Abstract This paper examines the treatment of ontology offered by critical realism. It addresses much of the material elaborated upon in two editions of this journal. Three main groups of criticisms are made here of the critical realist treatment of open systems. It is argued that critical realism, particularly in the project in economics emanating from Cambridge, UK, tends to define systems in terms of events. This definition is shown to be problematic. The exemplar of a closed system provided by critical realism of the solar system is shown to be flawed in that it is not closed according to the closure conditions identified by critical realism. Second, the negativity of the definitions adopted is problematic for heterodox traditions attempting to build positive programmes. Furthermore, the dualism of the definitions is also inconsistent with Dow's approach, which has ramifications for the coherence of post Keynesianism. Third, the definitions tend to polarize open and closed systems and ignore the degrees of openness evident in reality. The polarization of systems leads to polarized methodology and unsustainable arguments to reject so-called "closed-systems methods."
Keywords: open systems, closed systems, critical realism, post-Keynesianism, dualism
Increasingly, the term "open systems" is being used to describe the complex and unpredictable environment faced by economists and economic agents (Setterfield 2003; Lawson passim). "Open systems" has even been advanced as a potential basis for heterodox economics (Hodgson 1988; Dunn 2001; see also Downward 1999), and has arguably become a tacit assumption therein. However, Hodgson (2000) and Mearman (2002a) have argued that the concept is somewhat undeveloped and that, therefore, it would seem inappropriate for heterodox economics currently to be based squarely on it (cf. Dunn; Dow 2000). Even if an "open-systems methodology" remains only one of the pillars of heterodoxy (as in Downward 1999; Lee 2002), it still requires development. This paper considers particularly the ontology of open systems.
This paper aims to develop the concept of "open systems" by offering a constructive critique of the increasingly influential critical-realist view. Critical realism (CR) is, of course, a variegated literature. This variety is not a weakness per se. Indeed, uniformity of meaning can reduce flexibility and prevent change and therefore make an approach analytically poorer. However, while the critical-realist literature is somewhat diverse, a key presupposition of this paper is that there exists a "Cambridge school" view developed by Tony Lawson (1997, 2003) and others, mainly those (mostly at Cambridge University) closely influenced by Lawson. (1) This view has figured prominently in this journal (Winter 1996; Fall 1998). Again, there is variety within this group; however, there is sufficient coherence on key issues to justify the description of the group as a distinct school. Arguably, "open systems" is, along with ontological depth, one of the two most important concepts in this Cambridge school view. Crucially, moreover, this paper holds that CR in economics is dominated by the Cambridge view.
This paper argues that there are three problems in this Cambridge school's treatment of open systems: 1) it is dominated by event-level definitions--which also reflects an underdeveloped concept of "system;" 2) it emphasizes negative definitions; and 3) it tends towards polarizing definitions. These problems are shown to be problematic in many ways: It weakens the ability of CR to engage in constructive work, it raises questions about the possibility of coherence for post Keynesianism, and it leads to polarized methodological accounts on various issues. The paper proceeds in the order of these criticisms.
Before embarking on this critique, one other presupposition should be stated: In criticizing the existing treatment of open systems in CR in economics, an alternative conception of open system is held in mind. It is far beyond the scope of this paper to elucidate fully this alternative conception (see Mearman 2002a, 2004). Moreover, this alternative is not the subject of the paper; this paper does not stand or fall on that alternative. Furthermore, the alternative is as yet merely a sketch. Nonetheless, several key elements of this sketch can be noted. The alternative conception of an open system implicit in this paper (hereafter referred to as the "Open System Ontology," or OSO) is one that exhibits ontological depth, structures and causal mechanisms, multiple and interacting mechanisms, internal relations and emergent properties; it thus owes much to the ontology outlined by CR. However, reflecting the broader intellectual history of the term "open system", rooted, for example, in the General Systems Theory (GST) of von Bertalanffy (1968), the OSO also stresses a system boundary, which is fuzzy and permeable, so that mechanisms can affect the other mechanisms in the system. In such an open system, it is unlikely that strict regularities of events (of the kind "if event X occurs, then event Y will occur") will result; in this sense, the OSO shares the key defining characteristic of an open system held by CR. However, the definition of the OSO is not exhausted by that lack of event regularity; rather, in this alternative conception, a "closure" refers to situations in which, to some extent, the sources of openness, for example, emergence, are absent, are inoperative, have disappeared or have been removed. Openness, therefore, can occur, for example, in the nature of the object, its constituent mechanisms, relations between mechanisms, or in the nature of the system boundary.
DEFINITIONS OF SYSTEMS IN TERMS OF EVENT REGULARITIES
There is a range of critical-realist definitions of closed systems. Closed systems are variously defined as being "cut off" from external influences (Collier 1994: 128; Bhaskar, 1978: 69); "isolated" (C. Lawson, 1996); where outside factors are "neutralise[d]" (Collier 1994: 33); and in which all disturbances are anticipated and "held at bay" (Lawson 1997: 203). The net result, according to Collier (1994: 33), is that one mechanism alone operates, unaffected by other mechanisms (Lawson 1997: 203). (2) The obvious example of such a scenario is the experiment. Thus, Collier (249-50) and Lawson (1999a: 216) effectively equate closed-systems methods as experimental (and open-systems as non-experimental). (3) Strictly, the focus on isolation is incorrect, since experimentation imposes requirements inside the isolated area. Archer (1998: 190) notes that even in isolated environments, the problem-solving and perhaps capricious nature of humans means that the "closure" of experimentation cannot be achieved (see Mearman 2004). These problems are such that Bhaskar (1986: 101) claims that closed systems are "impossible" in social science.
Additional (partial) definitions of a closed system in CR are that relations within a system are stable (Beed and Beed 1996: 1099); and that conditions are imposed on the individuals in the system (Bhaskar 1978: 69). A fully closed system is where all individual and system criteria for closure are satisfied (Bhaskar 1978: 104), which suggests both a regularity of behaviour (Bhaskar 1978: 253, n. 1) and a homogeneous--unchanging and uniform--environment (Lawson 1997: 218); consequently, transformative action is impossible in a fully closed system (Bhaskar 1986: 31). According to CR, closure is achieved when specific closure conditions hold. The most significant of these are the Intrinsic Condition of Closure (ICC) and the Extrinsic Condition of Closure (ECC). The ICC requires that the object in question has a constancy or constancy of change, such that elements "inside" the system are stable enough to be identified. The ECC entails that no outside forces impinge on the particular object or system, or that any external effect is constant. It is significant that a strict distinction is usually made between "closures" and the ICC and ECC, the "conditions for closure". Both the distinction between closure and the conditions for it, and the closure conditions themselves, are prevalent in the "Cambridge" treatments of closed systems.
This paper argues that despite the apparent variety of definitions--which, to repeat, is in many ways to be valued--the Cambridge school definition of open and closed systems has been, and is being increasingly, restricted to one. That is, closed (and hence open) systems are defined in terms of events and their regularity.
Consider the following definitions of closed systems: Closed systems can be identified when the symmetry (of explanation and prediction) thesis holds (Sayer 1992: 130), or where there is a warrant for education (inference to particular instances) (Bhaskar 1986: 30). A closed system also means that there is a one-to-one relationship (isomorphism) between mechanisms and events (Lawson 1994a: 517), which implies the main definition of a closed system: a unique relationship between antecedent and consequent (Bhaskar 1978: 53); a stability of empirical relationships (Collier 1994); or a constant conjunction of events. This definition flows from the Humean argument that only regular successions between events--not underlying mechanisms (should they exist)--can be identified. Hence, causality is conceived as merely correlation, which in turn calls for the identification of event regularities between isolated atomistic states (Rotheim 1999: 73).
Lawson identifies closed systems as being where the formula "if event of type X occurs, then event of type Y will occur" (where X and Y can be scalars, vectors or matrices--Lawson 1994a: 507, n. 9). More recently, Lawson has modified this definition further to take into account the common practice of completely specifying the conditions under which closure holds. Thus, closed systems conform to the formula, "if X, then Y, under conditions E" (Lawson 1989a: 63, 1995a: 15). Such event regularities could be either deterministic or probabilistic (Lawson 1999b: 273), which in the latter case means that events will be in regular succession within some well-behaved probability distribution (Lewis and Runde 1999: 38; Lawson 1997: 76). (4) In this case, the closure is stochastic (Lawson 1997: 153-154). Lawson repeats this event-level definition in his most recent work (Lawson 2003: 5, 15, 23, 41, 103, 105, 143, 222, 306). (5)
Moreover, instances of this event-level definition of closed systems in terms of constant conjunctions of events have appeared in this journal: Lawson (1996: 407, 1998a: 359, 369), Pratten (1996: 439), C. Lawson (1996: 451, 459), Lewis (1996: 487) and Rotheim (1998: 326, 329-331) all use the definition. Of course, all of those authors affiliate in some way to Cambridge University. Clearly the event-level definition is not the only one offered by CR. However, it is argued that this is beginning to be the dominant definition, particularly in economics, and even more so for those (mainly at Cambridge) influenced by Lawson. This is shown most clearly in the definition of an open system.
Bhaskar (1978) defines open systems as the lack of "regular" (p. 33) or "invariable" (1978: 73) succession; no unique relationship between variables (1978: 53); or a non-invariance of empirical relationships (1978: 132). For Sayer (1992: 122), openness entails short-lived or non-existent regularities. Essentially, an open system is identified as, "Not 'if X then Y'", or as where there are no constant conjunctions of events (Bhaskar, 1989: 16). This definition has been adopted by the recent literature on CR in economics. Therein, open systems are mainly defined as where there are no event regularities (Pratten 1996: 423, Rotheim 1999; Lawson 2003: 79, 82, 119, 223-224) (6) or as systems lacking sharp (i.e. precise) stable event regularities (Lewis and Runde 1999: 38).
It was argued above that the Cambridge school of CR in economics mainly defines open/closed systems in terms of event regularities. Immediately, to pre-empt any critique, it could be argued that if this is how CR defines open systems, that is the end of the matter. However, this paper holds that such an argument is problematic because it ignores deficiencies in that definition of open systems. Moreover, as stated in the first section, there is an intellectual heritage of "open systems" that should be acknowledged and respected; furthermore, that heritage differs considerably from the critical-realist treatment. The existence of this heritage clearly raises the issue of whether CR could legitimately claim ownership of the term. Perhaps more significant is the issue of the relative merits of the competing conceptions. An implicit argument of the paper is that competing conceptions of open systems--some of which influence the OSO of the first section--have merit and may be superior to the critical-realist treatment. However, that discussion is beyond the scope of this paper.
The paper restricts itself to examining the critical-realist definition of open systems. Before showing how the critical-realist definition is problematic, one other point should be made. Pratten (1996: 426) writes, "In critical realist contributions such regularities are referred to as closures." Further, Pratten p. 431) criticizes neo-Ricardian economics for its use of "givens" as closures, where such givens are dependent on regularities. Rotheim (1998: 331) also claims that closures are constant conjunctions of events (see also Lawson 2003: 41). However, this equation of closures and event regularities can only be correct circularly, i.e. by defining closures as event regularities. However, to define closures as regularities is inconsistent with the definitions of closed and open systems (above) from CR, which stress the nature of the object, conditions placed on it, its location, its being cut off, or situations in which relations within a system are stable (Beed and Beed 1996: 1099). Under these definitions, it cannot be correct to define closed systems as event regularities. Rather, an event regularity is not equivalent to closure: it is suggestive of closure.
As evidenced by the ICC and ECC, closure occurs beneath the level of events: the nature of the objects and their relation to other objects are the defining factors in creating a closed system. For example, if mechanisms are of a particular type; if their actions are uniform; if their structure is of an unchanging form; or if they are effectively isolated from other mechanisms, they can be regarded as closed, and an event regularity will be generated. This ontology of closure beneath the level of events is consistent with other CR accounts (as above). It is consistent with GST and other narratives of open/closed systems. Unsurprisingly, it is also consistent with the OSO sketched in the first section. These accounts stress both closure beneath the level of events, and the notion of the "system".
This confusion of closure and its evidence is potentially serious. The claim conflates the empirical with the real; this is known as empirical realism, a flattening of ontology. It also suggests that the epistemic fallacy has been committed: what exists is reduced to what is known. It also suggests actualism, defined as the denial of the existence of underlying mechanisms and acknowledges only actual events or experiences (Collier 1994: 7). Empirical realism, the epistemic fallacy and actualism are all explicitly denied and rejected by CR. These contradictions arise here because of the event-level definition of closed systems.
It could be argued that the present argument has underplayed the ICC and ECC in the critical-realist account of open systems. Specifically, it could be argued that Lawson et al. are fully aware of both closure conditions, that the closure conditions occur below the level of events, and that therefore, the views of Pratten and others above--which clearly state an equation of closures and event regularities--are anomalies and that the Cambridge view does not equate regularities and closures. It is an empirical question as to whether Pratten's view is anomalous or is in fact widely held. (7) However, this paper holds that the Cambridge view mostly makes a clear distinction between closures and conditions for closures: Therefore, the ICC and ECC are not generally considered as closures.
Several other defenses of the "Cambridge approach" can be offered. First, the use of the event-level definition might be defended on strategic grounds. Bhaskar (1979: 138) maintains that CR is the only philosophy of science that takes constant conjunctions of events as neither necessary nor sufficient for explanation in natural and social science. In economics, Lawson and others have argued that the discipline is dominated by methods which unjustifiably presuppose the existence of event regularities; they have identified the search for event regularities as the sine qua non of orthodox economics. (8) Econometrics is the best example (Lawson 1989b, 1997, Chaoter 7; Pratten, 2005). Therefore, the event-level definition has high rhetorical value. Strategically, it is probably simpler to organize a coalition of heterodox economists around a critique of event regularities (and the associated (neo-classical) methodology of predictive tests and mathematical modeling at the expense of explanation), rather than enter more controversial territory beneath the level of events. (9) Therefore, it is to be expected--and it is consistent with the aims of the critical-realist project--that the Cambridge group emphasizes the event-level definition. Second, it might be argued that the classification of systems in terms of event regularities merely reflects the Critical-Realist logic of retroduction, which begins at the level of events and moves downward (in ontological terms) to the level of the real, generative mechanisms. The centerpiece and genesis of CR is the transcendental deduction of the stratified, open nature of the world from the experimental activity of scientists (Bhaskar 1978). The actual activity of experiment in natural science (and the relative failure of experimentation in social science) shows the generality of openness and effectively allows CR to presume the presence of open systems. This presumption holds unless there is evidence to the contrary, namely event regularities, however local or brief. Thus, again, it might be expected for CR to focus on the level of events.
Third, it might be argued that the existence of event-regularities is key: that the use of the term "closed system" is incidental. However, if the use is incidental, why does it occur? Lawson (1989b, n. 11) claims that the concept owes much to GST (von Bertalanffy, 1968). However, it is clear that GST and CR differ greatly. For example, the former stresses the existence of entropy as defining a closed system, while the latter makes virtually no reference to entropy. Presumably, then, the influence of GST is by analogy; but if CR's definition of closure is contrary to GST's, is the analogy rendered inappropriate? Clearly, analogy does not require identity: The analogy would have no work to do; however a contradiction would seem problematic. For example, entropy (and closed systems) is associated with disorder and presumably a messy event level, whereas for CR the closed system is defined in terms of event regularity. This difference does not make CR incorrect, but it does raise the issue of the basis on which CR claims ownership of the term "closed system." (10)
Furthermore, neither of these arguments in defense of CR (if they stand) would justify the apparent ignorance of the level of the real in the definition of the open/closed system. Critical-realist methodology is two-sided. Certainly, in critical-realist practice, phenomena, be they crises, localized event regularities, or rough and ready patterns of events (so-called "demiregularities"), etc., are usually the starting point in investigation. From the empirical level, real mechanisms are retroduced. However, retroduction is followed by the empirical assessment of these hypothesized mechanisms, which requires that the empirical is reconstructed from the real. Thus, some attention must be paid to the real level, to determine how those mechanisms produce the empirical. This consideration of the real, in turn, requires extensive deliberation on the workings of the system in question, i.e. its real causal mechanisms set in structures; and its structure, boundaries, etc., to illuminate exactly how the empirical is generated. Therefore, there is a need for a definition of the open system, which includes both the real level and the event level. (11)
Mearman (2002a,b) argues that the above problem stems from the use of the term "system" in (particularly the Cambridge school of) CR, which effectively ignores the two-part nature of the term. Specifically, an open system is a system that is open. CR's notion of the system is relatively underdeveloped. Indeed, Bhaskar (1978: 73) claims that system "carries no independent semantic force". (12) This seems to prove the point. Moreover, there is no clear picture of the term in CR (13): At times in CR, "system" refers to mechanisms, or the structures wherein they reside. Under some definitions, this is accurate, since an open system is one in which a mechanism operates but is subject to the operation of other mechanisms. However, this definition ignores the fact that the mechanism is at the real level, but generates events at the empirical level. Both of these levels are part of the system; thus, just as the system cannot be reduced to the events, nor can it be reduced to the mechanism. Moreover, when Lawson (1999c: 5) writes, "The aim [of experiment] is to engineer a system in which the actions of any mechanism being investigated are more readily identifiable," he differentiates clearly between a system and a mechanism. Lawson (1994b: 279) identifies system and structure separately, further weakening the equality. In short, the notion of system in CR seems underdeveloped. As argued above, this makes constructive research and the reconstitution of the real more difficult. (14)
The problem of the event-level definition can be illustrated by reference to the claim, common in CR, that the solar system represents a rare example of a closed system found outside experimental control (Bhaskar 1978: 65; Lawson 1996: 407, 411; Runde 1996: 472-473; Lawson 1999c: 4; Pratten 1996: 423).
This claim follows because the solar system exhibits, or at least approaches, complete event regularity. However, if one criterion for closure is, as stated above, that the symmetry thesis holds, the solar system cannot be a priori a completely closed system, since there is the possibility that an unpredictable asteroid could disrupt planetary motion. Runde (1996: 474) recognizes this: "Of course, even the regular movements of the planets is itself contingent on the planetary system remaining undisturbed (and by most accounts, eventually, the system will be disturbed). But it is a system that, relative to our own life histories, changes so slowly as to be imperceptible." It is commendable that Runde considers the issue; however, his claim about imperceptible change is inconsistent with CR. To claim that because one perceives the universe as stable, it must be so, reduces existence to knowledge and commits an epistemic fallacy. Furthermore, Runde classifies the system according to what has happened, rather than what might potentially happen. This reasoning is somewhat inevitable, because we rely on ex post descriptions; but his is a description of the events only and ignores the potentialities within the objects of the system. His reasoning suggests actualism.
Moreover, Runde has shown that the solar system is not closed: It is at least potentially subject to disturbances. Thus, the ECC does not hold. Yet the ECC is held to be necessary for a regularity and hence a closed system. Only by defining the system in terms of past events--i.e, that, in the past, disturbances have not occurred and have not, therefore, changed the pattern of event--does the ECC hold. Moreover, the ICC does not hold either. Specifically, an assumption of the ICC in the solar system would require the underlying constancy or constant rate of change of the entire system and indeed the universe. However, such an assumption would be bold: Monastersky (2002) claims that proponents of the so-called "inflation" and "M" theories of physics agree that the rate of expansion is unknown. Furthermore, Collier (1994: 244) admits that cosmology studies changing entities--new tendencies emerge as the structure of the cosmos changes. For example, he notes, "Big Bang" theories postulate different mechanisms operating immediately after that event. Of course, assuming a constant rate of expansion of the universe seems reasonable; however, it entails an assumption that is known to be quite possibly at odds with reality. When others do this, CR accuses them of instrumentalism (Lawson 1989b). The key point is this: in the solar system, neither the ECC nor the ICC seem to be satisfied; these conditions are necessary for a closed system; yet CR claims that the solar system is a closed system because of the regular actual recorded movements of the planets. The claim appears to be actualist, contrary to depth realism. Clearly, the event-level definition of systems is limited and limiting.
Finally, it should be noted that Lawson (1995b: 267) leaves open the possibility of local closures, even if by "chance", i.e. even if there was no prior basis to believe that regularity determinism or stochasticism held (see also Lawson 2003: 15). This admission is significant. If closures are defined as "if X, then Y," by chance, then a system can be called closed, without knowing anything about that system except its outcomes. Moreover, as a general case, it is possible for two external forces, acting on the mechanism inside a system, to exactly cancel out each other. If the nature and action of the mechanism inside the system were constant, such that, when triggered, it always produced the same effect, the outcome would be to produce a regularity. Yet the system is clearly open (in a broader sense) because of the impact of the external mechanisms. Thus, again, "not, if X, then Y", or perhaps, in terms of causality, "if X then not Y" seems not the best way to define an open system.
As a constructive point, there are several possible ways to prevent the problems highlighted here. One simple way is to define closed system as an event regularity and to use a different term to denote the underlying segment of the world in which that regularity obtains. (15) In that way, at least, the intellectual capital accumulated by CR could be protected. The confusion apparent over the term "open system" in the evidence presented above could be reduced or removed. It should be remembered that an even wider range of uses of a term "open system" exists outside CR than within it. Is this solution viable? It affords intellectual priority to the critical-realist claim which might, because of the wider usage of the term, be invalid; and, more seriously, it perhaps allows the persistence of a problematic term. Rather, this paper argues, a more positive definition of an open system is required, perhaps closer to that sketched in the first section. The need for a more constructive or positive definition is bolstered when one examines the way in which negative definitions of open systems predominate in the Cambridge view, as will be examined next.
NEGATIVE DEFINITIONS OF OPEN SYSTEMS
Several of the definitions of open systems offered above were of the form "if X, then not Y" or variations on that. Here, these are called negative definitions because they stress an absence of a condition, specifically of a regularity. A positive definition would be classed as one that stresses the presence of some particular. In addition to the negative definitions, presented above, open systems have been defined in CR as where closure conditions fail to hold (Downward 1999: 17), or that internal and external parameters are non-constant (Sayer 1981: 138). In partial defense of the "Cambridge school" view, other definitions of openness are similar. Kaldor (1972) suggests openness, in his concern that, contra orthodox models, constraints (for instance on consumers) would not be binding (see Hahn 1989: 55). Grunberg (1978: 542) equates openness with a lack of constants (and with complexity) and therefore with the inability "to ascertain invariant relationships." Keynes (1973: 262-263) conducts thought experiments on the effect of money wage reductions in "closed" and then "unclosed" systems, which are national systems affected by foreign economic factors.
Dow (passim) defines an open system, again, effectively as "not closed," as does Downward (1999). Dow offers essentially the reverse of her definition of the closed system. Thus, in an open system, not all constituent variables are known, structural relations are not all known or knowable, and traditional logic is not applicable (Dow 1996: 14). Relatedly, Olsen (2000) implies that an open system as being incomplete, or not fully specified by the theorist. This definition mirrors Setterfield (2003), but also some orthodox definitions, which define a closed system as complete, where all variables are modelled Hendry (1995: 310).
In fairness, CR does offer "positive" definitions. For example, Collier writes, "In open systems ... a multiplicity of mechanisms is operating, conjointly bringing about a series of events, which would not have been brought about by any proper subset of those mechanisms" (1994:43-44). Thus, outcomes are complexly co-determined (Collier 1994: 62) by a "plurality and a multiplicity of causes" (Bhaskar 1978: 72). Therefore the same mechanism can lead to different outcomes (i.e. rather than "if X, then Y," the result would be "if X, then any one of [Y.sub.1] ..., [Y.sub.n],"); and an outcome can be produced by a number of mechanisms.
Several authors, for example, Lewis and Runde (1999), have argued that openness of a social system can be identified by the existence of real choices for individuals (in the sense that the outcome of the choice is not predetermined). (16) Bhaskar (1979:114) defines choice as where "the agent's activity makes a difference to the state of affairs that would (normally) otherwise have prevailed". This conception contrasts to orthodox economic models of the consumer. Given preferences, prices, income, the assumed rationality of the consumer, and the assumed goal of utility maximisation, the rational agent in orthodox economics has no real choice: There is only one possible outcome for the consumer; moreover, they are assumed to be unable to change any of the variables relevant to their decision (DeUriarte 1989-1990). Additionally and significantly, Lawson (1995b: 265) argues, "social structure is human agent-dependent: it is only ever manifest in human activity. Thus, given the open nature of human action, the fact that any agent could always have acted otherwise, it follows that social structure can only ever be present in an open system." Clearly, social structure would disappear without humans--although, it is not created by the specific humans present at that time (Archer 1995)--but to assume this, it is necessary that humans have choice. Furthermore, much work has been done in elaborating the social ontology of open systems, in terms of, for example, emergence and internal relations. This work has mostly come from sociologists informed by CR.
In spite of the large apparent variety of definitions presented, it is argued that in economics, the definitions that dominate the discourse of open systems are negative; and that this is particularly the case with regard to the "Cambridge school" of CR in economics. The reader should refer back to the event-level definitions presented in section I. To reaffirm this, it should be noted that Lawson (2003), while offering definitions of open systems in terms of, for instance, multiple mechanisms (above), he also defines open systems in terms of unpredictability (2003: 100), unsusceptibility to closure (2003: 62), lack of event regularities (see above), and the impossibility of experiments (2003: 84). Moreover, most of the positive, i.e, not simply the opposite of a closed system, definitions which exist, such as by Dow (1996) and Kapp (1968), are from outside CR. They also go further than the negative critical-realist definitions above, which remain basically at the level of events, by discussing the domain of the real, specifically the nature of the structures to be found there. (17)
It has been argued that the dominant "Cambridge view" definitions of open systems are negative. The key issue is whether this negativity is problematic. Arguably it is the nature of argument that the development of concepts involves negative definitions. Clearly, it is common to define an unfamiliar object in terms of the familiar. Moreover, in the development of CR called Dialectical CR (after Bhaskar 1993), a key element is that absences can be causal and that their existence is significant. However, this paper finds the negative definitions to be problematic for two main reasons. First is the issue of the nature of heterodox economics. It is the nature--indeed, the literal definition--of heterodoxy that it opposes the current orthodoxy. Moreover, there is nothing to preclude, say, Marx, Keynes and Veblen, from being different, yet opposed to the same thing. As noted above, there might also be a strategic benefit in uniting around opposition to a particular issue, in order to create an ideological bridgehead in a debate. The desire to persuade others of a view can often be assisted by a negative description of an alternative. Where the rallying point is simple (for example, an event-level definition of openness), it might be even more effective. However, the strategic benefit of uniting around a negative concept might be lost if that act of unity leads to the positive in the program being obscured. The (persistent) danger for heterodoxy is that its would-be critics often reduce its definition to being "not orthodoxy" or "not neo-classicism" (cf. Walters and Young 1997; Lawson 1994a). Clearly, heterodox economics is more than this: Marx, Veblen and Keynes (for example) were all involved in criticizing the orthodoxy, but they also offered criticisms of contemporary society, and (all perhaps except Veblen) offered a constructive alternative program. Nonetheless, the negativist perception persists and is damaging to the heterodoxy. (18)
The second problem of negative definitions relates to their substantive consequences. The mode of construction and development of the definition are significant. Often a "dualist" process occurs. From Dow (1996: 16-17), dualism is, "... the propensity to classify concepts, statements and events according to duals, as belonging to only one of two all-encompassing, mutually-exclusive categories with fixed meanings". The unfamiliar is defined in terms of the familiar by placing it in opposition to it. Often the similarities between the two are ignored. For example, one might define irrationality in terms of rationality, missing intermediate concepts. Indeed, as Mearman (2005) argues, a central point of Dow's (1990, 1996) work is that such dualism leads to errors. The argument here is that the standard realist definition of openness is often dualistic, i.e. it has an unfamiliar concept, "open systems," and a familiar (via orthodox economic and scientific practice) concept, "closed systems."
Hence, definitions of open systems tend to begin with definitions of closed systems. Mearman (2005) shows that even when the concepts are poles, i.e. defined in terms of each other, as in the case of open and closed systems, the relationship between the two terms can be severed and two strictly distinct categories can emerge. No claim is made here that CR, or CR in economics, always and everywhere engages in such thought. For example, as the fourth section argues, the concept of "demi-regularities," which falls between complete (event-level) closure and complete openness, and which is a significant development of CR, was developed in economics by Lawson. However, it is argued that this dualism occurs often enough in critical-Realist argument regarding open systems for its implications to be considered. In any case, it is helpful to be alerted to the potential dangers of such dualistic reasoning.
There are two general and serious consequences of this dualism for CR. First, Mearman (2004, 2005) argues, following Dow, that the conditions for dualism are usually not met in open systems. Dualism makes both ontological and epistemological presuppositions, such as that objects are atomistic, relations between objects are external, and therefore that they can be characterized as distinct and therefore potentially characterized as mutually exclusive. Moreover, fixed categorizations require assumptions of certainty and infallibilism. Clearly none of these conditions are normally found in open systems. Thus, by adopting a dualist definition of open systems, open-systems proponents are engaged in closed-systems thinking. Of course, dualism is not incorrect per se: a door might be open or closed; however, it could also be "ajar". Crucially, though, it is held that very often dualistic categories are incorrect. For example, dualistic thought about rationality/irrationality might be inaccurate, in that humans do not fit neatly into either category. Equally seriously dualism can rule out the development of and/or eliminate useful possible categories; for example, bounded or situated rationality (see Lawson 1998b).
A second consequence of dualism, therefore, is a conflict between Dow's position and CR, which has implications for the goal of coherence for post-Keynesianism (cf. Mearman 2001a; Downward 1999; Dow 1999) and which, therefore, reinforces the problem of negativity identified above. Thus, for these two reasons, more positive definitions of open systems are needed. From the second section, these positive definitions should not be restricted to the event level.
Before moving on to the fourth section, two defenses of CR should be noted. First, if it is acknowledged that CR in economics does tend towards negative definition--and this, naturally, remains open to question--it could be argued that CR in economics is merely in a "negativist" phase, perhaps for strategic reasons, and that it is now moving out of that phase. Lawson (2003), given his generally more constructive tenor (cf. Dow 2004), perhaps provides evidence of a new positive phase. If this is the case, this author applauds that movement. It is imperative that the movement towards positive definitions continues. Second, it could be argued that, in fact, CR is not dualist in its treatment of open/closed systems. Again, that is an empirical question, which shall be considered in the fourth section.
POLARITY OF OPEN/CLOSED
The argument in the third section implies a conflict between Dow's position and CR on the question of a dualist categorization of open and closed systems. However, this is only strictly the case if the categorization is invalid in reality: Some dualistic categories might in fact be correct; however, for Dow, that is either unlikely, and/or there are not epistemological grounds for arriving at those certain categories. So, the immediate question is whether open/closed systems are treated as strictly separate. For, often in CR, it seems that, if a system is not completely closed, then it is inescapably open, rendering closed-systems methods totally impotent. For example, Lawson (1999d) insists that econometrics is only valid in strictly closed systems. Thus, there is a need to investigate two points on two different levels. It needs to be established whether CR is unjustifiably dualist in its treatment of the distinction of open-systems/closed-systems 1) ontologically and 2) methodologically. If CR is dualist on either count, this is problematic, for the reasons given earlier. If the treatment is dualist on one count but not on the other, then this is a disjuncture between the two, which seems problematic for a realist perspective.
Open/closed systems might appear to be a clear dual, given that the two concepts are defined most often simply as opposites of each other. (19) This matter can be investigated further by examining the polar extremes, perfect openness and perfect closure. Recently, the polar view has been suggested by, for example, Rotheim (1999: 75). Lawson (1994b: 276) suggests "two extremes--strict event regularities or a completely non-systematic flux--merely constitut[ing] the polar extremes of a potential continuum". Later (p. 277) he proposes "a continuum of outcomes ... ranging from closed systems of constant conjunctions of events to an inchoate random flux". Consistent with the second section, the definition is in terms of events. Arguably, it is unlikely that either extreme actually exists in reality. First, Collier (1994: 33) claims, "no system in our universe is ever perfectly closed". Above, it was shown that even the solar system is not closed. Therefore, the prime example of naturally occurring closure given by CR is invalid. Lawson (1997: 203) demurs somewhat, claiming, "the goal of perfect closure ... cannot always adequately be engineered; indeed it may very rarely be".
At the other end of the spectrum, critical-realist authors have clarified that openness refers neither to a complete arbitrariness of events (Rotheim 1999: 75) nor to an inchoate flux (see Lawson 1994b: 276). Indeed, Cottrell (1998) criticizes CR, in that significant regularities are in fact found in the social world. An example might be that one works and then gets paid. Lawson (1998a) replies that, indeed, people go to work and are paid after working; but they can go home and perform the same activity for no pay. Therefore, the strict regularity "if work, then get paid" fails to hold. (20) That is not to deny that much of the time "if work, then get paid" does hold, but not always and everywhere. Other reasons might be that people work voluntarily, or that some crisis occurs which prevents payment (Argentinean public service workers and Iraqi soldiers are recent examples). In reality, this is the openness that Lawson (et al.) discusses, not inchoate flux. Again, all these arguments are presented in terms of event regularities, supporting the argument in the second section.
Practically, there is no prospect that either perfect openness or perfect closure exists. Between the two theoretical extremes lies everything of practical interest. Nevertheless, the language of CR has in general, particularly in its earlier work, focused on the extreme cases. CR accuses orthodoxy of clinging to methods based on the perfect (non-existent) closed system. However, CR also uses the perfect (but unachievable) closed system in order to construct its alternative. The contrast between on the one hand astronomy, and on the other, every other discipline, serves a rhetorical purpose and has rhetorical value. Rather than envisaging the spectrum of open--closed systems as a continuum, bounded by theoretical if not practically attainable extremes, critical-realist treatments tend to begin with the notion of a (perfectly) closed system and look for instances whereby the event regularity fails to hold. Where this is the case, the system is classified "open". However, there is clearly a difference between a system in which there exists a mechanism that occasionally operates, whereas the system is otherwise stable; and one in which there is a chaotic mess of sporadically active mechanisms, continually combining in novel ways. In the former case, there remains a good chance of developing knowledge, whereas in the latter, that chance seems remote. However, both would be called "open" and both would fail to exhibit event regularity. Again, the discussion is in terms of events (second section) and negativity (third section).
Of course, CR is not a monolith: Even if CR has predominately focused on the extreme cases of perfect openness and perfect closure, in fairness to it, it should be acknowledged that at other times, critical-realist treatments have acknowledged the existence of such "partial closure". Partial closure can have a variety of meanings. Sayer (1992: 124) defines quasi-closed systems as "producing regularities that are only approximate and spatially and temporally restricted". The definition suggests two types of partial closure. One is defined in terms of spatial or temporal specificity. This definition suggests a large, open mass segmented into smaller closed systems, which is close to the critical-realist notion of "local closures". This is a sense in which CR often speaks about partial closure. Bhaskar (1978: 78) argues, "for experimental science to be possible the world must be open but susceptible to regional closures". Similarly, "A closure is of course always relative to a particular set of events and a particular region of space and period of time" (Bhaskar 1978: 73). This treatment corresponds with Sayer (1992) above. This notion of historically and spatially specific closures is consistent in Bhaskar's work (Bhaskar 1978, 1979: 128, 1986: 27). Moreover, this possibility is recognized by Lawson in one of his most important contributions. With regard to econometrics, Lawson says it is legitimate to investigate whether, "in certain conditions some closed-systems methods or whatever could contribute to enlightenment" (Lawson 1999d: 8) (emphasis added). The conditions in question suggest local closures.
In addition to these notions of local (but complete) closure, there is evidence that CR recognizes, to some extent, the possibility that systems can be partially open/closed, i.e. open/closed to a degree. For example, Bhaskar (1989: 185) claims that biology deals with quasi-closed systems. (21) An example given is the study of the life cycle of an organism; however, the reasoning is not explicit. This definition of partial closure suggests, therefore, that partially closed systems are ones in which the event regularity is apparent yet not strict. Such a partially closed system might have merely evolved; or the system has had closure introduced, making it more closed than before. One such source of greater closure might be an institutional feature, such as a rule, habit, custom, or convention. Indeed, Sayer (1992: 124) writes, "Many forms of social organization tend to approximate regularities in patterns of events by enforcing rules ...". This treatment captures the concept of partial closure in the first sense very well. Also, when Lawson (1993: 175) suggests that people create "a significant degree of structural stability", even under uncertainty, by resorting to conventions, he suggests closure which is partial: The convention represents a partial closure at the level of the agent (a structure), which reproduces rather than transforms other structures, and therefore creates more closure in the broader structural milieu. Again, it should be noted that these claims can only be made by going beneath the level of events.
What should be clear immediately is that this nuanced approach does not justify any simple strict dualistic treatment of the ontology of open systems. It would seem that any critical-realist treatments which hold that once perfect closure is impossible, distinctions between the different open systems available are lost, would seem to be invalid. Moreover, it would seem that Lawson in particular has understood this in another major contribution to CR, the concept of demi-regularities, which are "a special situation of the open world [in which] certain mechanisms (whether natural or social) reveal themselves in rough and ready patterns ... [but] it is a special case of this special situation that the patterns produced correspond to strict event regularities ..." (Lawson 1997: 219) (emphasis in original). (22) More simply put, demi-regularities are rough and ready patterns of events, which fall short of being strict regularities. The notion of demi-regularities captures both the concepts of local and partial closure. Clearly demi-regularities are the result of a form of partial closure.
Therefore, demi-regularities lie somewhere in between the closed system of the experiment and the chaos of a perfectly open system. They also suggest that it is useful to think of systems as lying at some point on a continuum, which suggests, in turn, that there are degrees of openness of systems. Thus, once one moves from the perfection of the experimental closure, there are a vast number of slightly different points at which one can stop. The concept of demi-regularities fits into this schema, and would appear to rebut any criticism that the Cambridge view is dualistic with respect to systems. However, this view of a continuum of degrees of openness leads to another criticism of the Cambridge approach. For, if the perfect closure is viewed as impossible, then the difference between possible closures and demi-regularities seems to be much smaller. Moreover, Cottrell (1998: 352) argues that Lawson overstates the difference between what he views as orthodox (complete, strict) event regularities, which must be based in perfectly closed systems, and his own demi-regularities. For Cottrell, orthodox economists do not seek the strict regularities Lawson claims they do. In terms of the present analysis, Cottrell is accusing Lawson of using a false impossible standard in order to criticize the orthodoxy and to establish his own concept of demi-regularity. Cottrell's argument does seem to be valid: There appears to be a difference in degree rather than in kind between the two types of regularity.
Overall, ontologically, CR presents a somewhat mixed picture: On one hand, CR is developing notions of openness which transcend the strict dual of open/closed system, which had been prevalent; yet on the other, there does seem still to be evidence that the notion of perfect closure, though impossible in reality, is used as a benchmark for the assessment of reality, and as a means of criticizing various economic theories. This tendency to overstate the possibility of closed systems, and to understate the similarity of CR's view of closed systems and the views of those whom they criticize, is shown even more clearly when examining methodology.
The most powerful contribution of the critical-Realist project in economics has been to demonstrate the importance of ontology and to reorient economics such that the clear disjuncture (particularly found in orthodox economics) which exists between reality and the methods employed to investigate it should be at least reduced. Consequently, in attempting to understand open environments, methodologies and methods that presuppose closure should not be relied upon. The aspiration for the economist should then be to adopt an open-systems methodology: one that accepts and takes seriously an ontology of open systems and does not merely ignore disjunctures between ontology and methods used. This methodology seems a reasonable aspiration for CR, given its ontological focus. However, it is argued that the Cambridge school of CR in economics has tended to adopt a strategy of rejection of what shall be called here "closed-systems" methods, i.e. techniques which presuppose closure, be that defined in terms of event regularities or beneath the level of events. (23) In contrast, this paper argues that a central tenet of an "open-systems methodology" is that it can still employ "closed-system methods", because the former will take seriously into account the weaknesses of the latter in open environments and employ them more cautiously and limitedly. (24)
There are several arguments against this rejection of closed-systems methods (see Mearman 2004) but one in particular is implied by an ontology of partial closure. If systems are open to differing degrees, then it is likely that methods are too; and therefore the key is to fit methods to situations. Moreover, just as it is unjustified to conflate systems that are not completely closed into a single category of "open-systems", it is unjustified to conflate methods that presuppose some degree of closure into a category of "closed-systems methods" and to reject these. The components of this argument will now be examined.
The first component of the argument is that the Cambridge view entails a strategy of rejection. Two examples are employed to make this case. Pratten (1996) provides one. He argues that ultimately, neo-Ricardian economists are faced with the choice of abandoning many of their methods or face being rejected by post-Keynesians (p. 439). His argument is that although they wish to do realist, open-systems work, neo-Ricardians are trapped into thinking in terms of closure (pp. 435, 437). That is, neo-Ricardians are wedded to both a closed-systems methodology and to closed-systems methods. These conclusions are based on his argument that neo-Ricardians engage in closure as a "first step" in analysis (pp. 43-432). This claim follows from such features of neo-Ricardian analysis as the use of "given" values of certain factors. Pratten is correct to argue that there are certain assumptions in the neo-Ricardian analysis that are questionable from the perspective of open systems. The assumption of a pre-determined long period would be one example.
Lawson's (1997) treatment of econometrics provides the other example of rejection. Lawson argues that regression analysis is based on the unwarranted assumptions of underlying homogeneity and of being able to exclude factors not selected for the analysis. This analysis leads Lawson to argue that a) econometrics should be restricted to conditions under which there is complete closure; b) econometrics should be redefined in terms of descriptive statistics; and c) other methods should be employed which are not guilty of closure. One such example is his notion of "contrast explanation" (see Lawson 1997, 2003).
The objections to both arguments are the same. In both cases, it seems to be argued that because the techniques and theories in question seem to impose or assume some closure, but the reality is open, the techniques are automatically invalid. However, that argument is problematic for several reasons. First, these arguments are effectively collapsing together techniques and effectively ignoring differences between them. Neo-Ricardian analysis does indeed involve some imposed closure; however, this is significantly less than in neo-classical analysis. There is indeed considerable realism in neo-Ricardian analysis. Moreover, Trigg (2003) argues that in neo-Ricardian analysis, methods are employed so as to suit circumstances and that this is evidence of an awareness of openness. Thus, drawing on Trigg, contrary to Pratten, even though neo-Ricardians do employ closed-systems methods, they do not engage in closed-systems methodology, because they do take into account the open nature of reality. One charge against Pratten here is that he ignores the way in which the closures, such as they are, have been introduced. In short he ignores what Mearman (2002a) has termed "process openness", under which the model or technique should not be judged merely on its form but on the method of its creation. "Process openness" implies that a model can appear to be closed, but has been constructed with openness in mind. Setterfield's (2003) formal model of demand is one such example: it contains equations, but ones which are the consequence of an open-systems thought process; many of his equations are constructed to reflect contingencies, have evolution built into them, and have "time-variant moments ... the evolution of [which] ... is not described within the confines of the model" (2003: 79). Clearly, "process openness" is an essential characteristic of an "open-systems methodology".
With respect to Lawson, clearly, there are many different types of econometrics, yet they are conflated and rejected except in highly specific circumstances. However, as Downward et al. (2002), Downward and Mearman (2002) and Mearman (2004)--responses derived from CR but reacting against the strategy of rejection entailed by the Cambridge view--have argued, there are elements of closure in all methods, including contrast explanation, such as the introduction of closure (ICC) necessary to envisage an entity as relatively enduring (Downward 1999), or the assumption of qualitative invariance involved in quantification. Furthermore, to analyse open systems, strategies must be developed, which inevitably amount to partially closing, in either sense, in thought, an open system. Dow (1996: 14) claims "an open system can be segmented into sub-systems which can be approximated to closed systems for partial analysis, but which are always open organically to influences from other parts of the overall system". Setterfield (2003) adopts the same tack, citing Kregel's (1976) stratagem of "locking up elements without ignoring them". Setterfield describes this as a "conditional closure". An obvious example of these techniques is that a critical-realist abstraction necessarily involves a focus on what is real and essential (Lawson 1989a) to the temporary exclusion of other factors, regarded as transient or insignificant. However, these influences cannot be legitimately completely excluded, as they might interfere at any point with the operation of the real and essential (Mearman 2001b). Moreover, by abstracting from other factors, one assumes that those factors are behaving in particular, consistent ways. To focus on the single player on a football field is not to ignore the fact that other players are also moving, but it is to make assumptions about those movements.
The present argument shows that although there might well be techniques, which can claim to impose relatively less closure, they are not free from closure. Thus, Lawson cannot claim to have open-systems techniques and Pratten cannot claim to avoid any first steps of closure in substantive analysis. Moreover, given that there are degrees of openness of reality, there are circumstances under which some more "open" techniques are less suitable than more "closed" variants. By treating all types of econometrics as the same, this ignores the fact that non- or semi-parametric techniques, for example, involve less closure than parametric techniques (Finch and McMaster 2002). Thus, as Mearman (2004) argues, a strategy of a priori rejection on the basis of openness is unsustainable. Nonetheless, such a strategy of rejection seems to have been argued by the Cambridge view.
This paper has examined the influential treatment of "open systems" by critical realism, which has been prominent in this journal. This paper has argued that a "Cambridge view" of CR in economics is identifiable and distinctive. It is argued that this Cambridge view has three problems in its treatment of open systems: 1) it is dominated by event-level definitions--which also reflects an underdeveloped concept of "system"; 2) it emphasizes negative definitions; and 3) it tends towards polarizing definitions, particularly in methodology. These problems create difficulties in trying to develop methodology and substantive work informed by critical realism. The event-level definition is certainly effective for criticizing orthodox economics and for focusing discussion onto ontology. The event-level definition might also act as a rough guide for identifying possible closed systems. However, it hinders analysis by ignoring the nature of the system. It is also shown to be an imperfect guide to openness. It is certainly ironic that CR is to be criticized for focusing on the event level: One of its main achievements has been to focus attention away from the event level and to the deeper levels of causal structures and mechanisms. However, there seems to be a contradiction in the critical-realist treatment: The concern with ontological depth has been accompanied by a relative ignorance of that depth. Perhaps this has occurred because of the central concern to criticize the orthodoxy. Alternatively, possibly it occurs because CR in economics lacks a coherent notion of system, and therefore does not think of the system as a whole.
Possibly the greatest problem with the critical-Realist treatment is its polarizing treatments of existing methods. This mode of argumentation is clearly intended to criticize the orthodoxy. However, it has been shown a number of times that this argument is unsustainable. Furthermore, given that CR has been shown to have recognized the inappropriateness of dualism ontologically, it is problematic for a realist perspective that, according to the evidence in this paper, it maintains a dualist stance on methodology.
What is needed for the project informed by critical realism is the construction of a positive, nuanced treatment of systems. These treatments will, in turn, involve a more complete definition of system, which moves beyond simply classifying systems by their event patterns. Such a definition will likely incorporate treatments of systems from other literature (Mearman 2002a,b).
Archer, M. (1995) Realist Social Theory: The Morphogenetic Approach, Cambridge: Cambridge University Press.
Archer, M. (1998) "Introduction: Realism in the Social Sciences," in M. Archer, R. Bhaskar, A. Collier, T. Lawson and A. Norrie (eds) Critical Realism: Essential Readings, London: Routledge.
Beed, C. and Beed, C. (1996) "Polarities Between Naturalism and Non-naturalism in Contemporary Sciences," Journal of Economic Issues 30(4): 1077-1104.
Bertalanffy, L. von (1950 ) General System Theory: Foundation, Development, Applications, New York: George Braziller.
Bhaskar, R. (1978) A Realist Theory of Science, Brighton: Harvester.
Bhaskar, R. (1979) The Possibility of Naturalism, Hemel Hempstead: Harvester.
Bhaskar, R. (1986) Scientific Realism and Human Emancipation, London: Verso.
Bhaskar, R. (1989) Reclaiming Reality, London: Verso.
Bhaskar, R. (1993) Dialectic: The Pulse of Freedom, London: Verso.
Brown, A., Slater, G. and Spencer, D. (2003) "Driven to Abstraction? Critical Realism and the Search for the 'Inner Connection' of Social Phenomena," Cambridge Journal of Economics 26(6): 773-788.
Caldwell, B. (1982) Beyond Positivism: Economic Methodology in the Twentieth Century, London: Allen and Unwin.
Collier, A. (1994) Critical Realism: An Introduction to Roy Bhaskar's Philosophy, London: Verso.
Cottrell, A. (1998) "Realism, Regularities and Prediction," Review of Social Economy 56(3)(Fa11): 347-355.
DeUriarte, B. (1989-1990) "On the Free Will of Rational Agents in Neo-classical Economics," Journal of Post-Keynesian Economics 12: 605-617.
Dow, S. C. (1990) "Beyond Dualism," Cambridge Journal of Economics 14: 143-157.
Dow, S. C. (1996) The Methodology of Macroeconomic Thought, Aldershot: Elgar.
Dow, S. C. (1999) "Post-Keynesianism and Critical Realism: What is the Connection?" Journal of Post-Keynesian Economics 22: 15-33.
Dow, S. C. (2000) "Prospects for the Progress of Heterodox Economics," Journal of the History of Economic Thought 22(2): 157-170.
Dow, S. C. (2004) "Reorienting Economics: Some Epistemological Issues. A Review Essay on Tony Lawson, 'Reorienting Economics'," Journal of Economic Methodology 11(3): 307-312.
Downward, P. (1999) Pricing Theory in Post-Keynesian Economics: A Realist Approach, London: Edward Elgar.
Downward, P., Finch, J. and Ramsay, J. (2002) "Critical Realism, Empirical Methods and Inference: A Critical Discussion," Cambridge Journal of Economics 26(4): 481-500.
Downward, P. and Mearman, A. (2002) "Critical Realism and Econometrics: A Constructive Dialogue with Post-Keynesian Economics," Metroeconomica 53(4): 391-415.
Downward, P. and Mearman, A. (2003) "Presenting 'Demi-Regularities' of Pricing Behavior: the Need for Triangulation," in M. Forstater and L. R. Wray (eds) Contemporary Post-Keynesian Analysis: A Compendium of Contributions to the Seventh International post Keynesian Workshop, New York: Elgar.
Downward, P. and Mearman, A. (2004) "On Tourism and Hospitality Management Research: A Critical Realist Proposal," Tourism and Hospitality Management Planning and Development 1(2): 107-122.
Dunn, S. (2001) "Whither Post-Keynesianism?" Journal of Post-Keynesian Economics 22 (3): 343-364.
Faulkner, P. (2002) "Some Problems with the Conception of the Human Subject in Critical Realism," Cambridge Journal of Economics 26: 739-751.
Finch, J. and McMaster, R. (2002) "On Categorical Variables and Non-parametric statistical Inference in the Pursuit of Causal Explanations," Cambridge Journal of Economics 26: 753-772.
Fleetwood, S. (ed.) (1999) Critical Realism in Economics: Development and Debate, London: Routledge.
Fleetwood, S. (2002) "Why Neo-classical Economics Explains Nothing at All," Post-Autistic Economic Review 17.
Grunberg, E. (1978) '"Complexity' and 'Open Systems' in Economic Discourse," Journal of Economic Issues 12(3)(September): 541-560.
Hahn, F. (1989) "Kaldor on Growth," in T. Lawson, J. Palma and J. Sender (eds) Kaldor's Political Economy, Dorchester: Academic Press.
Hendry, D. (1995) Dynamic Econometrics, Oxford: Oxford University Press. Hodgson, G. (1988) Economics and Institutions: A Manifesto for a Modern Institutional Economics, Cambridge: Polity.
Hodgson, G. (2000) "What is the Essence of Institutionalist Economics?" paper presented to the A fEE Sessions at ASSA, Boston, MA, January.
Ingham, G. (1996) "Money as a Social Relation," Review of Social Economy 54(4): 507-530.
Kaldor, N. (1972) "The Irrelevance of Equilibrium Economics," Economic Journal 82 (December).
Kapp, W. (1968) "In Defense of Institutionalist Economics," Swedish Journal of Economics 70: 1-18.
Keynes, J. M. (1973) Collected Writings of John Maynard Keynes, London: Macmillan for the Royal Economic Society (Vol. VII).
Kregel, J. (1976) "Economic Methodology in the Face of Uncertainty," Economic Journal 86: 209-225.
Lawson, C. (1996) "Realism, Theory and Individualism in the Work of Carl Menger," Review of Social Economy 54(4): 445-464.
Lawson, T. (1989a) "Abstraction, Tendencies and Stylised Facts: A Realist Approach to Economic Analysis," Cambridge Journal of Economics 13(1): 59-78.
Lawson, T. (1989b) "Realism and Instrumentalism in the Development of Econometrics," Oxford Economic Papers 41: 236-258.
Lawson, T. (1993) "Keynes and Conventions," Review of Social Economy 51(2)(Summer): 174-200.
Lawson, T. (1994a) "The Nature of Post-Keynesianism and its Links to Other Traditions: A Realist Perspective," Journal of Post-Keynesian Economics 16(4): 503-538.
Lawson, T. (1994b) "A Realist Theory for Economics," in R. Backhouse (ed.) New Directions in Economic Methodology, London: Routledge.
Lawson, T. (1995a) "A Realist Perspective on Contemporary 'Economic Theory'," Journal of Economic Issues 29(1)(March): 1-32.
Lawson, T. (1995b) "The 'Lucas Critique': A Generalisation," Cambridge Journal of Economics 19: 257-276.
Lawson, T. (1996) "Development in Hayek's Social Theorising," in S. Frowen (ed.) Hayek, the Economist and Social Philosopher: A Critical Retrospect, London: Macmillan.
Lawson, T. (1997) Economics and Reality, London: Routledge.
Lawson, T. (1998a) "Clarifying and Developing the Economics and Reality. Project: Closed and Open Systems, Deductivism, Prediction, and Teaching," Review of Social Economy 56(3)(Fall): 356-375.
Lawson, T. (1998b) "Situated Rationality," Journal of Economic Methodology 4(1): 101-125.
Lawson, T. (1999a) "Critical Issues in Economics as Realist Social Theory," in S. Fleetwood (ed.) Critical Realism in Economics: Development and Debate, London: Routledge: 209-258.
Lawson, T. (1999b) "What Has Realism Got to Do With It?" Economics and Philosophy 15: 269-282.
Lawson, T. (1999c) "Developments in Economics as Realist Social Theory," in S. Fleetwood (ed.) Critical Realism in Economics: Development and Debate, London: Routledge: 3-20.
Lawson, T. (1999d) "Connections and Distinctions: Post-Keynesianism and Critical Realism," Journal of Post-Keynesian Economics 22(1)(Fall): 3-14. Lawson, T. (2003) Reorienting Economics, London: Routledge.
Lee, F. (2002) "Theory Creation and the Methodological Foundation of Post-Keynesian Microeconomics," Cambridge Journal of Economics 26: 789-804.
Lewis, P. (1996) "Metaphor and Critical Realism," Review of Social Economy 54(4): 487-506.
Lewis, P. and Runde, J. (1999) "A Critical Realist Perspective on Paul Davidson's Methodological Writing on and Rhetorical Strategy for--Post-Keynesian Economics," Journal of Post-Keynesian Economics 22(1)(Fall): 35-56.
Mearman, A. (2001a) "On the Coherence of Post-Keynesian Economics: Is critical Realism Dow-dualist?" Paper presented to the conference of the International Association for Critical Realism, Roskilde, August.
Mearman, A. (2001b) "Review of P. Downward (1999)," Review of Political Economy 13(3): 383-387.
Mearman, A. (2002a) "To What Extent is Veblen and Open Systems Theorist?" Journal of Economic Issues 36(2): 573-580.
Mearman, A. (2002b) A Contribution to the Methodology of Post-Keynesian Economics, Ph.D. thesis, University of Leeds.
Mearman, A. (2003) "Review of Sutton "Marshall's Tendencies: What Can Economists Know?'," mimeo.
Mearman, A. (2004) "Open Systems and Economic Methodology," University of the West of England Discussion Papers in Economics, 04/01.
Mearman, A. (2005) "On Sheila Dow's Concept of Dualism: Clarification and Development," Cambridge Journal of Economics 29(4): 619-634.
Monastersky, N. (2002) "Recycling the Universe: New Theory Posits that Time has no Beginning or End," Chronicle of Higher Education, 7 June.
Morgan, M. (2002) "How Models Help Economists to Know," Economics and Philosophy 18(1): 5-16.
Olsen, W. (2000) "An Open Systems Interpretation of Simultaneous Equation Models: Realist Theory and an Example Using Farm-Level Data," Paper presented to the IACR Conference, Lancaster, August.
Pinkstone, B. (2000) "Econometrics versus Contrastives: Alternative Approaches to Explanation in Economic History: A Case Study," Paper presented to the Cambridge Realist Workshop Conference, Critical Realism, What Difference Does It Make?, Cambridge University, May.
Pratten, S. (1996) "The 'Closure' Assumption as a First Step: Neo-Ricardian Economics and Post-Keynesianism," Review of Social Economy 54: 423-443.
Pratten, S. (2005) "Economics as Progress: The LSE Approach to Econometric Modelling and critical realism as Programmes for Research," Cambridge Journal of Economics 29(2): 179-205.
Rotheim, R. (1998) "On Closed Systems and the Language of Economic Discourse," Review of Social Economy 56(3)(Fall): 324-334.
Rotheim, R. (1999) "Post-Keynesian Economics and Realist Philosophy," Journal of Post-Keynesian Economics 22(1)(Fall): 71-104.
Runde, J. (1996) "On Popper, Probabilities and Propensities," Review of Social Economy 54(4): 465-486.
Sayer, A. (1981) "Abstraction," Radical Philosophy (Summer): 6-15, reprinted in M. Archer, et al. (eds): 120-143.
Sayer, A. (1992) Method and Social Science: A Realist Approach, London: Routledge.
Setterfield, M. (2003) "Critical Realism and Formal Modelling: Incompatible Bedfellows?" in P. Downward (ed.) Applied Economics and the Critical Realist Critique, London: Routledge.
Siakantaris, N. (2000) "Experimental Economics Under the Microscope," Cambridge Journal of Economics 24:267-281.
Sutton, J. (2000) Marshall's Tendencies: What Can Economists Know? Cambridge, MA.: Cambridge University Press.
Trigg, A. (2003) "Quantity and Price Systems: Towards a Framework for Coherence Between Post-Keynesian and Sraffian Economics," Paper presented to the conference of the Association for Heterodox Economics, Nottingham, July.
Viskovatoff, A. (1998) "Is Gerard Debreu A Deductivist? Commentary on Tony Lawson's Economics and Reality," Review of Social Economy 56(3): 335-346.
Walters, B. and Young, D. (1997) "On the Coherence of Post-Keynesian Economics," Scottish Journal of Political Economy 44(3): 329-349.
School of Economics, University of West England, UK
(1) Other significant members of this category are Clive Lawson (1996), Pratten (1996), Runde (1996), Lewis (1996), Ingham (1996), Siakantaris (2000), Faulkner (2002) and Fleetwood (1999, 2002), all of whom teach, taught or studied at Cambridge. Pinkstone (2000) and Rotheim (1998, 1999) were temporarily based at Cambridge in the 1990s and 1980s respectively. Clearly, there is a geo-historical justification for the identification of this collection of scholars as a group. Additionally, there is sufficient coherence in terms of their influences and approach on issues of economic practice to warrant their being grouped together. It is beyond the scope of this paper to prove this point fully. However, it should be stressed that the coherence of approach on key issues is more significant in identifying the Cambridge group than is some common geography: For example, Setterfield (2003) also studied at Cambridge, but has arguably moved in a different theoretical direction.
(2) Bhaskar (1978:51-52) defines a (generative) mechanism as "nothing other than a way of acting of a thing. It endures, and under appropriate circumstances is exercised, as long as the properties that account for it persist."
(3) Here, "experimental" refers to a situation in which a scientist intervenes to isolate and control a specific mechanism. Thought experiments are not included, which is slightly unsatisfactory because thought experiments are also important to science; and because thought experiments might also involve the isolation and control of mechanisms, albeit in thought.
(4) Neither Lewis and Runde nor Lawson define clearly what is meant by a well-behaved probability distribution; however, both seem to mean a distribution that is symmetrical; for example, the normal distribution. This interpretation is consistent with Lawson's (1989b) treatment of the ECC in econometrics.
(5) Lawson (2003: 15) does develop his concept of closed systems. He distinguishes between "closure as concomitance" and "closure as causal sequence." However, the closure is still identified via the event level.
(6) In fact, Lawson (2003) spends little time defining open systems. However, he does also define them in terms of the presence of multiple causes or conditions (2003: 42, 56, 125, 229, 233) or where reality is highly internally related (p. 229).
(7) Lawson (1997) is somewhat ambivalent on this point: frequently he strictly separates closures and the conditions necessary for them (ICC and ECC), which matches Pratten's position; however, he (pp. 76-78) also discusses closure as being achieved by the ICC and/or ECC, for example in "isolated systems" (p. 78) or "closed economies" (p. 78).
(8) Viskovatoff (1998) questions Lawson's characterization, arguing that there are (at least) two methodologies in orthodox economics: one, he claims, pays no attention to empirical outcomes. Lawson (1998a) disagreed. Mearman (2003, 2004) argues that the language of some orthodox economists, such as Sutton (2000), Morgan (2002) and others, implies a concern for mechanisms.
(9) For example, whilst Austrian and Marxist economists might agree with the need to go below the surface of events, they would disagree strongly about what lies beneath. An event-level definition avoids such disagreement, albeit temporarily.
(10) As an illustration, two pertinent definitions to consider are Kapp's (1968) and Dow's (1996). Kapp's definition corresponds exactly with an aspect of the GST definition in one way: that an open system is one that receives (and survives on) impulses from outside. It is difficult to identify this aspect in the critical-realist definition since in it, a) "system" is not well defined, and b) the spatial aspect is de-emphasized (Bhaskar 1978: 76-77). Another aspect of Kapp's definition is that in open systems there is an interaction between sub-systems, which is akin to the common notion that larger systems (often considered open) do comprise sub-systems. For the same reasons as stated immediately above, it is difficult to conceptualize this notion in CR, for, although Kapp's concept of the boundary might be thought of as similar to the ECC, clearly a) there is no concept of boundary in CR; and b) there is no notion in CR of the system requiring external impulses for its survival. For the same reasons, Dow's (1996) definition of an open system, which stresses fuzzy or indeterminate boundaries, is a departure from the CR definition. Dow also notes that an open system is identifiable by imperfect ordering, i.e. with a degree of disorder (Dow 1996: 14). This is opposite to the GST concept of closed systems being associated with randomness (in the usual sense). Disorder is perhaps only present in CR in that a non-invariant empirical relationship might imply disorder.
(11) This argument is complementary to Brown et al's (2003) argument that critical-realist abstraction is weak in terms of reconstituting the concrete.
(12) Lawson (2003: 15-16) also downplays the significance of the term "'system," claiming that the definition of the system is merely dependent on the area or time over which an event regularity can be found.
(13) Above, it was argued that such diversity of views might be regarded positively, for example as a sign of flexibility. However, in the case of the definition of the system, there is simply a lack of coherent treatments. Other economists drawing on CR, such as Downward (1999), Lee (2002), Brown et al. (2003) and Mearman (2002b) are trying to address this apparent lacuna.
(14) Lawson (1989a: 71) does offer a definition of a "system", in terms of a combination of structures. Second, he defines a totality and differentiates between it and a system: arguably in terms of the extent of internal relationality between structures. Furthermore, the dialectical turn of CR has produced the concept of "totality," which is a system of internal relations (Bhaskar 1993: 405), with an intensive and extensive margin (p. 125). Clearly this captures better the notion of "system". Bhaskar (1993: 95-96, 127, 269, 273; 357) distinguishes between open and closed totalities but these are distinct from the concept of open (and closed systems) used, which remains: "Systems where constant conjunctions of events do not occur" (p. 401).
(15) I acknowledge the suggestion of this point by an anonymous referee.
(16) Others, such as Dow (1996) and Setterfield (2003), offer a similar definition.
(17) In terms of economic applications, Mearman (2002a) has argued that Veblen's work exemplifies an open-systems approach. Veblen does not discuss systems, or use that terminology; however, there is clear evidence of his thinking in terms of multiple mechanisms, ontological depth and complex patterns of events. Mearman argues that Veblen's approach is consistent with the OSO outlined in the first section.
(18) One of the principal tasks of heterodox economists and organizations is to develop coherence and/or develop arguments against the necessity of coherence. This remains, in spite of the clear fragmentation of the orthodoxy (cf. Viskovatoff 1998; Caldwell 1982).
(19) Mearman (2005) shows that technically, polar concepts strictly cannot be duals under Dow's definition. However, polar concepts can be rendered duals by imposing an artificial dividing point between the poles.
(20) Clearly this assumes a broader definition of work than, for example "leave the home to work on another's property". This is contentious but consistent with the critical-realist treatment of "work" (Bhaskar 1978: 194-195).
(21) Bhaskar (1978: 253, n. 1) makes the same point: "... it is clear that some systems, such as biological ones, are more nearly closed (reveal a greater degree of regularity of behaviour, or recurrence of syndromes) than others ..."
(22) Lawson (2003: 105-106) clarifies his concept of demi-regularity. It can apply to any rough and ready pattern, even when the pattern involves a deviation from an expected regularity.
(23) It should be noted immediately that closed-systems methods are not a synonym for what Lawson called the "mathematical-deductive" approach. Although both mathematics and deduction are "closed-systems methods" the term is not restricted to them. Moreover, whereas in Lawson, deductivism implies nomological arguments; and further, that such laws are formulated in terms of universal event regularities; "closed-systems methods" do not necessarily presuppose event regularities.
(24) See Mearman (2004) and Downward and Mearman (2002, 2003, 2004), who argue that a strategy of triangulation of methods--which employ varying degrees of closure--might be preferable in open systems.…