The economic problem of a community: ontological reflections inspired by the Socialist Calculation Debate

Abstract

Implicitly, the Socialist Calculation Debate is about the connection between neoclassical theory and reality. It involves several ontological presuppositions, with wide-ranging implications. Our purpose is to discuss one of these presuppositions. We want to scrutinise the conception, implicit in the ‘market socialist’ arguments, that things have economic roles or functions independently of the social environment in which they exist. Primary factors and techniques of production are taken to exist independently of social arrangements; jointly with consumer preferences, they are supposed to define the economic problem of a community. In criticising this conception, we borrow from F. A. Hayek, complementing his reflections with those of Donald Davidson, Tony Lawson and others. We also discuss a second problem, which follows directly from the first: if, in contrast with the market socialists’ conception, the economic roles of things depend holistically on complex social processes, how can we compare alternative forms of economic organisation?

1. Introduction

In 1920, Ludwig von Mises argued that socialism would be impossible. According to Mises, solving complex economic problems presupposes an ‘intellectual division of labour’ and a price system resulting from competitive exchanges against money. The removal of means of production from monetary exchange would imply, accordingly, that only simple, household economic problems are rationally solvable under socialism.

In the 1930s, several economists attempted to refute Mises’ impossibility argument in what came to be known as the Socialist Calculation Debate. They theorised the workings of a system of economic organisation said to be socialist insofar as means of production are not privately owned, but in which alternatives are evaluated by way of prices that result from the interaction of decentralised decision units.

Their counterargument assumes, however, that primary factors and techniques of production are explicitly or propositionally known, or at least knowable. F. A. Hayek’s most influential objection to these ‘market socialist’ proposals targets precisely this. As there exists neither a list of all primary factors nor a ‘book of blueprints’ with every technique of production, any form of economic organisation must discover what things take these economic roles. The competitive market process leads to reasonably efficient outcomes precisely because it integrates a vast amount of dispersed, often tacit knowledge and continually renews the structure of knowledge that makes a complex economy work.

The Socialist Calculation Debate is, of course, much more than a debate on the feasibility of ‘socialism’ or the merits of the competitive market process. Implicitly, the discussions turn on the nature of economic theory and, specifically, on the connection between neoclassical theory and reality. Though not all the participants seem quite aware of this, the debate involves a series of fundamental ontological presuppositions, with wide-ranging implications. Our purpose in this paper is to discuss one of these ontological presuppositions.

We want to scrutinise the conception, implicit in the market socialist arguments, that things have economic roles or functions—e.g. that that is a means of production—independently of the social environment in which they exist. In other words, for the market socialists, primary factors or techniques of production are not just taken as known, or knowable. They are also given in the ontological sense of existing independently of social arrangements. Jointly with consumer preferences, they are supposed to define the economic problem of a community.

In discussing this conception, and articulating an alternative ontological position, we borrow from Hayek’s writings in the aftermath of the debate, in which categories in the moral sciences are associated with action and hence with attitudes. But we complement Hayek’s reflections with those of Donald Davidson, Tony Lawson and other authors. We do not aim to offer either an account of Hayek’s intellectual evolution, as Fleetwood (1995) or Caldwell (2005) do, or a comparison of Hayekian and other perspectives on social ontology, as provided by Lawson (1997, pp. 134ff) or Runde (2001).

We also comment on a second problem, which follows directly our discussion of the first. In the alternative ontological position that we sketch, the economic roles or functions of things depend holistically on complex social processes. But, if this is so, how is it possible to compare alternative forms of economic organisation, or indeed to assess the effects of policy measures?

The paper is structured as follows. We begin by briefly reviewing Mises’ impossibility argument (Section 2) and the market socialists’ response to Mises (Section 3). We then elaborate on Hayek’s most essential criticism of the market socialist proposals (Section 4). This epistemic criticism will lead us to the implicit ontological commitment that we want to assess, i.e. that things have economic roles independently of the social environment in which they exist. In scrutinising this market socialist presupposition, we articulate an alternative position, built on the recognition of the role of attitudes in the definition of economic roles (Section 5). Finally, we discuss the possibility of comparing forms of economic organisation, attempting to steer a course between the Scylla of an engineering, or ‘technical’, approach to economics and the Charybdis of extreme holism or extreme kaleidics (Section 6). Brief concluding comments appear in Section 7.

2. Mises’ impossibility argument

Around the beginning of the twentieth century, it was often implied that the problems studied by economists would be specific to capitalism and should not bother socialists or collectivists. In the society that some believed the future was bound to bring, ‘economic’ problems would be replaced by ‘technical’ ones. Economics as traditionally understood would eventually become obsolete; social science would be replaced by ‘engineering’.

Mises, however, believed that the insights of economics were applicable to any form of social organisation, as human action entails economising. There is an economic problem whenever to achieve an end another end must be relinquished, and this occurs when means are scarce and have alternative uses (see Robbins, 1935, p. 13; Hayek, 1935A, pp. 122–23). To choose one course of action as opposed to another involves a judgment as to which one is best: ‘every man who, in the course of economic life, takes a choice between the satisfaction of one need as against another, eo ipso makes a judgment of value’ (Mises, 1920, p. 95).

Mises’ (1920, pp. 96ff) definition of an economic problem encompasses both the commonplace—whether to have apples or pears—and the complex—whether, and how, to build a new rail line between Paris and Vienna. But, crucially, commonplace and complex problems differ in their nature. Whereas most people can roughly solve the former, solving the latter requires an amount of data collection, computation, speculation and entrepreneurial judgment that transcends the capacity of an individual consciousness. A human mind would stand perplexed ‘among the bewildering mass of intermediate products and potentialities of production’ (1920, p. 103). According to Mises, then, in an economy in which complex economic problems are constantly being solved—more or less satisfactorily—there must exist an economically meaningful process of intellectual division of labour.¹

For Mises, this process is possible if there are prices resulting from ‘the interplay of the subjective valuations’ (1920, p. 97) in exchanges against money. Insofar as prices are the cumulative outcome of numerous such exchanges, they are reliable signposts for making economic judgments.² To be sure, the more the value of money fluctuates, or the fewer the goods that enter direct exchange against money, the less reliable is monetary calculation. Still, as long as inflation remains low and the range of monetary exchange covers a significant part of the productive structure, Mises tells us that the price system is sufficient (1920, p. 101).

But, moreover, he argues that widespread monetary exchange is actually necessary for the coordination of all actions resulting from local evaluations—i.e. that no advanced economy can exist without a price system of significant scope. Thus, he contends that the economic meaningfulness of prices stems from their being the result of competitive market exchanges and that no system of calculation can have similar economic meaningfulness without being tied up to a vast array of ‘market dealings’ (p. 111).³

If we define a socialist economy as an advanced economy with communally owned means of production and without monetary exchange against the means of production, Mises’ famous result that socialism is ‘impossible’ clearly follows.⁴ Without private property and market prices, rational choice is impossible if we accept his assumptions.

His definition of socialism was hardly a straw man. Economic calculation in natura was indeed advocated (see Chaloupek, 1990, pp. 668ff; Lavoie, 1985, p. 49), and Mises’ 1920 argument was primarily framed against this type of argument. It was certainly not intended as an argument against neoclassical economics, which, as late as 1932, Mises found close to his own approach (see Boettke, 2000, p. 5; Lavoie, 1985, p. 3; Kirzner, 1988, pp. 9ff). And, as we shall see, neoclassical economists who were sympathetic to the socialist ideal understood, and accepted much of, Mises’ argument. They sought to counter his claims on the impossibility of socialism on the terrain of ‘modern’, not Marxist, economic analysis.

3. Answering Mises’ challenge: the market socialists of the 1930s

The most important debate spurred by Mises’ 1920 article is the Socialist Calculation Debate of the 1930s.⁵ Economists like Abba Lerner, Evan Durbin, H. D. Dickinson and Oskar Lange tried to refute Mises’ challenge by proposing a partly decentralised model of socialism that could solve the complex economic problems of advanced societies. Their reasoning was rigorously neoclassical.

The most famous answer to Mises is arguably a two-part article by Lange (1936, 1937). Lange (1936, p. 54) accuses Mises of mixing up two meanings of ‘price’. According to Lange, in Mises, price sometimes stands for exchange ratio, i.e. for monetary prices issued from competitive exchange, other times for ‘terms on which alternatives are offered’, i.e. for the trade-offs the economy faces.⁶ Lange argues that, if we are given a preference scale and a list of primary factors, those ‘terms’ or trade-offs ‘are determined ultimately by the technical possibilities of transformation of one commodity into another, i.e. by the production functions.’ For him, though it is true that exchange ratios satisfactorily reproduce the ‘terms on which alternatives are offered’, they are nevertheless not identical with these terms, and therefore it is conceivable that we could construct a way of recovering those terms alternative to competitive exchange. In other words, Lange does not find in Mises an argument that establishes more than the sufficiency of competitive exchange for complex economic calculation.

In fact, Lange thinks that there is no sound argument for the necessity of prices derived from competitive exchange. A few years earlier, Fred M. Taylor, in his presidential address to the American Economic Association, had expounded a non-tâtonnement procedure that, for Lange and others, amounted to a counterexample to Mises’ challenge. According to this counterexample, the problem of imputation—‘that is, the problem of ascertaining the effective importance in the productive process of each primary factor’ (Taylor, 1929, p. 6)—is solvable by a trial-and-error procedure. First, a list of the basic, pre-existing primary factors is constructed by the Central Production Board, and tentative valuations (numbers) are given to each factor. Then, the production units, managed by functionaries, choose the quantity of each product they are to produce by taking the prices of primary factors as given and following some predefined rule, for instance equalling price to marginal cost. Finally, citizens are allowed an income to be freely spent in a market for consumption goods. Depending on the preferences of consumers, expressed in purchases, some goods will be in excess supply, others in excess demand. This is transmitted up the productive network by the mechanical rule-following of the functionaries, and results in surpluses and deficits of primary factors. New valuations of these factors consonant with these deviations are produced (following prescribed rules), and the procedure goes on until equilibrium is established. Once the problem of imputation is thus solved, everything else is easy (p. 8).⁷

For Lange (1936, pp. 63ff), the price structure resulting from perfect competition has significance because it is parametric. In fact, the competitive price structure is the only one satisfying ‘the objective equilibrium conditions, i.e. [which] equalises demand and supply of each commodity’ (p. 63). If the ‘parametric function of prices’ (p. 59) is imposed by the Central Production Board in the trial-and-error procedure—if agents are made to take every accounting price as given—then the tried and corrected prices resulting from the trial-and-error procedure will also reproduce the objective equilibrium conditions of the economy, the ‘terms on which alternatives are offered’ (p. 54). Lange (p. 66) goes so far as to write that ‘the accounting prices in a socialist economy can be determined by the same process of trial and error by which prices on a competitive market are determined’. Thus, he envisages market socialism based on the trial-and-error procedure as an improved form of what is fundamentally the same process as the private property based, competitive market system—improved because, for Lange (1937, pp. 125ff), problems like externalities could be solved more easily under market socialism by making prices more comprehensive, and business cycles could be reined in by insulating local disturbances and preventing systemic contagion.

Inevitably, many socialists did not find the conception of a market socialism compelling. Roberts (1971, p. 564) surmises that no socialist ever ‘believed in, placed hopes in, or fought for’ market socialism. Dobb (1940, p. 277) criticises ‘those who dream of marrying collectivism to economic anarchy’ and advocates proper central planning, based on the ‘fuller knowledge of the data’ available to the Central Planning Board, which is ‘a crucial element in the superiority of a socialist over a capitalist economy’ (Dobb, 1939, p. 726). No doubt many socialists would not even be interested in neoclassical models to begin with. Yet Lange’s model of a socialist economy shows that, in purely economic matters, some avowed Marxists did begin to argue like non-Marxists, as Schumpeter (1954, p. 884) approvingly remarks: ‘[T]hey learned the truth that economic theory is a technique of reasoning; that such a technique is neutral by nature … [and that there existed] a piece of ground on which it was possible to build objectively scientific structures’.

Indeed, as has often been observed, dissent between the market socialist and the Austrian camp in the Socialist Calculation Debate was ultimately due to different methodological perspectives, which implied different views on the nature of economic problems and, in particular, on the status of neoclassical theory (see e.g. Hayek, 1945, p. 89). Significantly, Schumpeter, who was sympathetic to neoclassical economics but not at all to socialism, was persuaded that Lange and the market socialists had exposed Mises’ argument as an error (see Schumpeter, 1942, pp. 172ff; 1954, pp. 985ff).

4. A Hayekian objection

According to Lange (1936, p. 55), since Mises holds that ‘without private ownership of the means of production no determinate index of alternatives exists’, then, for him, ‘the economic principles of choice between different alternatives are applicable only to a special institutional set-up.’ Lange provocatively claims that only an ‘institutionalist’ would hold such a position.

But does Lange have a point? We can perhaps conceive of notions that are universally relevant in the study of social phenomena, i.e. relevant regardless of the historical or institutional context. Consider, for instance, the notion of choice. Presumably, most of us cannot conceive of a human community, whatever its institutional set-up, where there is no agency and, therefore, no choice. Nor can we identify choice without identifying an agent. From notions such as these—agency, choice, etc.—we may be able to construct theories as universal as the underlying notions themselves. Yet this does not entail that such theories straightforwardly fit any actual states of affairs—i.e. that identifying the referents of, or the denotation of, designators is unproblematic, or even possible. A theory that wishes to have any bearing on actual states of affairs requires sentences such as ‘actions are caused by reasons’ as well as sentences such as, say, ‘Penelope undoes Laertes’ shroud because she wishes to fool her suitors’.

The market socialists are insouciant in the face of problems of this kind. They fail to realize that there is a gulf between the ‘purely theoretical’ problem that they formulate in neoclassical terms and the real economic problems that a socialist community would actually need to solve. Lange (1936, pp. 54–55) has no qualms in stating that ‘the amount of resources available’ is as much a given under capitalism as it is under socialism, whilst Taylor (1929, p. 3, italics supplied) defines primary factors as ‘for example, the land itself; the water powers; the original raw materials such as metallic ores; the different kinds of labour services, etc.’ But how can this ‘etc.’ suffice if a continually updated list of all the primary factors is required for the socialist plan to be feasible?

Hayek’s best-known objection to the market socialists (see e.g. Hayek, 1937; 1940, pp. 197ff; 1945) turns precisely on the fact that, in real life, as opposed to the neoclassical world, there is no agent who does, or could, command the knowledge necessary to solve most economic problems. On the contrary, the study of the creation, discovery, treatment and transmission of knowledge is crucial for understanding intersubjective economic coordination. To comprehend how economic problems are being solved in real life, as well as to assess the merits of alternative forms of organisation, we need to consider who is informed of what, whence that information came, how agents transformed information and other inputs into true and false beliefs, and how knowledge is created, transmitted and coordinated.

In other words, the knowledge necessary to solve most economic problems is actually dispersed in people’s heads, implicit in their routines, integrated in social rules of conduct; and we cannot make this knowledge either propositional or simple enough, or get the right incentive structure, for it to be transmitted and centralised, even in the limited degree required by the market socialists’ trial-and-error procedure. This knowledge can only be discovered through the decentralised, competitive market process.

For Hayek (1948), accordingly, the notion of competition that the market socialists make their own ‘assumes the situation to exist which a true explanation ought to account for’ (p. 94). Indeed, if it is already known what should be produced and what is the best way of producing and distributing it, what is competition about? These are precisely the questions that the process of competition is to answer. ‘It is only through the process of competition that the facts will be discovered’, whereas ‘“perfect” competition means the absence of all competitive activities’ (p. 96).⁸

Hayek’s epistemic criticism of Lange’s trial-and-error procedure and its textbook notion of competition is clearly very powerful, not to say definitive.⁹ But Lange does not just implicitly assume that primary factors could be listed and techniques of production collected into a ‘book of blueprints’—i.e. that they are known or knowable. He also takes it for granted that primary factors or techniques of production are given in the ontological sense of existing even if the concrete form of organisation were changed. Jointly with consumer preferences, these ontologically given primary factors and techniques are supposed to define the economic problem of a community.

This implicit ontological presupposition is not directly addressed by the Hayekian argument elaborated in this section, nor indeed is it discussed in the context of the Socialist Calculation Debate. We dedicate the next section to scrutinising it, building on Hayek and others. In assessing the view that things have particular economic roles, and that communities face the same allocative economic problems, regardless of the social environment in which they exist, we also articulate an alternative view that underscores the dynamic creation and recreation of economic roles.

5. The economic problem of a community

Enrico Barone and Vilfredo Pareto, identified by the market socialists as their predecessors, conclude that, at the level of fundamental theory, the economic problems of what they call ‘individualist’ (or ‘free competition’) and ‘collectivist’ societies—or, roughly, of ‘capitalism’ and ‘socialism’—are identical: ‘all the economic categories of the old [individualist] regime must reappear’ (Barone, 1908, p. 289). But it is not only categories that reappear. Their referents in an actual economy, i.e. the actual stock of primary factors, the actually available techniques of production and the tastes agents actually have, are taken to be invariant to its form of organisation. The economic problem of a community at a point in time is determined by these facts (see Barone, 1908, pp. 265–67; Pareto, 1909, p. 363, VI.58).

The market socialist interpretation of Mises’ arguments is grounded on a similar view of economic reality. Of course, primary factors, techniques of production and tastes may differ across communities and perhaps even depend on their historical ways of solving economic problems (see Barone, 1908, pp. 288–89). But given that there are so many gallons of acetic acid, there is a procedure for pasteurising milk, and agent X prefers milk to whisky, the economic problem of a community does not depend on its present form of economic organisation. After all, with a change in the form of economic organisation, acid would not evaporate, pathogens would not become resistant to pasteurisation and, at least in the short term, there is no obvious reason why agent X should reverse her preferences. In this view, then, forms of economic organisation are ways of implementing ‘technical’ or ‘engineering’ solutions to an economic problem independent from, and pre-existent to, present forms of organisation. Hence, the merit of a solution brought about by a particular form of organisation can be gauged and compared with the solution that maximises the satisfaction of all concerned, given the existing preferences, primary factors and techniques of production.

This ontological presupposition of the market socialists’ is not necessarily displaced by the epistemic argument we have relied on to criticise their trial-and-error procedure. That Hayekian argument establishes that the creation, discovery, treatment and transmission of knowledge required for an advanced economy to function depend on the decentralised, rivalrous market process. However, the fact that the economic problem of an advanced economy cannot be solved without the market process need not entail that that problem is not even defined independently of the form of economic organisation.¹⁰ Drawing on Hayek’s writings in the aftermath of the Socialist Calculation Debate, we now want to argue in some detail that, in truth, there is no economic problem of a community independent of its form of economic organisation.

What, to begin with, defines the economic role of something? Why, for instance, is something a means of production? It is easy to see that no properties inherent to things—the properties of the stretch of space-time they occupy, which we may call intrinsic properties—are sufficient to determine their role. Few would affirm that an undiscovered oil reserve belongs to a community’s set of means of production. Before people learn of its existence, and perhaps go so far as to decide that it is worth their while to do something about it, it should not be counted as a means of production. The discovery of the reserve is a change in what is true of it: it used to belong to the set of things that have not been discovered, and now it belongs to the set of things that have. The physical and chemical properties of the stretch of space-time where the reserve is found are, however, unchanged.

A similar, albeit more general, reasoning leads Hayek (1942, 1943B, 1944) to argue that the facts with which the moral sciences in general, and economics in particular, are concerned depend on the subjectivity of agents, on their attitudes: on what they believe and desire, which in turn explains what they do. It was a change in what agents believed about the world that gave the oil reserve a role as a means of production, although nothing in it changed;¹¹ what changed was the agents’ coming to know of the reserve’s existence, and their forming appropriate beliefs about its suitability and profitability as a well. The same point may be reinforced by considering a technique of production, e.g. fracking. Clearly, fracking had to be invented, described, known by agents, and its implementation had to be judged an option, before it could have any economic role. That fracking has a role as a technique of production is thus dependent on the beliefs and desires of agents, over and beyond such properties as featured in the descriptive content of a blueprint or patent.¹²

The dependence of economic roles on attitudes indicates that any thorough understanding of economic affairs must take agents’ concrete attitudes, in context, into account (see Boettke, 1995, p. 28; Lavoie, 2011, p. 115; Storr, 2010, p. 31). But what precisely are attitudes?

Attitudes can certainly not be understood as a purely individual matter: they are not things people have in their brains, irrespective of their interaction with others in a shared world (see Davidson, 1995A, p. 15). The possibility of having attitudes is intimately related to our immersion in a community of fellow creatures. Without communicating with others, and adopting their reactions to a shared world as a standard to gauge our own reactions to it, we could never have acquired an awareness of the notion of objective truth (p. 7). But only a creature with a grasp of the notion of objective truth could be said to think at all: how could a creature that does not grasp that notion believe that some proposition is, or desire that it be, true (p. 9)? It is because of the interpersonal origin of any distinction between the objective and the subjective that Davidson (p. 18) writes that ‘knowledge of other minds is basic to all thought’.

We communicate, we form expectations about the behaviour of others, and try to understand what we observe them doing, by applying skills we have developed throughout our personal histories—skills the development and application of which are mediated by social rules informing and coordinating human interaction.¹³ The exercise of these skills cumulates in an understanding of others that is expressed in a vocabulary of beliefs and desires, of reason and emotion, of choice and, if need be, of madness and mental illness. Attitudes are these ways of individuating the psychological states we and others are in, of expressing their content and their interrelations: ways that capture what is relevant for interpersonal communication (see Davidson, 1991, p. 60).

Of course, any understanding of others we come up with is contextual and provisional. It results from integrating the evidence we have available when interacting in everyday life, much of which is behavioural—what we hear others say and watch them do; and it reflects our limited, human capacity for processing this evidence in real time, in the face of ever new circumstances (see Davidson, 1995B, p. 132; Hayek, 1942, p. 91). As with any theory, this understanding is underdetermined by the evidence we employ, and even by all the evidence potentially available: there are bound to be many sets of sentences that are compatible with the available evidence and yet imply sentences that are in contradiction (see Quine, 1951, 1964).¹⁴ Any understanding of others, and thus our own attitudes, is also path-dependent, in as much as it depends on each person’s actual, idiosyncratic history. Our understanding could always have turned out differently, even if remaining true to the potentially available evidence, let alone to that which we found occasion to use or process.

If attitudes are provisional, contextual and path-dependent representations social beings come up with to keep track of psychological states and of behaviour; and if, as we have seen earlier, the economic role of things depends on attitudes; then these roles should also be taken as provisional, contextual and path-dependent. Moreover, insofar as attitudes are themselves the result of social interaction, they too depend on actual economic roles. Thus, we should not conceptualise the relation between attitudes and roles as one-sided: attitudes co-determine roles just as roles co-determine attitudes.

To this we must add that the economic role of something seldom depends on the attitudes of a single individual, but rather on appropriate relations between the attitudes of several agents (see Lawson, 2012, pp. 361ff). This dependence is not merely a result of the intersubjective dependence of attitudes themselves. It is also due to the fact that the role of things is defined by the articulation of several agents’ attitudes. Prices in a market order are a case in point: their role as ‘indicators of what ought to be done in the present circumstances’ (Hayek, 1982, p. 276) is hardly reducible to the attitudes of any individual or group, but rather stems from the dynamic, extended order of the attitudes and the consequent actions of a community of agents.

To sum up, we have established two propositions regarding the connection between attitudes and the economic roles of things. First, the economic significance of something is both created and discovered in the same process in which attitudes are formed, i.e. the economic role of things is co-determined with attitudes themselves. Second, economic roles are defined within constellations of attitudes.

These two propositions enable us to question the plausibility of a notion of the economic problem of a community that would exist irrespective of the specificities of the community’s social relations, such as its form of economic organisation. Instead, in real communities, characterised by specific social relations, various economic problems are simultaneously posed and tentatively solved, many such problems being integrations of simpler problems that agents implicitly and explicitly address (see Lawson, 2012, pp. 376ff). Different forms of organisation may lead to action being differently articulated and coordinated, to different determinations of economic significance and, consequently, to different individual economic problems—and also to a different integration of these problems, resulting in different supra-individual economic problems. Unless the social processes governing the creation of significance, the articulation of actions and the integration of individual problems were invariant to economic organisation, the market socialists’ ideal problem of a community need have no connection with the actual problems being solved in that community. Nor need the engineering significance of a particular good or technique of production have anything to do with the economic roles determined within a social web of significance.

6. Comparing alternative forms of economic organisation

The pre-existence of things with particular economic roles, perhaps even of the economic problem of a community, is often taken for granted—and not just by the market socialists. When we say that it would be better (or worse) if the present form of organisation (or a part of it) were changed, we usually (implicitly) assume that the comparison is meaningful because it compares solutions to the same economic problem of the same community.

When is this sort of reasoning warranted, though? What can reasonably be assumed to remain unchanged when we are discussing what the world would be like if the actual world were different in some respect? In other words, under what conditions is a counterfactual statement true? The truth of a counterfactual statement depends on what is the case in a really possible world that is different in the relevant regard, yet otherwise similar, to the actual world.¹⁵ Therefore, the use of different standards of similarity, as well as different understandings of what worlds are possible, generate different claims on the truth value of counterfactual statements (see Lewis, 1973).

In the market socialist outlook, the standard of similarity and the domain of possible worlds are based on the intrinsic properties of things. According to this outlook, there is a possible world in which a community has a market socialist form of organisation, different to its form of organisation in the actual world, and yet faces the same economic problem as in the actual world. This is so because primary factors and techniques of production are pre-determined economic facts. We have argued that this position is untenable. Yet, up to a point, it has some intuitive appeal. This hammer, for instance, will not change because people have changed their mind about it, nor will it hit nails less efficiently because of new minimum wage laws or of a change in the form of economic organisation.

In order to compare alternative forms of organisation, however, or to assess policy measures, a standard of economically relevant similarity and difference over really possible worlds seems to be required. In the previous section, we elaborated on the nature of this economic dimension, borrowing Hayek’s claim that economic properties depend on attitudes. We observed, furthermore, that attitudes are co-determined within intersubjective processes of interpretation and interaction, mediated by social rules; and that economic roles and problems are likewise co-determined within a social web of significance. No part of the social body is unrelated to the other parts, and nothing has significance in a social vacuum. Our preferences, viewpoint, even our sense of identity, depend on the society we live in, on the processes that turn us into acting agents and on our ways of understanding and interacting with others (see Lawson, 2012, pp. 373–74).

Incidentally, nothing in these observations is at odds with the contemporary scientific ethos that any change, including economic change, supervenes on changes in a physical dimension.¹⁶Lawson (2012, p. 346) defends an ‘ontological naturalism’, ‘a non-dualist orientation that entails that even features such as life, choice and intentionality are integrated with the (rest of the) natural world and not composed of some separate (non-naturalistic) stuff’. Thus, the discovery of the oil reserve (in the previous section) certainly involves physical events somewhere: in agents’ brains, perhaps. Yet, from the thesis that every economic event supervenes on physical events, it does not follow that we could individuate economic events with a vocabulary that does not express attitudes, that we could reduce types of economic events to types of physical events (see Davidson, 2009, p. 56). To understand others in an intersubjective world, we employ rules of inference and judgments of similarity that have no parallel in physical theories (see Davidson, 1970). The relative autonomy of economic (or, more generally, social) theory must therefore be underscored.

But how in the first place are we to have an economic theory of a world where everything is interrelated, and indeed in flux? As Lawson (1997, pp. 30ff; 2012, pp. 357ff) argues, we do successfully rely on relatively stable social rules and social structures, with relatively predictable local manifestations, to coordinate our actions with those of others. Intentional action is possible because such social rules and structures exist and human agents have a degree of (usually tacit) knowledge of them; and it is explicable because knowledge of such rules and structures, and of their agent-dependent mode of operation, can be theoretically formulated.

Such explanations are always partial and contextual—though in very different degrees—and their different purposes and goals imply that a considerable amount of complexity can (and indeed must) be disregarded.¹⁷ In our oil reserve example, there is no harm in ignoring the idiosyncratic history of the discoverers, the effect of ideology or special interests in the pursuits that lead to the discovery of the reserve and its acquiring an economic role, or the dynamic co-dependence between attitudes and economic roles. On the other hand, if we were to discuss the effects of a change in the economic organisation of a community on the development of oil-related technology, ignoring such issues would be unwarranted.

Let us take stock. We proposed that economic roles depend holistically on a social web of significance. But, in such a world, how can we form beliefs or act? How can we reasonably compare forms of organisation or assess policy measures? In line with others, especially Tony Lawson, we have argued that there exists a degree of stability in the social world, which enables agents to formulate plans and social theorists to construct explanations, however partial and contextual, of the complex phenomena that interest them. In particular, we now add, there is often a locally stable, conventional relation between economic roles and the intrinsic properties of things. The latter sometimes offer grounds for recognising economic roles in a way that is sufficient for most ordinary purposes (see Lewis, 2013, p. 391). In Hayek’s (1943A, pp. 65–66) words, ‘as long as I move among my own kind of people, it is probably the physical properties of a bank note or a revolver from which I conclude that they are money or a weapon to the person holding them.’

There is a path, then, that avoids an extreme holism and kaleidics, a pit where we pull a tiny thread and the whole fabric is altered, where all continuity ceases. Change often occurs on what is, for practical purposes, a relatively unchanging background of significance, enabling us to identify possible worlds different, yet sufficiently similar to ours in the relevant regards, for us to understand what is at stake. Small, localised changes in a community will not radically affect the economic roles of things because—or where—there is a relatively stable connection between these roles and the intrinsic properties of things. Within these limits, there are really possible counterfactual worlds in which a community addresses economic problems similar to those it faces in the actual world.

But, to return to our starting point, it is wrong to assume that, because some changes may plausibly be regarded, for certain purposes, as taking place against an enduring background of significance, the organisation of a whole economy could change whilst that background would somehow endure. There is no really possible world in which a community has a radically different form of economic organisation and yet faces the same economic problem as in the actual world.

7. Concluding comments

The Socialist Calculation Debate was a singular episode in the history of economic thought. It enabled the articulation of different methodological outlooks, and it spurred some of those involved, in particular the Austrians, to achieve a better understanding of the specificities of their worldview. For a long time, it was widely believed that the Austrians had lost the debate. This was due to the formal demonstration that in a model of a market socialist economy the same formal solution to the same formal problem was achievable as in a competitive economy. However, even if the economic problem of all communities is thought to be formally the same—i.e. if one accepts that it can be described with the same symbols and sentences (with the same formal theory)—the discovery of what those symbols refer to in the actual world is part of the economic problem solved by any real community. The practical necessity of ascertaining economic roles in the actual epistemic conditions of everyday life is the insight of Section 4.

But economic roles are actually co-determined within contingent forms of economic organisation. The social embeddedness and dynamic determination of the economic dimension, its dependence on intersubjective processes of interpretation and understanding, and on other social processes of supra-individual coordination, is the insight of Section 5.

The market socialist outlook is based on an engineering approach to the economy. This outlook is congruent with the intuition that we can coherently compare economic systems after changes have been made to them, or compare different forms of organisation. In scrutinising this intuition, we have argued that the relative stability of social rules and structures enables us to avoid the scepticism that might be entailed by too extreme an emphasis on the holistic and self-transformative nature of social processes. Still, one must not expect more than what Hayek (1942, p. 106; 1955; 1989) terms ‘explanations of the principle’ when discussing complex social processes. This is the insight of Section 6.

Footnotes

1

‘The distribution among a number of individuals of administrative control over economic goods in a community of men who take part in the labour of producing them, and who are economically interested in them, entails a kind of intellectual division of labour, which would not be possible without some system of calculating production and without economy’ (Mises, 1920, p. 102). In other words, Mises implies that in a complex economy each agent performs, and can only perform, a small fraction of the mental operations required for the economy to work. Instead of ‘intellectual’, the adjective ‘mental’ is used in the translation of Mises’ Gemeinwirtschaft (see Mises, 1932, p. 101). The translation ‘division of knowledge’ appears in Hayek (1937, p. 50). Mises’ original passage reads ‘eine Art geistiger Arbeitsteilung’.

2

Mises’ position has similarities with Pierson’s (1902). Pierson emphasizes the incommensurability of goods in the absence of a standard of value, writing that ‘we cannot subtract cotton, coal and the depreciation of machines from yarns and textiles’ (p. 70). This calculation requires ‘evaluation’ (pp. 82–83) and is done through monetary exchange (p. 72). ‘It is not a purely technical problem which is here in question, but rather a decision as to the most profitable way of employing material things; and the rightness of such a decision must depend upon the rightness of the evaluation which preceded it’ (p. 59).

3

In addition to this, Mises (1920, pp. 110ff; 1932, pp. 186–87) also discusses problems of ‘responsibility and initiative’, arguing that entrepreneurs can be made responsible for the profits and losses that result from their choices only with private property (see Boettke, 1998, p. 137, for a discussion of the pivotal role of private property in Mises’ warnings against socialism). As capitalism was, for Mises’ socialist opponents, more bureaucratic than socialism, this line of reasoning was largely ignored in the Socialist Calculation Debate. Lange (1937, pp. 123ff) writes that this is a problem for sociologists, not economists (see also Durbin, 1936, p. 676; Lerner, 1937, pp. 267ff). Interestingly, Lerner (1938, pp. 74–75) states that ‘to agree that managers will not manage prudently unless they manage with their own money is to agree with von Mises that socialism is impossible’. Shortly afterwards, incentive problems became a central concern in the economics of planning (see Bergson, 1948).

4

For Mises, the impracticality (Undurchführbarkeit) of socialism and its impossibility (Unmöglichkeit) are the same proposition. According to Salerno (1990), impossibility is closest to his intentions. But see Lavoie (1985, pp. 152–53). Significantly, authors who reason from a neoclassical perspective separate ‘theoretical’ and ‘practical’ problems: Joseph Schumpeter distinguishes the ‘logical’ soundness of socialism, whose blueprint is alleged to be more rational than capitalism’s, from ‘practical’ difficulties in implementing it (see Graça Moura, 2017).

5

For a survey of the German-language debate on socialism in the 1920s, see Chaloupek (1990), who agrees with Mises’ (1996, pp. 706ff) assessment that this debate was intellectually inferior to that of the 1930s. Keizer (1987) examines Mises’ interventions in this German-language debate.

6

Lange borrows the distinction of these meanings from Wicksteed (1910, p. 28): ‘“Price,” then, in the narrower sense of “the money for which a material thing, a service, or a privilege can be obtained,” is simply a special case of “price” in the wider sense of “the terms on which alternatives are offered to us” … (how much of this against how much of that?).’

7

Dickinson (1933) also develops a procedure based on ‘successive approximation’ (p. 241). His framework is like Taylor’s for the first iterations. But once ‘the system has got going’ (p. 242), he tells us that we may dispense with the trial-and-error setup, and simply solve the problem mathematically, with demand and production functions. Dickinson states that the problem is not even one of solving equations but merely of calculating ‘small deviations from an already established equilibrium’ (p. 242). After Robbins’ (1934, p. 151) and Hayek’s (1935B, pp. 152ff) criticisms that attempting to solve the economic problem with equations is absurd, to begin with because of computational difficulties, this sort of proposal was abandoned (see e.g. Durbin, 1936, pp. 677–78; Lerner, 1937, p. 254).

8

See also Hayek (1968, pp. 306–7). Hayek’s conception of the knowledge problem in an advanced economy is, of course, central to the Austrian vision, in which the processual nature of the market is brought to the fore (see e.g. Boettke, 2000, p. 5; Lavoie, 1985, pp. 22ff). Here, agents are not passive optimisers in a static setting; instead, they change the world in unique, contingent ways (see e.g. Mises, 1996, p. 26; Hayek, 1952, pp. 53, 63, 108). Operating within a relatively stable institutional setup, they continually recreate the intersubjective tokens that render some of their subjectivity—especially the dispersed, contextual knowledge that guided their actions—available to others (see Lavoie, 1987, pp. 599–600 for a discussion of the ‘capitalistic communication process’ and a comparison between conversation in a natural language and ‘the economic process of competition in the accounting language’). Because market prices result from competitive exchange, they are meaningful guides for the identification and selection of economically viable plans of action, facilitating the correction of errors and the coordination of plans. Knowledge is brought to bear that would have to be disregarded in a centralised setting, not just because of its non-propositional nature, or for reasons of complexity, but because its creation and transmission could not even exist outside the market process (see Boettke, 1998, p. 145). In short, the market process enables the discovery and use of knowledge sufficient to solve, though not ‘optimally’, complex economic problems (see e.g. Kirzner, 1990, p. 45; Lewis, 2015, p. 1187).

9

There have been attempts to develop sounder—i.e. post Hayek—forms of market socialism (see Bardhan and Roemer, 1993; Burczak, 2006). Whether these proposals are immune to the objections raised above is, however, disputed (see Storr, 2007).

10

Of course, this statement must not be read as to mean that Hayek would accept the existence of an economic problem of a community independent of its form of organisation. Whilst his writings at the time of (or shortly after) the Socialist Calculation Debate do not elaborate much on this (see, however, Hayek, 1935B, pp. 154–55; 1948, p. 101), he later underscores that competition is not just a procedure for discovering knowledge that would not otherwise be known. Rather, it enables the extended order within which economic significance is created (see e.g. Hayek, 1982, pp. 267–90).

11

The intrinsic properties of things, to repeat, are insufficient to define their economic roles. However, a few intrinsic properties may be necessary. For instance, potassium cyanide, which lacks any nutritional value, could not possibly be counted as food, even if a group of people believed it to be wholesome and treated it as such. It is also implausible that anything without mass could ever be a hammer, even if someone believed it was possible to hit a nail with a mental device of their invention. But necessity itself may depend on agents’ attitudes: whatever nutritional value something needs to have for an informed agent to conceptualise it as food may depend on the context.

12

It might be tempting to assert that something’s potential economic role in a community is independent of agents’ attitudes, even if appropriate attitudes in appropriate contexts are necessary to actualise a role. This, however, would be wrong, and would lead us to the implausible conclusion that any as yet undiscovered oil reserve is, today, a potential well, or that fracking was a potential technique of production in Ancient Greece.

13

These skills involve heuristics, judicious selection of prejudices and the following of tried and tested inferential paths. For instance, as many have remarked (see Hayek, 1942, p. 89; 1943B, p. 139; Quine, 1964, pp. 59ff; Davidson, 1967, p. 27), we usually take others to be much like us. We usually assume off-hand that others share most of our fundamental beliefs (‘cars are four-wheeled vehicles’), make similar judgments of salience (‘look over there!’) and agree with us on broad ethical and aesthetical matters (‘I cannot understand why someone would do such a thing…’).

14

Of course, the circumstance that many hypotheses fit the available evidence does not mean that there are no good reasons for preferring one to another. Two theories equally justified by evidence may differ in charity or humanity (see Lourenço, 2016).

15

In other words, the alternative, counterfactual world must represent a real possibility. As Lawson (2003, p. 19, italics supplied) notices, ‘the sort of fictitious assumptions thrown up within modern mainstream economics do not involve claims that could be true in some really possible counterfactual state of our world. I interpret the latter as part of the domain of the real.’

16

As Quine (1978, p. 98) puts it, ‘nothing happens in the world, not the flutter of an eyelid, not the flicker of a thought, without some redistribution of microphysical states’.

17

That is, only a subset of the operative causal factors matter for particular explanations: in Lewis’ (2013, p. 398) words, ‘those [factors] whose influence is indicated by the existence of differences or contrasts between the history of the phenomenon under investigation (termed the “focus”) and the history of another, similar (but not identical) phenomenon (the “foil”)’. Contrastive explanations, which Lewis is discussing in this passage, are often used to assess the truth of counterfactual statements such as ‘if the focus were like the foil in this or that regard, then the focus would etc.’