An International Symposium celebrating the centenary of Sir Karl Popper's birth in 1902 occasioned this edited volume. The contribution of Popper to economic methodology has been extensively debated but nowadays represents a marginal area in the philosophy and methodology of economics. It is thus legitimate to wonder whether such a book was at all needed. Or as the editors put it, ‘what is the ‘value-added’ of another addition to the literature?’ (p. 1). The title would suggest that the volume's objective is to re-examine Popper's contribution to economic methodology in light of contemporary challenges. This sounds like a worthy enterprise. There might be something to learn from re-engaging an influential and prolific philosopher like Popper in light of new developments in philosophy and economics. Yet contemporary developments in the science of economics (think for instance of the increasing popularity of behavioural and experimental economics), which certainly pose significant challenges to economic methodologists, are seldom mentioned. And work in the philosophy and methodology of economics of the last couple of decades is rarely discussed in detail. These absences give the volume a somewhat outdated flavour. More importantly, I think that seriously engaging contemporary developments in economics and in economic methodology would have significantly benefited some of the essays. (A few illustrations are given below.)
The suspicion arises that in spite of the book's title and its publication under the aegis of the International Network for Economic Method, the volume's contributions might not be in the field of economic methodology. In the Preface, the authors indicate that the aim of the volume is ‘to reconsider fundamental aspects of Popper's work in the light of major developments in twentieth-century philosophy’ (p. 1), namely the realism/antirealism debate and the induction/anti-induction debate. So formulated its aim bears only a loose connection to either economics or economic methodology. Indeed I think that scholars interested in critically examining Popper's thought will find this volume more relevant than general philosophers and methodologists of economics will do.
In the first chapter, the editors Thomas Boylan and Paschal O'Gorman make a case for the importance of the realism/antirealism debate both in general philosophy as well as in philosophy of science for examining the contribution of Popper to economic methodology. In so doing, the essay also sets the stage for the subsequent two chapters. In particular, the authors put forward four claims meant to demonstrate that the realism/antirealism debate is important for appreciating Popper's views on the methodology of economics. First, although early in his career Popper regarded the realism/antirealism debate as irrelevant for scientific methodology, later on he came to maintain the importance of an objective theory of truth to the methodology of any science, including economics. Second, the authors argue that Popper's realist view of logic has implications for economic methodology as well as for economics. The reason is that realism in logic is essential to the mathematical modelling of the economy in general equilibrium theory. Third, Popper's realism is also important for a later development in Popper's philosophy, namely the ‘thesis of the three worlds’. This postulates the existence of three worlds or universes: (i) world one: the world of physical objects or of physical states; (ii) world two: the world of states of consciousness or mental states; (iii) world three: the world of objective contents of thought (which include theoretical systems, critical arguments, the contents of journals and books). The authors argue that the three world thesis provides the indispensable framework in which to locate economics (and the social sciences): economics belongs to the third world, its development follows the Popperian quadruple of problem – tentative solution – error elimination – problem, and is constituted of theories that we know cannot be true. Neoclassical theory is a good example: it is objective in Popper's sense, and yet its role is that of a regulative principle – an ideal from which we fall short in actual decision making.
Boylan and O'Gorman's fourth claim is that for Popper models play an indispensable role in social scientific explanations. I'll examine this claim in more detail, for it is one of the areas where I think that paying more attention to contemporary debates in economic methodology would have made the contribution more relevant and topical. Popper's view on the role of scientific models is tied to the explanation of kinds or types of events. If we want to explain why lunar eclipses occur when there is a full moon, we build a mechanical model (e.g. with balls and lamps) of the sun, earth, and moon system. In contrast to an explanation of singular events that conforms to the D-N model of explanation, here there is no explicit reference to initial conditions. It is the model itself that represents typical initial conditions. To function in an explanatory fashion, however, the model needs to be ‘animated’ by laws (in the previous example these are Newton's laws). In the social sciences models are more important than in the natural sciences because explanation in the social science proceeds in terms of human actions and social situations and it is often impossible to provide a complete description of a specific social situation. This means that in the social sciences models play a crucial role both in explanations of singular events and in explanations of kinds of events. The model describes a ‘typical agent’, who is anybody who shares the relevant situation, and then specifies the agent's relevant aims and typical knowledge, the physical things and their properties that are relevant to the situation and the relevant social institutions. The explanation of the event then comprises the model and the animating law, which in social science is the rationality principle: ‘the assumption that the various persons or agents involved act adequately or appropriately, that is to say, in accordance with the situation’ (Popper Reference Popper1994: 169).
Boylan and O'Gorman suggest that Popper's view of models might offer new insights to recent debates on modelling in the philosophy of science and the philosophy of economics. But how, precisely? Unfortunately Boylan and O'Gorman leave it to us to figure this out. Speculatively one possibility would be to consult Popper's account on the issue of the explanatory function of models. Two features of Popper's account, however, lead me to doubt the extent to which it may be relevant to contemporary concerns. First, Popper's account is geared to the D-N model of explanation, which nowadays virtually everyone regards as an inadequate account of explanation. In its stead, gaining increasing currency is the idea that the explanatory function of models is tightly connected to their manipulability (see for example Woodward Reference Woodward2004), a feature that is absent from Popper's account. Second, it is unclear from Boylan and O'Gorman's discussion whether and how Popper's account addresses the crucial question of how false models explain. In other words, how idealized models of typical social situations, animated by the rationality principle, may genuinely explain real-world phenomena. In chapter three, O'Gorman suggests that the models animated by the rationality principle serve to estimate the deviation of actual behaviour from ideal behaviour (p. 70). If this interpretation is correct, there appears to be a tension between Popper's account, where models function as ideals against which to contrast actual behaviour, and the contemporary view that idealized models serve to isolate theoretically the operation of an actual causal factor or a mechanism of interest (see for instance Cartwright Reference Cartwright1999, Mäki Reference Mäki2004). I will not try to resolve this tension here, but I believe this is a topic that may have deserved further attention. In general, since theoretical modelling and their role in economics occupy contemporary philosophy and methodology of economics a great deal more than Popperian themes, this feels a lot like a missed opportunity.
In the second essay, O'Gorman examines four issues. First, he deals with the apparent contradiction between Popper's falsificationism and his situational analysis as the appropriate methodology for the social sciences. O'Gorman claims that they can be reconciled. Situational models are in fact testable. In particular, different models of the typical situation can be tested for the accuracy of the description of the relevant physical factors, of the knowledge base attributed to the representative agent and of the relevant institutional factors. What is not testable is the rationality principle. But this is not a problem for Popper. In testing competing explanations (constituted by the rationality principle and different models of the situation) it is sound methodological policy to regard the different models of the situation accountable, and not the common rationality principle. The second issue concerns the consequences of Popper's three world thesis for his ideas on piecemeal engineering. O'Gorman convincingly demonstrates that the three world thesis supports Popper's subscription to piecemeal social engineering. The third issue explores the consequences of the three world thesis for Popper's methodological individualism, and claims that it requires its abandonment. Finally, the fourth issue examines the relationship between hermeneutical approaches and Popper's methodology of the social sciences: O'Gorman identifies a complementarity between the two.
Let me dwell on O'Gorman's claim regarding the consequences of the three world thesis for Popper's methodological individualism, for I take it to be another instance in which engaging recent work in philosophy and methodology of economics would have been helpful. O'Gorman distinguishes between a qualified and an unqualified version of methodological individualism. The unqualified version states that all statements about the social can be reduced to statements about the individual, whereas the qualified version claims that where possible reference to social structures and institutions should be eliminated (p. 43). Although the early Popper seems to wave between the qualified and the unqualified version, O'Gorman claims that the Popper of the three world thesis is bound to subscribe to the qualified one. The reason is that social structures and institutions belong to world three, which is real and autonomous from both world one and world two. For O'Gorman this ontological view ‘unequivocally’ implies qualified methodological individualism, which admits of the possibility that certain social structures cannot be eliminated and ‘remain on the scientific books’ (p. 43). The conclusion is that Popper's situational analysis ‘does not reject a judicious recourse to relevant institutions properly understood’ (p. 45).
O'Gorman's argument contains a few parts lacking in clarity that, I suspect, could have been avoided if more attention had been paid to contemporary literature on methodological individualism. O'Gorman claims that qualified methodological individualism differs from the unqualified formulation because the former allows for the possibility that certain social institutions might not be reduced. Which ones? Why? Is the reason ontological, that is, certain social entities have irreducible properties? Or is it epistemological, that is, we are not in a position to carry out the necessary reduction? Or is it pragmatic, that is, for certain purposes reference to social structures is just fine?
Contemporary philosophical literature on reductionism and individualism has taught us that theses about explanation and ontology should be kept distinct. There is no ‘unequivocal’ move from ontology to methodology. For instance, one can hold on to a strong ontological individualist thesis and still maintain that social-level explanations are legitimate, say because they provide information that is unavailable from individual-level explanations (Jackson and Pettit Reference Jackson and Pettit1992). If we take this lesson seriously, it is not in Popper's views about ontology but in Popper's views about explanation that we should look for hints about his methodological individualism. It is possible that for Popper reference to social structures and institutions remain in the scientific books because for certain explanatory purposes the reduction of social structures and institutions is unnecessary, and not because their mode of existence makes it impossible. Let me emphasize that I am not suggesting that O'Gorman is mistaken in claiming that the later Popper endorses qualified methodological individualism. My point is rather that O'Gorman's claim needs more careful argumentation and that consulting contemporary philosophical discussions on methodological individualism could have provided a wider range of resources to do so.
In the third chapter, Thomas Boylan aims to examine the reasons why Popper's concept of rationality remains privileged in orthodox economics. This is a place where I would have appreciated some discussion of the recent experimental and behavioural literature that has become so important in contemporary economics and economic methodology. Haven't these developments challenged the privileged position of the concept of rationality? I guess Boylan thinks they haven't. Even so, the choice of glossing over these developments altogether is questionable, especially in light of the behavioural economists’ recent pronouncement that their theories are within the mainstream (see for example Camerer and Loewenstein Reference Camerer, Loewenstein, Camerer, Loewenstein and Rabin2004). Boylan then moves on to look at how Popper's concept of rationality and his methodology of the social sciences would have been modified had Popper taken seriously the work of two of his contemporaries, namely Quine and Wittgenstein.
In the fourth chapter, Tony Lawson addresses the question that motivates critical realism and his particular contribution: how can social explanatory work proceed in an open system that lacks the possibility of experimental intervention? As is well known, critical realists subscribe to three theses: (i) the nature of the material we study bears significantly on how we can study it; (ii) social reality is both open and structured; (iii) causal explanation, as opposed to deductivism, is the proper method for the social sciences. The critical realist view openly contrasts with Popper's ideas on social scientific method, which prescribe the use of situational analysis and methodological individualism. Lawson is aware of this but proposes that the two views converge on the idea that scientists seek for errors and learn from them. Most of Lawson's essay is then devoted to exposing his ideas about social scientific explanation. The morale of the essay is that, if we agree with Lawson who agrees with Popper, that seeing a problem is important, and that a condition for seeing the problem is the experience of disappointed expectations, than we should promote a more inclusive academy than we currently have. The novelty of this essay vis-à-vis previous works by Lawson lies in the attempt to establish a connection between critical realism and Popper's critical rationalism. Yet, given the disproportionate space allocated to critical realist prescriptions, which although interesting are not new, and their obvious contrast with Popper's explicit methodology for the social science, one wonders whether Popperian themes are there mainly to justify the inclusion of the essay in this volume.
The chapter by Giorello and Motterlini takes a critical stance towards Popper's views on metaphysics and on the growth of knowledge through criticism. For Popper, metaphysics is an external source of scientific ideas. The authors argue against this view to claim that disagreement and debate concerning the metaphysical core and heuristic power of different research programmes is a prerequisite for scientific progress and that the framework for scientific testing is itself defined by the clash between competing research programmes. The argument is supported by nice examples from the history of science, none of which unfortunately concerns economics. Finally, the essays of John McCall and K. Vela Velupillai show how recent advances in probability theory and mathematics can vindicate the role of induction in science. Both essays however lack a specific connection to either economics or economic methodology.
To conclude, the volume contains some interesting ideas but the general impression is that for a collection whose title is ‘Popper and Economic Methodology. Contemporary Challenges’ there is too little both of contemporary economic methodology and of contemporary economics. On the assumption that scholars with a general interest in economics and economic methodology are among the intended audience of this volume, I believe that consulting and discussing contemporary work in these fields would have considerably increased the value-added of this (further) contribution on Popper and economic methodology.
