Kahneman and Tversky and the framing effect literature by being interested solely in natural framing, i.e. in the absence of “manipulations” designed by the theorists to affect individuals’ decision making (2001, 4). Bacharach attempts to understand the “natural” process of framing in order to appreciate the influence of frames on individuals’ decision-making and “practical” rationality in their everyday life.

18Bacharach (2001, 1) defines frames as the set of concepts an individual handles when thinking about the world; it is most simply her view or her perceptions of the world. It is thus analogous to Hayek’s subjective representations of reality:

[I]n order to explain how someone acts, we have to take account of the representation or model of her situation that she is using as she thinks what to do. This model varies with the cognitive frame in which she does her thinking. Her frame stands to her thoughts as a set of axes does to a graph; it circumscribes the thoughts that are logically possible for her (not ever but at the time). In a decision problem, everything is up for framing. The preferences on which she acts, her alternatives … So far from finding herself with given preferences over outcomes, as traditional theory holds self-evident, these preferences depend upon the evaluative concepts that are uppermost in her mind. (Bacharach, 2006, 69)

19By implication individuals’ frames shape their representation of commodities. Therefore, commodities are “things” or “objects” under description (Bacharach, 1990, 351), and description necessarily relies on language:

Commodities depend for their existence on words to express the concepts under which fall consumers’ desires. Call the set of the community’s verbalized concepts its conceptual repertoire. (Bacharach, 1990, 366)

20Understanding how frames influence individuals’ decision-making requires for Bacharach to identify: (i) what are determinants of individuals’ frames, (ii) the process by which frames come to individuals’ mind, and (iii) and the internal structure of frames (Bacharach and Bernasconi, 1997, 5). Bacharach’s purpose is to draw a model of individual decision-making permitting a full account of this mechanism. To this purpose he builds the Variable Frame Theory (VFT) (Bacharach, 1991; 1993; 1995; 1997; 1999; 2006).

21Individuals’ frames, i.e. individuals’ conceptual repertoire—which is structured into families of concepts (Bacharach, 2001, 5)—is based on their “everyday experience”, e.g.—as for Hayek—on their experience of effective consumptions, of market functioning, of trade, of strategic or social interactions on markets, etc. This everyday experience in turn shapes individuals’ “everyday theory”, i.e. the way they theorize the world. Framing is therefore context-dependent (Bacharach, 1991; 1993; 1997; 1999; 2001; 2006). Each context induces the (involuntary) instantiation of a set of pre-existing concepts (or families of concepts), (see, e.g. Scazzieri, 2008, 197; 2011). A stock of concepts is progressively stored according to individuals’ everyday experience and this in turn gradually shapes individuals’ cognitive structure. In other words, this process shapes agents’ conceptual repertoire, i.e., agents’ frames (Scazzieri, 2008; 2011). According to Bacharach

It is reasonable to think that framing would primarily be associated with the cognitive and linguistic ability to grasp specific problem situations through the activation of particular set of “naturally connected” features. (Bacharach, 1997, 196-197)

22These “naturally connected features” are specific to each individual with respect to her experience and her corresponding cognitive frame. Besides, individuals are not “conceptually omniscient” (Bacharach, 1991, 29). As a consequence individuals’ frames are necessarily incomplete (Bacharach, 2001, 5). Framing is to a large extent a matter of attention. Because of their non-omniscience, depending on each individual, certain concepts (or families of concepts) are more salient than others in specific contexts (Ibid.). In this perspective, congruence and similarity are two attention-focusing mechanisms (Scazzieri, 2008; 2011). They may drive the attention on specific features of the world (Ibid., 195).

24Since framing is prior to reasoning (Bacharach, 2001, 5), such an accurate account of the process of framing allows Bacharach to better “circumscribe” the rationale of individuals’ decisions (regarding their frames), and in particular the “possibility space” of individuals’ beliefs and expectations (e.g. of economic variables, of others’ perceptions and eventual responses in case of interactions). Individuals’ beliefs are defined within their frames and may be expressed through propositions that are included within those frames (Bacharach, 1986, 182). The incompleteness of frames therefore necessarily induces “truncated beliefs” (Bacharach, 2001, 9). Thus, individuals’ rationality is “circumscribed” by their frames: “the space of propositions or event on which an agent’s subjective probabilities are defined is always incomplete” (Bacharach, 2001, 5). Subsequently, Bacharach’s account of rationality stands for subjective rationality (Bacharach, 1992, 248). As explained in the following quotation, this statement has important consequences for Bacharach’s account of market exchange.

Let [O] be a predicate [i.e. a concept] of the language of the economy E which denotes the type F of [commodity]. When a member of E believes of a thing that it is “a O”, she not only believes that it is an F but also, crucially, she believes (perhaps only implicitly) to apply to it numerous generalizations describing tendencies of Fs… If she does not acquire a sufficiency of commonplace, tendential beliefs of this sort, she does not understand the sentence “it's an [O]” and she cannot be correctly ascribed the belief that the thing is an [O]. I shall call the network of propositions that embed the concept F and inform in this way competent speakers' beliefs about what they believe to be Fs, the everyday theory of Fs. (Bacharach, 1990, 357)

25The above quoted generalizations are specific to each individual, in a given time and a given place. And since every individual’s frame is incomplete, every individual’s “[e]veryday F-theory is partial” (Ibid., 358). Individuals’ and eventually groups’ conceptual repertoire(s) do not contain all the potential “properties” of the F commodity. However, individuals can learn new concepts, bringing about new families, new associations of concepts, etc. and this process modifies existing frames, or eventually shapes new frames (Bacharach, 1990, 367). Yet, in order to be memorized and enrich individuals’ frames, the eventually added predicates—or concepts—must be activated by new experiences and then enter individuals’ “everyday theory” (see also Scazzieri, 2008).

If a concept F is to enter the repertoire, people must learn the addendum to everyday theory which treats of Fs—they must “do” a new topic; they must often acquire new recognitional capacities; they must enlarge their vocabulary. All these impose cognitive strain. Once acquired, new concepts must be maintained, expensively, in working order. The larger the existing stock, the greater may be the marginal costs of acquiring and maintaining a new concept. (Bacharach, 1990, 367)

26All of this implies that Bacharach does not believe in a unique and “objective” classification of goods and in the homogeneity of individuals’ perceptions of those goods as in the GEE. Like wise, because perceptions of commodities are heterogeneous, “the law of one price” and the optimality of competitive equilibrium in GEE cannot be sustained for Bacharach (1990, 346).

27To summarize, the introduction of framing, i.e. of individuals’ representations in decision-making, has important consequences especially for us in this contribution as it paves the way to the consideration of agent’s subjectivity and heterogeneity as well as to the introduction of the influence of psychology in the realm of economics. This has effects not only for the treatment of individuals’ beliefs and rationality, but also for the role of strategic and social interactions in markets exchanges. We will address this issue in the second and third parts of this paper.

2. Economics and Psychology: Subjective Minds, Cognition and Behavioural Rules

28In this section, we will highlight how both Hayek and Bacharach favour a very specific use of psychology. On the one hand, Hayek builds his theory of human cognition by contrast with the predominant psychological framework of his time. On the other hand, Bacharach combines various subfields of cognitive and social psychology, yet distancing himself from the predominant trends prevailing in these subfields, an attitude which is shared by Hayek.

29The first stage of Hayek’s use of psychology in economics after 1937 is based on the role he attributed to knowledge and beliefs.

30His conception of knowledge is clearly expressed in his Constitution of Liberty:

The growth of knowledge and the growth of civilization are the same only if we interpret knowledge to include all the human adaptations to environment in which past experience has been incorporated. Not all knowledge in this sense is part of our intellect, nor is our intellect the whole of our knowledge. Our habits and skills, our emotional attitudes, our tools, and our institutions—all are in this sense adaptations to past experience which have grown up by selective elimination of less suitable conduct. They are as much an indispensable foundation of successful action as is our conscious knowledge. (Hayek, 1960, 26)

31This quotation of Hayek shows that, in his view, individual knowledge cannot be limited to “explicit and conscious knowledge” (Ibid., 25). It also stresses a perspective common to Hayek and Bacharach which consists in pointing out the role of individuals’ personal experience in the structuration of their mind.

32Hayek distinguishes three main forms of knowledge.

33The first form corresponds to general explicit knowledge. Hayek refers to it when he notes that explicit knowledge is expressed through “formulated generic rules that can be communicated by language from person to person” (Hayek, 1960, 33).

35The third form of knowledge that Hayek refers to differs entirely from the previous ones since it is at least partially tacit and unconscious. Hayek’s consideration of this type of knowledge needs to be emphasized. Hayek distinguishes rule-following behaviour which he characterized as “knowing how” and the knowledge of something which he characterized as “knowing that” (Hayek, 1988, 78). This distinction is also present in Bacharach’s contributions. By contrast with both first types of knowledge, the third form can only be acquired directly because it does not necessarily imply the preliminary definition of an objective of voluntary learning. Day after day, it is absorbed through social interaction. It is not acquired with the help of artefacts or formal institutions but by the permanent resort to rules of social behaviour, the meaning of which is not always and entirely understood by agents. Individual agents, in general, are not conscious to use and implement particular rules in order to face specific daily circumstances. Tacit knowledge cannot be transferred with the help of a codified language. Its existence is directly related to the subjective personality of agents.

36As we have seen above, Bacharach characterizes individual frames as history-dependent constructions and as the source of agents’ subjective knowledge. Agents’ perceptions are dependent on their experiences of market exchanges. Progressively, because of the recurrence of particular contexts, agents become more and more able to conceive general types of situation. This process of generalization also allows them to establish behavioural rules which they will follow in the future.

37More generally, Hayek’s conception of knowledge must be understood in relation to his subjectivist methodology, which essentially rests on two arguments.

38The first, which is cognitive, is found in The Sensory Order (1952). In this book, Hayek champions the idea that according to the connectionist approach, the point of departure of mental representations is not the physical order of things as “scientistic objectivism”—to use Hayek's expression (cf. Hayek, 1952, chapter V)—would have put it, “but the product of abstractions which the mind must possess in order to be capable of experiencing that richness of the particular [of the reality]” (Hayek, 1978, 44). The conscious experiences that individuals regard as relatively concrete and primary and which they attribute to the intrinsic properties of the physical order “are the product of a superimposition of many ‘classifications’ of the events perceived according to their significance in many respects” (Ibid., 36).

39Hayek’s point of departure is Ernst Mach’s sensationalist approach for which the set of human perceptions is constituted by pure and elementary units of sensations. According to Mach, these sensations are coming from the external world and transmitted to the internal world, namely the human body and its nervous system. In this context, the brain is a tool which permits to receive external data. Mach assumes a strict correspondence between external data and internal sensations. However, elementary sensations receive a new organization in the brain. At this internal level, new relations develop and new meanings are generated. Individual perceptions therefore result from two different origins: the external data and the relations set between sensations by the brain.

40Hayek also assumes that there are two “orders” in which the human mind arranges the objects in the world: the physical order, which classifies events as similar or different according to whether they produce similar or different external events; and the sensory order, which classifies events according to their sensory properties. In contrast to Ernst Mach, Hayek however holds that there is no simple one-to-one correspondence between the elements of the two orders:

Since Mach had qualified so many of the connexions between sensations as “relations”, I was finally forced to conclude that the whole structure of the sensory world was derived from “relations” and that one might therefore throw out altogether the concept of pure and simple sensations, which plays such a large role in Mach. (Hayek, 1967, 174)

41In other words, sensationalism has to be rejected as a way to explain the formation of agents’ perceptions. Hayek indeed excludes the existence of a direct correspondence between external events and the organization of the brain. According to him, the explanation of the social world (and the relationship between psychology and economics) should only rely on the internal organization of the mind. External events have only an indirect influence on the mind. The human mind is only a classifier of external impulsions and gives them a meaning. The classification of external impulsions implemented by the brain form a physiological memory, which is the real point of departure of mental phenomena. Therefore, each specific brain creates specific mental phenomena which provide the basis of Hayek’s subjectivism. Agent specific or subjective sensations depend on the position of neuronal fibres which lead impulsions to the set of the central neuronal system. This is a typical connexionist approach stressed in 1982 by Hayek himself:

My conclusion at an early stage was thus that mental events are a particular order of physical events within a subsystem of the physical world that relates the larger subsystem of the world that we call an organism (and of which they are part) with the whole system so as to enable the organism to survive (Hayek, 1982, 288).

42This quotation is an interesting summary of the relation between sensory and social orders. As we already noted, these two orders differ since the first is purely subjective and therefore this is the reason why individual agents are fundamentally unable to understand the working of the overall social order. The Sensory Order also includes an evolutionary dimension which contributes to explain the contents of a real process of subjective learning and helps to understand how subjectivism is compatible with the evolutionary aspects of human and economic behaviour.

43These neuronal foundations also help to understand why Hayek’s subjectivism entirely differs from standard individualism. A comparison between The Sensory Order and Hayek’s theory of the economy and society suggests why for him the social order cannot be interpreted as a direct result of aggregating rational choices. As Hayek noted in 1967,

Such spontaneous orders as those of societies, although they will often produce results similar to those which could be produced by the brain, are thus organized on principles different from those which govern the relations between a brain and the organism which it directs. Although the brain may be organized on principles similar to those on which a society is organized, society is not a brain and must not be represented as a sort of super-brain, because in it the acting parts and those between which the relations determining the structure are established are the same, and the ordering task is not deputized to any part in which a model is pre-formed (Hayek, 1967, 74).

44There are thus as many subjective forms of knowledge as there are individual “nervous systems”, i.e., as there are heterogeneous agents. For Hayek, this approach does not always conflict with some of the tools commonly used in microeconomics, what he called the formal “logic of the choice”. However, he thinks those tools require a serious adaptation.

45The second justification of Hayek’s subjectivism is found in what Hayek called the “social division of knowledge”. For Hayek, as a civilization develops, the knowledge of its society becomes more complex and specialised. However, no individual agent can have access to such knowledge alone. Knowledge is dispersed within society and each of the individuals constituting it can have access only to a very small part of the social knowledge and of the processes by which social and economic activity is regulated and reproduced globally. This conception of dispersed knowledge is similar to Bacharach’s one. As we have seen above, Bacharach (1986) considers that the larger the social “conceptual repertoire” is, the more difficult it is for individuals to have access to it, i.e. to learn and maintain new concepts.

46Hayek claims that the economic behaviour of each individual agent is embedded in the framework of his own theory of how the world works. This idea is also vividly defended by Bacharach (1986; 1989; 1991; 1993; 1997; 1999; 2001; 2006). For Hayek, it means that each individual takes decisions according to his own set of “structural” individual beliefs. These structural beliefs shape what we could call “circumstantial” beliefs, namely, beliefs that govern specific decisions related to particular circumstances and to particular expected economic results at a given point of time.

47These decisions give rise to actions that, in turn, produce results that the agent compares to her own expectations. The divergence the individual perceives between her own expectations and the actual outcomes of her actions leads her to revise her circumstantial beliefs and sometimes, albeit less frequently, her “structural” beliefs. This revision process is stronger when the observation of actual results convinces the agent of the “errors” in his “attempt” and provides her with the opportunity to eliminate some beliefs. If, on the other hand, the 'attempt' is successful, a process of selection of beliefs takes place. Indeed, in the long-run, as pointed out by Garrouste (1999, 891), “structural” beliefs become stronger and are gradually transformed into individual abstract rules, comparable to genuine routines. In fact, the agent becomes aware that her structural beliefs help her to obtain results that are superior to other possible beliefs.

48However, despite the fact that individual beliefs are subjective, they are not independent from those of other agents. This provides the last step of Hayek’s construction related to the transition from individual to collective beliefs.

49Similarly, Bacharach differentiates several forms of knowledge, which are related to his vision of subjectivism. Bacharach (1991, 15) defines two types of knowledge: (i) occurent and (ii) non-occurent. The former entails that “a necessary condition for someone to know that p is that the question of p should come to her mind or occur to her; that she should think of p”. This knowledge is deliberative but nevertheless constrained by framing. It appeals to individuals’ reflection and awareness. Individual frames determine individual inference and epistemic conditions (see Scazzieri, 2011). It appeals to individuals’ reflection and awareness. The later is tacit “but nevertheless action-guiding”. Tacit knowledge is expressed in terms of unconscious rule following. In other words, tacit knowledge corresponds to stored and generalized knowledge enabling individuals, by congruence between similar contexts to follow patterns of behaviors (Scazzieri, 2008, 2011). We can assume that in Bacharach’s framework, tacit knowledge stems from individuals’ “everyday experience” which then, by generalization, forms an “everyday theory”. As we have seen above, Bacharach views agents as non-omniscient and cognitively bounded. In an uncertain world, unconscious rule following behaviour therefore plays an essential role in his characterization of economic behaviours. In some circumstances, the rationale of agents’ decisions relies on their non-occurent knowledge, i.e. on the rule-following form of individual knowledge. Integrating the role of cognitive frames in individual decision-making—in order to provide a pragmatic account of economic rationality—requires to investigate the interplay between short term and long term memory (Scazzieri, 2008; 2011). It requests understanding how short term and long term memory structure individual cognitive frames, and how occurent knowledge progressively generalizes to a non occurent knowledge.

50In addition, the culture in which individuals are embedded endows them with “social” and tacit knowledge. Bacharach (Bacharach and Hurley, 1991, 3) believes in the existence of a “cultural common sense”, i.e. “the fact that every real [individual] has the general knowledge her cultures gives her, such as knowledge of which arrangements are salient or traditional in that culture”. This “cultural common sense” may allow individuals to coordinate or at least can orient them toward an effective decision in an uncertain world peopled by cognitively bounded individuals and in which the cognitive load of computation is to heavy (Aoki, 2001). Thus, as in Hayek, Bacharach’s subjectivism and his interpretation of agents’ heterogeneity combine to their cultural background in order to form a “social order” enabling individuals to coordinate.

51In view of this, Bacharach draws upon psychology to understand how this non-occurrent knowledge is built and stored, and how it can affect individual decision-making.

52For Bacharach, “the study of rationality has much to gain by triangulation from different disciplines” (in Bacharach and Hurley, 1991, 4). This includes, in particular, cognitive psychology and social psychology. Understanding how players’ representations matter in their mode of reasoning and accordingly affect their beliefs toward the way market exchange can be conducted goes beyond the scope of economics. Recall that for Bacharach the role of individuals’ beliefs is central to the analysis of market exchange. Bacharach in fact, declares

There is, I shall argue, a serious gap in our theoretical understanding of how economies work. What is missing is a satisfactory Theory of Belief for individual economic agents.  (Bacharach, 1986, 175)

53In this perspective he uses and crosses various literatures in psychology and social psychology. Alongside the role of framing, a large part of psychology that matter for Bacharach is linked to the role of sociality, and of social and strategic interactions, in human cognition.

54At the time of Bacharach’s contribution to Decision Theory and Game Theory through his Variable Frame Theory (VFT) (1990; 1991; 1993; 1997; 1999; 2001; 2006), the work of Kahneman and Tversky (1979; 1986), and the “framing effect literature”, started to influence the work of economists in these fields. As already mentioned Bacharach’s use of framing differs from Kahneman and Tversky’s “prospect theory” (1979). Indeed, in the VFT Bacharach attempts to offer a descriptive theory of the process by which frames come to individuals’ mind and the way frames structure individuals’ mind. Bacharach develops the variable frame theory from 1991 within the analytical framework of game theory. However, his interest on perceptions—and frames—and the way they drive individuals beliefs, reasoning and behaviours dates back to 1986, the date on which he criticizes the standard economic approach of beliefs and then in 1990 when assessing “the traditional account of commodities” in the GEE, and the analyze of market behaviours (Bacharach, 1990, 346).

55Because frames are structured, different levels of this structure require different mechanisms of activation (Scazzieri, 2008). Bacharach therefore draws on different frameworks in psychology to identify what determine the structure of frames and the different levels of activation that prevail in individuals’ recognitional capacities.

56A first element of framing that matters for Bacharach is “entification" since concepts come in bundles and are classified in families (Bacharach, 1991; 1993; 2001; 2006). In this perspective, he refers to Gestalt psychology (e.g. Wertheimer, 1923; Campbell, 1958) which focuses on visual perceptions. In this approach, ‘entification’ is associated with the following characteristics: “‘contiguity’, ‘common fate’ (moving in parallel over time), ‘good figure’ (forming a recognizable pattern) and ‘similarity’ [...] ‘closeness and impermeability’” (Bacharach, 2006, 70-71). One of the previous characteristics, “similarity”, leads to a second trend of research that Bacharach uses: Post Gestalt psychology (e.g. Campbell, 1958; Tajfel, 1969; Tversky, 1977; Rosch, 1978). This approach defines similarity as the “criterion of grouphood” i.e. “the meta-contrast principle [which] explains categorization, that is, the cognitive activity of dividing a domain of items” (Ibid., 71).

57On the one hand, by referring to Gestalt psychology, Bacharach offers a first element in the structuration of frames. This first element explains the horizontal classification of “things” through families of concepts. On the other hand, similarity is a determinant characteristic to form generalizations in context of uncertainty and to draw generalizations from recurrent practical situations and experiences. Accordingly, similarity between situations, commodities, etc. allows individuals to form their everyday theory and in turn to appreciate their response despite the uncertainty surrounding their decisions (Scazzieri, 2011). This criterion can therefore refer to a second element in Bacharach’s account of the structure of frames: the vertical classification. It explains why some concepts, say some frames, acquire progressively higher level of generalization. It explains why some concepts become deeply anchored in individuals’ cognitive structure. Specific situations involve the instantiation of some concepts or frames. The more those specific—or quite similar—situations occur the more the preexisting associated network of concepts will instantiate, and the more this network will therefore be deeply anchored. In turn, this will make the behavioural response more and more unconscious.

58In this perspective Bacharach’s account of framing brings closer to the post—behavioural—Skinnerian (1985) relational frame theory. This psychological theory is centered around the relationship between networks of stimuli—including visual perceptions, smells, noises emotions etc.—their meaning and associated semantics for people and their behavioural responses. This theory postulates that the different networks of stimuli will be stored in long term memory and any context involving one of the stimuli will provoke the instantiation of the whole network as we previously show for Bacharach’s VFT. Indeed, as emphasized by Scazzieri (2008, 196-197), in the VFT

framing would primarily be associated with the cognitive and linguistic ability to grasp specific problem situations through the activation of a particular set of “naturally connected” features … framing would generally rely upon pre-existing cognitive structures (the different subsets of naturally connected features), but only specific (contingent) circumstances could turn virtual frame into an effective one.

59For scholars in the post Skinnerian account of relational frame theory

the act of relational framing is thought of as a process, an ongoing way of responding to stimuli as they are presented. People frame events relationally in the moment as an active process that is a function of their extensive learning history and stimulation in the present environment. (Blackledge, 2003, 429).

60Besides, the recognitional capacities mentioned by Bacharach (1986) are though as in relational frame theory, i.e.

With a vast amount of training, using multiple relations across many stimuli, words come to share the functions of a wide variety of experiences and events. At first, this occurs through direct training, and along formal stimulus dimensions. After repeated experiences of doing so across multiple exemplars, we learn to bring relational responding to bear on non-formal, or arbitrary, relations between stimuli. Once we do so, our verbally constructed worlds become increasingly complex as we derive more and more relations between virtually every stimulus we discriminate. (Blackledge, 2003, 427)

61Drawing on the Post Gestalt Psychology, Bacharach aims also to show why individuals’ identities matter in their decision-making. For Bacharach the characteristics identified by this framework can refer to groups of humans. Subsequently, and here in contrast with Hayek’s subjectivism, he states that “psychological, cultural and social” similarities can provide the basis for an individual to entify a group of individuals and in turn activate her sense of group identity (Bacharach, 2006). As we shall see in the next section, this feature of human cognition plays an important role since it allows individuals to interact, communicate, and coordinate in their everyday life. However this statement does not mean that individuals’ heterogeneity vanishes like Bacharach’s subjectivism.

Personhood is the resultant, to the extent that it is so constituted, of a set of group identities; more exactly, the person is defined by the intersection of her group identities. But it is only to some extent, since there are plenty of person-defining features which do not correspond to group memberships. (Ibid., 88-89)

62Bacharach refers to the theory of self-categorization in psychology in order to understand how each individual constitutes her own identity through interactions between her multiple group identities. Self-categorization theory postulates that self-identification is a matter of framing. This means that salience in specific contexts explains which of the self-identities are activated (Ibid.). Bacharach relies on Bruner (1957) and Gurin and Markus (1988) to postulate that

Which of my collective personae is activated depends on the current accessibility of the categories to which I belong the relative accessibility of a category depends upon many things, which include the perceiver’s current expectations, tasks and purposes. In human interactions, the accessibility of categories is a special case of the notion of availability of frames at the heart of the variable frame theory of games. (Bacharach, 2006, 74)

64Besides individuals’ group identities are only one of the self-identities that matter for Bacharach. He quotes Brewer and Gardner (1996) who identify three self-identities: (i) “personal”, (ii) “relational” and (iii) “collective” (Bacharach, 2006, 74). An individual’s personal identity is linked “to aspects of her representation of herself that differentiate herself from others” (Ibid.), while her relational identity concerns her “self conception in terms of relationships with other individuals with whom she interacts” (Ibid.). In Bacharach’s conception of sociality, individuals’ relational identity affects their preferences, goals, mode of reasoning, etc. This is explained by the fact that when interacting, the issue of individuals’ decision or behaviour depends on the people with whom they interact, their own perceptions, beliefs and goal. Individuals have to guess in what dispositions are these others—i.e. what are their mental states—to make their own decision. This means for Bacharach that the issue of an interaction is never a priori defined but endogenously determined during interaction. Interactions induce a dynamic of revision of individuals’ frames and beliefs according to those of the others. This matter relies on social cognition and more specifically on the attribution theory within the theory of mind (see the contributions of Fiske and Taylor, 1991; Hewstone, 1983; all quoted in Bacharach, 1989, 180-181). Attribution Theory involves individuals’ second order beliefs, i.e. their beliefs about the beliefs of the other individuals with whom they interact (Ibid., 181). For Bacharach, these second order beliefs intervene in strategic and social interactions, as in market exchanges. In his model of trade, Bacharach

connect[s] two sets of beliefs: those of consumers about the graders of that which they would acquire if they engaged in commodity trade in a type F; and those which make up the “everyday theory” of Fs and so fix the meaning of the predicate 𝜙 that denotes F. (Bacharach, 1986, 347)

3. Social Coordination and Markets

65In considering Hayek’s cognitive approach, Birner (1999, 67-68) underlines the analogy, developed by Hayek himself, between the organization of the human brain and that of society. This marks the limits of Hayekian subjectivism, because “knowledge and beliefs of different people” are not completely diverse but “possess a common structure which makes communication possible” (Hayek, 1952, 49). Even if Hayek’s and Bacharach’s contributions do not resort to the same explanation of this phenomenon, and even if their subjectivisms differ, we may stress that both accept the idea that there exists a common mental structure of individuals which plays an important role to explain coordination. In this perspective:

individuals which compose society are guided in their actions by a classification of things or events according to a system of sense qualities and of concepts which has a common structure and which we know because we, too, are men. [...]. Not only men’s action toward external objects but also all the relations between men and all the social institutions can be understood only by what men think about them. Society as we know it is, as it were, built up from the concepts and ideas held by the people, and social phenomena can be recognized by us and have meaning to us only as they are reflected in the minds of men (Hayek, 1952, 57-58, our emphasis)

66The rules of conduct that guide agents’ behaviour are clearly dependent on the mental “common structure” of human beings. This explains why Hayek is ready to acknowledge that individuals belonging to the same historical and/or socio-cultural environment tend to share common individual beliefs. This stance is clearly similar to Bacharach’s view of the impact of individuals’ social knowledge. Social knowledge allows the formation and acceptance of conventions or institutions and the corresponding stable pattern of behaviours that induce common individual beliefs.

67As some commentators have noted (Ioannides, 1999, 874-876; Garrouste, 1999, 887-891), the features of social rules of conduct that in Hayek underline the formation of a spontaneous social order similar to the one found in market societies, are threefold. First, they must be tacit, that is, “supra-conscious”, to use Hayek’s expression. To put it another way, individuals follow rules of conduct, without knowing explicitly that they are doing so. Second, these rules must be abstract. Third, tacit and abstract rules must necessarily be general. This means that they must be effective for all individuals and apply to an infinite number of particular cases. The content of these rules is independent of the particular individuals who adopt them or of the particular types of actions in which they are put into practice: these rules are the result of a process of adaptation which tends to gradually erase its origin (Hayek, 1960, 27). With the passing of time and the repeated use of individual rules of conduct, their tendency to become more and more abstract and general creates the conditions of their growing autonomy vis-à-vis the individuals who have implemented and/or adopted them. In the long run, these forms of conduct “consist of what we call ‘traditions’ and ‘institutions’, which we use because they are available to us as a product of cumulative growth without either having been designed by any one mind” (Ibid., 27).

68This interpretation is confirmed by Hayek’s analysis of the “properties” that social rules are supposed to possess in order to produce a global order that appears to be independent of individual actions. These “properties” explain how individual rules (even shared rules) can be transformed into social rules, i.e. “normative” rules which “tell” individuals what they can or cannot do. As we already noted, the social process of standardization of rules first results from similarities between individual mental processes. Its origin can also be found in the existence of a “common cultural tradition”. Finally, it can be enforced by, say, the State or the law, and imply sanctions in case of violation. The above remarks confirm that social facts or norms have gradually acquired such a large autonomy with respect to individuals that they appear to them as strictly exogenous (customs, convention, culture, law, etc.). This is why they appear to be the real causes of the process of social standardization of individual rules. Therefore, if shared individual beliefs often imply “true” social beliefs, it is mainly because individuals usually choose their individual rules of conduct within the repertoire of the available social rules. As we already noted, this view is shared by Bacharach. Traditions and institutions imply shared individual perceptions and therefore provide the basis for a common structure of frames. These institutions are interpreted by individuals as exogenous coordination devices because agents think of them as reliable. They are therefore progressively stored in individuals’ “everyday theory” of how the social world works. That is why step-by-step successful coordination devices, by generalization, induce tacit rule following and are incorporated in the deep structure of individuals’ frames.

69It then remains to understand how the mechanism of selection between rules and available social beliefs works. This mechanism is essential since it allows us to understand the emergence of rules or efficient social beliefs. The answer Hayek puts forward corresponds to his theory of cultural evolution:

All that we can know is that the ultimate decision about what is good or bad will be made not by individual human wisdom but by the decline of the groups that have adhered to the “wrong” beliefs. (Hayek, 1960, 36)

70However, one can still wonder how the process of selection of social rules leads to a situation of relative autonomy with respect to individual knowledge beliefs, even if Hayek strongly advocates subjectivist individualism. This autonomy is taken into account by Hayek through his reference to the “knowledge of society” (Ibid., 25). Individuals have no direct access to this kind of knowledge (Ibid., 25). On the other hand, for Hayek, the simultaneous recourse, by individuals, to their own particular knowledge leads each of them to benefit from what one could call positive externalities stemming from social interaction. All that remains is to understand how individuals can indirectly benefit from this knowledge. Hayek’s answer is straightforward: tradition provides agents with a set of rules that are superior or “wiser” than human reason (Hayek, 1988, 73).

71To sum up, Hayek actually allows for the existence of a “knowledge of society”, or “of the impersonal process of society”, which, as already noted, differs from the mere juxtaposition of individual kinds of knowledge (Hayek, 1960, 65). This means that, for Hayek, we must distinguish between two analytical levels (see for instance, Ibid., 28). The first is entirely governed by the methodology of subjectivist individualism and, therefore, only refers to individual beliefs, be they shared or not. The second corresponds to a kind of knowledge that individuals cannot directly access. It is the outcome of the interactive effects of their actions. In order for this impersonal knowledge to be as efficient as possible, it is first necessary that a social process of selection generates rules allowing men to live together in an open society, in other words, in a type of social order that permits individuals to make free but compatible decisions. It is also necessary that these selected rules produce the largest and best ‘knowledge of society’.

72The rules that govern market coordination and selection processes are similar to those that contribute to generate this knowledge. They also set a kind of limit to subjectivism, because, is subjectivism was too systematic, coordination would have become too uneasy to understand. This is why in Individualism and economic order, Hayek notes:

Competition is essentially a process of the formation of opinion: by spreading information, it creates that unity and coherence of the economic system which we presuppose when we think of it as one market. It creates the views people have about what is best and cheapest, and it is because of it that people know at least as much about the possibilities and opportunities as they in fact do. It is thus a process which involves a continuous change in the data and whose significance must therefore be completely missed by any theory which treats these data as constant (Hayek, 1948, 106).

73This quotation summarizes the fact that, according to Hayek, market coordination is only a “case study” of his general theory of cultural evolution. On the one hand, it therefore permits to understand why markets are considered to be optimal institutions. On the other hand, as Vanberg (1986) noted, Hayek never developed entirely and satisfactorily the research program that argues that the market order is the best possible social order.

74Now, according to Bacharach, a pervasive trait of humanity is the ability of individuals to coordinate (and eventually cooperate). For him, real individuals are perfectly able to coordinate for two broad reasons we also met in Hayek’s approach: cultural and evolutionary. First, in an environment characterized by a multileveled society, with multiple types of interactions between heterogeneous agents—or groups, conventions or institutions drive toward convergent pattern of behaviours. A common social background stored in cultural knowledge supplements the non-omniscience of individuals and the bounds of their cognitive abilities in decision-making (see Bacharach and Hurley, 1991; Schmidt and Livet, 2014). When the cost of computation is too high, or the uncertainty surrounding who are the others with whom someone interact, she tend to rely on her knowledge of conventions or social rules in order to decide and to act (Schmidt and Livet, 2014). Second, in long period, the “natural” selection operates at a group level. This selection is in favor of groups which perform well, i.e. structured and organized groups in which the members coordinate and cooperate—through cultural coordination devices (Bacharach, 2006, chapter 3).

76Since commodities are imperfectly characterized or defined in Bacharach’s account of trade, the eventual buyers need to know that there is common understanding of what can be a given type of products—say, F. They need to be sure that everybody in this community shares the same understanding—semantics—of the words determining what is an “F commodity”. This is the necessary condition for a trade to be eventually conducted (Bacharach, 1990, 357).

Your confidence that accepting the k-offer would bring you an F depends in the first place upon your understanding of the offer and your belief in the supplier's understanding of it. (Bacharach, 1990, 353)

78The ST states that to understand others we appeal to “the resources of [our] own minds to simulate [...] others” (Davies and Stone, 1995, 3). We use our own mental states, our own mental scheme, to explain and predict others’ behaviour, by putting ourself in place of them and by simulating, from this angle, their eventual perceptions, intentions, and beliefs. This implies in this perspective that the individuals’ prediction is firmly dependent upon their own perceptions—i.e. their own frames—, feelings, etc. (see e.g. Goldman, 2006). The accuracy of our predictions concerning others’ mental states (intentions, beliefs etc.) depends on the proximity of others’ mental states from our own (Harris, 1995, 226). In order to make accurate predictions toward others’ behaviour, the others’ mental states we incorporate in our decision making must not be too different in the same way as our reasoning process must not be incompatible. Even though, “[e]ven when simulation is insufficient for making decisions in the role of other, it might allow one to discriminate between those options likely to be attractive to the target and those likely to be unattractive.” (Goldman and Shanton, 2012, 10) Besides, a simulation failure can be compensated by resorting to focal points or social conventions. When the uncertainty concerning whom is the other is two high, one device allowing the reduction of this uncertainty is to rely on social conventions, institutions (Aoki, 2001, 13; Schmidt and Livet, 2014).

79Moreover, the above-mentioned empirical results, showing that within a common language community frames tend to be shared, reveal “that beliefs about framing propensities are also shared” (Bacharach, 2001, 9). Furthermore, Bacharach postulates the existence of mutual beliefs (Ibid.).

Among the members of a single-language community, meanings, and so the network of beliefs which fixes them, cannot vary much. There is a core of tendential beliefs such that anyone who believes “it's a k” of an object acquires most or all of them of it. Furthermore, the communicative functions of language are served only because everyone believes (at least implicitly) that other members believing the sentence possess them. (Ibid.) 

82In his work, Bacharach puts also a great emphasis on self-perception or self-categorization as a fundamental process in human cognition. He argues that the way individuals perceive themselves (and others) determines their capacity to coordinate and cooperate (see e.g., Bacharach, 2006, 70). As we have seen, for Bacharach individuals’ group identities matter in their decision-making. Individuals are constituted by multiple group identities which in turn enhance collaborative and cooperative behaviours. This paves the way for the introduction of Bacharach’s evolutionary argument. He postulates “that group identification is the fundamental evolved proximate mechanism for collaboration in man” (2006, 112). He thinks that

[M]an has evolved to have a repertoire of cooperative behaviours geared to different types of situations with scope for cooperation … Call the hypothesis that people are endowed with such a range of cooperative, situation-dependent behaviours the cooperative repertoire hypothesis. (Ibid.)

83Each individual inherits a cooperative conceptual repertoire, which explains why individuals are able to successfully coordinate and cooperate (for instance, to engage in exchange transactions). Finally, selection works at a collective level, so that groups whose members are effective at cooperation have a better chance to perform well and survive in the long term (see Bacharach, 2006).

84Here is, in some way, the basis of Bacharach’s social philosophy. He seems to implicitly support a social vision in which individuals should better be cooperative in order for this community to survive and grow. He implicitly favors pro-social and cooperative conducts because of this collective selection. The more the member of a community rely on “social” focal points, conventions, etc., i.e. the more individuals’ cultural background is developed the more this community is structured and the more individuals will successfully coordinate. And in turn the whole community will benefit. In his work on Team Reasoning, Bacharach (1999) shows that when the members of a given community act by concert for the realization of shared objectives it increases at the same time their expected payoff and the community’s payoff while simply acting as a “benefactor”—i.e. altruistically—is generally less efficient in terms of collective and individual payoff. Altruistic members when acting alone, i.e. without coordinating their behaviour and cooperating with the other members, do not provide the highest possible collective payoff.

4. Conclusion

85Hayek’s and Bacharach’s conceptions of economic congruence and coordination are far from incompatible. Both economists share a close conception of the relationship between subjectivity and psychology. However, their different psychological assumptions lead to two distinct conceptions of social coordination. It would therefore be not such an obvious task to build a synthetic research program including both approaches.

86However, Bacharach’s and Hayek’s emphasis on the need of connecting economics and psychology is today shared by a growing number of behavioural microeconomists and game theorists, as they are increasingly interested in agents’ heuristics and cognitive biases on the one hand, and in social preferences and social psychology on the other hand. We may also stress that Hayek’s and Bacharach’s views on social and market coordination require introducing complexity in economics and therefore considering heterogeneous agents and forms of social interactions much more sophisticated than in the Walrasian-Paretian framework. Both these research tendencies express a strong discontent towards the old general economic equilibrium research program and point to a substantial change in the foundations of microeconomic analysis. From this point of view, the contributions of Hayek and Bacharach had a pioneering role whose implications have yet to be fully explored.

87Hayek and Bacharach suggest that the economic analysis of decision-making and game-theoretical situations should be embedded in a cognitive approach involving philosophical, psychological and sociological backgrounds and the use of the recent discoveries in neurosciences.

88Adopting a cognitive account of individual decision-making can allow integrating the interplay between historical construction of individuals’ mind and a contextual approach of decision-making. They enable to draw economic models based on “situated rationality” (Scazzieri, 2008). Nevertheless, the fact is that even nowadays there are very few economic approaches based on this methodology. From an epistemological point of view, integrating the substantial advances made by both Hayek and Bacharach requires to develop dynamic models in which uniqueness and optimality—of the equilibrium or of the solution—must be abandoned. This epistemological turn is still at its early stage. This requires integrating a new form of strong interdisciplinarity in economics that has just began.

We are very grateful to the participants to the “Economics and Psychology in historical perpective” Conference and especially J-S. Lenfant, M. Rizzo and R. Sugden for their helpful comments and suggestions. The usual disclaimer applies, meaning that the authors alone are responsible for any errors that may remain and for the views expressed in the paper.