The Particular Difficulties of the Social ‘Sciences’

Theorists have often sought to differentiate the social disciplines from the natural sciences on the grounds that the latter are more objective. Indeed, an invidious comparison is often made between the social disciplines and the natural sciences. This follows from a tendency to idealise the natural sciences and to see Newtonian physics as the exemplar of scientific practice. It is then assumed that the production of universal laws characterises the natural sciences in general—but this is far from being true.[59] Such a sweeping generalisation does not do justice to the diversity of scientific practice in the natural sciences or to the variety of criteria of success negotiated within those diverse fields.[60] Simple law-like behaviour and predictability are elusive in the natural sciences also—though in the natural sciences it is possible more often to get away with simple idealisations, to isolate a system and to treat its properties as context independent.[61] In any event—as the above account makes clear—the broad claims of the natural sciences to objectivity—in the sense advanced by the Enlightenment tradition—cannot be sustained.

Nevertheless, there are particular difficulties with the social disciplines, which add to the above problems, and which are reflected in unease about their status. The result has been the development of a separate theoretical discussion of the philosophy of the social sciences, which can be quite esoteric. This discussion is at pains to distinguish itself from the positivism criticised earlier—though there are still unreformed positivists in economics. William Outhwaite categorises this discourse into three schools—involving realist, hermeneutic and pragmatic perspectives—though there appears to be significant overlap between them.[62] Nevertheless, the main issue separating these perspectives is the extent to which any social discipline can describe a social reality independent of the observer and her or his description of it. This discourse overlaps with that in the natural sciences described earlier. Critical realism—following Roy Bhaskar—is possibly the current dominant school. It agrees that a distinction is to be made between the natural and the social sciences, that the latter do not operate in the same way as the former and cannot be studied with the same methods, and that social life is constructed continually through practice.[63] Nevertheless, in neglecting the limitations of language, they attempt ‘to privilege a concept of the real that can be definitely discovered, described and activated under definable conditions’.[64] In this, they appear to be too optimistic. As educationalist and methodologist John Schostak explains, symbolic representation—including through language—can never be the full measure of the ‘real’.[65] There is something missing of the ‘real’ in any representation that we cannot recover, however much we try to tame it. Schostak suggests that for critical realism to be useful, it has to deal successfully with representation in all its possible articulations, and with the emergence of understanding as acts of creative imagination shared through discourse. This is why I lean towards the pragmatic and hermeneutical schools.

None of the above positions suggest that we should not try to understand the social world. The disagreement is about the extent to which we are likely to succeed and the confidence with which we are prepared to apply the resulting insights. No one is claiming that in any particular investigation there is a single, ultimately true theory that is accessible to us. Nor can we ever fully escape the language with which we describe the social system. In short, the ‘TRUTH’ about society is not available to us.

For example, Kincaid, who describes himself as a realist—believing the idea that things exist and act independently of our descriptions—claims that because there is no simple logic of science we cannot evaluate social science by looking at simple formal traits. At the same time, he believes that good social science cannot be ruled out on a priori conceptual grounds. Rather, he claims we have to look in detail at the methods used and the kinds of evidence adduced. Importantly, he concludes that large parts of the social disciplines have failed to produce such good science.[66] He goes on to claim that the philosophy of science can contribute to the study of society only if it eschews a priori armchair theorising in favour of a philosophy tied intimately to the real practice of social science research. In respect of that social science practice, Fiske and Shweder tell us:

It is obvious that social science is not a single integrated discipline; rather it is a collectivity of endeavours sometimes working cooperatively, sometimes borrowing from each other, and only occasionally collaborating in joint enterprises. It is a range of disciplines and methodologies, above and beyond the somewhat anachronistic categories in university catalogs.[67]

Kincaid agrees that the social disciplines employ methods that are not found anywhere in the natural sciences.[68] Nevertheless, he claims that the social disciplines can be good science by the standards of scientific adequacy of the natural sciences—describing basic patterns found in nature—but only by meeting those standards. This is because he believes that human beings are part of the natural order and are amenable to scientific understanding. This, he declares, is simply an extension of an Enlightenment tenet. Given our critique of the Enlightenment, this is hardly a persuasive argument. Furthermore, he believes that behind the diverse methods of the natural sciences there is a common core of ‘scientific rationality’, which the social disciplines sometimes share.[69] Importantly, he believes that social science is distinct from psychology—with its own domain of inquiry largely to do with understanding large-scale social structures—and in the process rejects the methodological individualism of much of the social disciplines. Interestingly, Kincaid goes on to define those scientific standards in terms of ‘scientific virtues’—virtues promoting confirmation and those promoting explanation—standards that deny that scientific justification can be reduced to a certain method. It should already be clear that Kincaid agrees with Rorty and that methods do vary across the sciences and do not provide a foolproof, mechanical basis for choosing theories. Nor does Kincaid believe all is well with social research. Nevertheless and confusingly, Kincaid appears to believe that there is something special about science, that, in effect, it possesses a privileged form of justification—a belief I have already discounted.

Kincaid’s belief that human beings are part of the natural order goes to the heart of the problem of social inquiry. This is a belief that we must reject as being much too strong. In effect, Kincaid seeks to defeat the dichotomy made in our vocabulary between the natural and the social—a vocabulary he uses while denying its import. While the phenomena studied in the natural sciences could have an existence independent of the concepts used to describe them, this might not be true very often, if at all, of the social disciplines.[70] Rather, the social disciplines are concerned with human beings who—as we saw in Chapter 2—construct their social reality, defining themselves in symbolic forms with shared understandings of the world, which they use to structure their actions.[71] Consequently, it is not the way the world is, but the way we conceptualise it, that influences our actions.[72]

This is the reason why leading Canadian political philosopher Charles Taylor makes a distinction between human ‘behaviour’ and human ‘action’, in which the former is caused by forces over which the individual has no control—analogous to the forces of nature—and the latter results from that person’s intentions. He then points out that the language describing human conduct is mainly an intentional one and it is about human action rather than human behaviour. It is the language of reasons and not of causes. This is important because—as we have seen already—the interpretation of any phenomenon depends on the language available to us, bringing with it particular theoretical entailments. The meaning of everyday behaviour and even the very fabric of society are woven into our ordinary vocabulary.[73] It is also clear that the meaning we attach to human actions depends on the particular circumstances with which we are dealing. Importantly, social structures and institutions play a large role in determining our actions. Secord seeks to clarify the situation, telling us that while social structures have real effects, they are different from natural structures in that they do not exist independently of our conceptions; nevertheless, they precede the individual. Such ‘structures preceded the entrance of individuals into society, and individuals act within them as a medium’.[74] This is a view I endorsed in Chapter 2.

Additionally—as has been pointed out already—language, including the language used in the social disciplines, is inherently metaphorical. Similarly, the interpretation of any text and of any situation is dependent largely on historically situated conventions. Gergen draws our attention to the way in which the particular literary figures used dominate the process of interpretation.[75] He reminds us that, once a particular metaphor is selected, it restrains what else can be said. The root metaphors differ across the social disciplines, providing different perspectives—ideologies even—which are difficult to reconcile.[76] These stories—these definitions of ourselves—reflect to some extent the stories that social researchers tell. Our stories, therefore—our language games—cannot be objective or normatively neutral, as we will see in greater detail in the next chapter. While this is also true of the natural sciences, there is almost a qualitative difference in the extent to which these respective disciplines can aspire to objectivity. One further consequence is that generalisations in the social disciplines are generally narrow in scope.[77] Nagel suggests:

[The] conclusions reached by controlled study of sample data drawn from one society are not likely to be valid for a sample obtained from another society. Unlike the laws of physics and chemistry, generalisations in the social sciences therefore have at best only a severely restricted scope, limited to social phenomena occurring during a relatively brief historical epoch within special institutional settings.[78]

Similarly, Gergen tells us that there are few patterns of human action that are not subject to significant alteration, while cultural anthropologist Roy D’Andrade records that the different fields of science have different canons of generalisation.[79] While researchers aspire to tell integrative stories, it could be simply inappropriate for social researchers to seek to emulate the natural sciences in an attempt to derive ‘fundamental general laws’ describing human conduct. Cronbach argues that this particular idealisation of scientific research—the development of general lasting laws on the model of parts of physics[80] —is not achievable in the social disciplines.[81] It might also not be achievable in much of the natural sciences. Nor is there any good reason to expect a unity of method across the social disciplines. On the contrary, Fiske tells us that such knowledge is fragmented, composed of multiple discrete parcels—a consequence of the different objects of inquiry and different methods of knowing. As a result, these bodies of knowledge are likely to always remain separate.[82] In particular, generalisations and theories in the social disciplines are rarely abandoned because most conceptual statements in those disciplines are formulated in such a way that they cannot be falsified. Fiske suggests that, in part, some of these difficulties arise because of too high a level of aspiration on the part of the social researcher.

All of this suggests that a strong onus lies with the theorist intent on developing systems of interrelated generalisations in a particular area of human activity to demonstrate that such generalisations do exist and then to delineate their scope. Consequently, the question arises as to whether neoclassical economics has discharged that obligation. I think not. As Ormerod tells us, the idea that people respond to economic incentives could be a universal generalisation, but the strength of any response to any particular set of incentives is emphatically not universal; it depends on the social, institutional and historical context. Human beings are not compelled to act by social ‘forces’ in the same deterministic way that natural phenomena respond to natural forces. Weber suggested therefore that the natural sciences were concerned with erklären or explaining focused on causality, while the social disciplines were concerned with verstehen or understanding. Such things as meaning, intention, ideas, values and emotions were, according to Descartes, non-things and were beyond the reach of the mechanical sciences.[83]

One approach used in an attempt to get around this problem is to consider reasons as causes. While Weber agreed that there was a logical distinction between natural and social reality, he did not believe that these differences required different scientific methods. He believed that uniqueness and historicity were features of natural as well as social phenomena. In any event, with his positivist, rationalist bent, Weber sought a rigorous method that would enable claims made about the social world to be subjected to empirical validation. While Weber accepted that no conceptual system could do full justice to the complexity of particular social phenomena, the tool he adopted for this purpose was the concept of an ‘ideal type’—an idea used also by Mill and his contemporaries. This idea—a reflection of the perfectionism and transcendentalism embedded in Western thought and in particular the positivism popular at the time—is the conceptual source of the idealisation of the market in economics. An ideal type is an analytical construct, a rationalised reconstruction, a stereotype, a fiction even, deliberately exaggerating what are thought to be typical actions to produce a coherent whole in an attempt to get to the essence of a social reality—assuming in the process that there is such an essence to be got at. As such, it looks suspiciously like an attempt to revive Plato’s forms in the context of the social disciplines. The ideal type was to be derived inductively from historical reality, though it would never correspond with reality. Importantly, Weber thought this tool could be applied only to social behaviour that was rational and goal oriented, which he believed was increasingly dominating Western society. In this regard, it is important to remember that Weber conceived of four different orientations towards social action—instrumentally rational, value rational, affective and traditional—though these categories were not intended to provide an overall classification. As we will see in the next chapter, the rational, instrumental nature of much economic activity is open to devastating criticism. In these circumstances—on the basis of Weber’s own qualification—it can hardly be assumed to apply to economics.

Furthermore, the technique is open to misinterpretation resulting from the common metaphysical assumption that ‘scientific laws’ are authoritative—that is, that they determine the way the world is (that scientific generalisations, ‘laws’, are causal agents) rather than being simply descriptions of the way the world is. In the absence of a god—conceived of as a lawmaker, dictating the laws of nature and of human conduct in the way that the Enlightenment and Smith had assumed—it is hard to imagine where any authoritative force could come from. No one these days, however, thinks that the invisible hand of the market is the hand of God. Perhaps, given Weber’s restriction of this method to the analysis of rational social action, it is rationality that is to provide this authoritative force. If so, it will just not do. To claim that the world is inherently rational or even mathematical, as the Pythagoreans thought, is only to postpone the question momentarily, as well as to overlook the problematic nature of those concepts. What gives rationality or mathematics an authoritative force? In any event, the critique of rationalism and mathematics in Chapter 5 undermines all such pretensions. Additionally, the work of the Nobel Prize-winning economists Simon, Daniel Kahneman and Vernon Smith on our cognitive limitations has undermined it at the empirical level. The fact that scientific laws do not have authoritative force has another important implication: the natural and social worlds are not, in principle, ultimately explicable.

If, on the other hand, one assumes that social laws are simply attempts at describing the way the social world is, the meaning to be ascribed to any such ‘laws’ based on unrealistic idealisations is problematic. While it might be interesting to some people to speculate about how people might behave if they were entirely economic beings, the value of such speculation and their ‘tendency laws’ to policy decision is far from certain when we all know that the assumption is false. Such idealisation is a highly reductionist strategy, with its origins in ancient Greek atomism, which attempts to reduce physical reality to fundamental and identical particles. One can complain justly that economic systems cannot be dissected in this fashion. Weber was not aware of the difficulties later theorists found in modelling interdependent complex systems. They are not open to this reductive strategy. Importantly, the idea that the factors left out can be added back in to form a more complete description—for example, the idea that economic analysis deals with ‘tendency laws’—assumes that such entities are separable in the first place and are independent. They might not be if, for example, we are dealing with non-linear dynamic systems.[84] If they are not independent and it is improbable that they would be, such influences cannot simply be added together. Complex or non-linear dynamics could produce multiple possible solutions, while even very small changes in initial conditions could produce drastic changes in outcomes.

What this means is that human behaviour is not describable by simple deterministic models. Such reductionism has a systematic bias in that it ignores or over-simplifies the importance of the context of the system being studied.[85] This simplification of the context ‘also often legislates higher-level systems out of existence or leaves no way of describing inter-systemic phenomena appropriately’.[86] Indeed, ‘assumptions that appear benign at such an individual level may be dangerous over-simplifications when viewed from a higher level’.[87] This is, of course, what we find with Margaret Thatcher’s claim that there is no such thing as society and with the methodological individualism practised by economics. It is this reductionism and methodological individualism that leads directly to the modelling of society as if it is based on self-serving individuals. I have already drawn attention to the fact that this modelling is not a neutral strategy, and have expressed concern about the potential impact of such modelling on society itself. This concern leads me to question whether methodological individualism is a legitimate, albeit potentially dangerous, analytical strategy or simply a cloak disguising the ideological prejudices embedded in neoclassical economic analysis. If our study is intended to influence our policy decisions—and if there are reasons to believe that there are higher-level social structures that impact on our social problems—we are honour bound to study them. Additionally, the conclusions drawn from such simplifying assumptions could simply be the artefacts of those assumptions with little or no connection with the phenomena that we are supposed to be studying.

A further problem surrounds how to choose the ideal type—what is to sit within the system to be examined and what sits outside as the ‘context’. This is hardly a normatively neutral exercise and it is a problem for which no persuasive answer has been given. A further and fundamental question surrounds whether such ideal types lead to generalisations that are, in fact, empirically falsifiable. Given what was said above, they certainly cannot be verified. We will return to this topic in Chapter 8 when we discuss the content of economics more directly. It is important in the interim to remember that while Weber was a positivist, he never intended these ideal types to be used as normative ideals. For Weber, any understanding of causation in the social disciplines is a result of ‘an interpretative understanding of social action and involves an explanation of relevant antecedent phenomena as meaning-complexes’.[88] This seems a far cry from the deterministic, mechanical modelling of neoclassical economics.

In 1953, American economist Milton Friedman (1912–2006) offered a radical and highly influential new defence of economic idealisation in The Methodology of Positive Economics. Friedman accepted that experience reflected the complex influences of numerous causes and therefore could support numerous interpretations. This is consistent with the position taken here. Friedman believed that it was possible to subject our policy beliefs to empirical test and that the role of economic theory was to provide a system of generalisations able to generate predictions that could be checked against experience. Because no decisive disproof was possible of any hypothesis, however, we should have confidence only in those hypotheses that survived many tests and performed consistently better than the alternatives. Friedman was interested only in empirically meaningful and testable hypotheses as an engine of analysis for the problem at hand, rather than as a description of reality. For Friedman, the reality of a hypothesis was simply irrelevant. He claims that ‘the more significant the theory, the more unrealistic the assumptions…To be important a hypothesis must be descriptively false in its assumptions…the relevant question to ask about the “assumptions” of a theory is not whether they are descriptively “realistic”, for they never are.’[89]

To most of us this seems transparent nonsense—a further ad hoc rationalisation to defend the indefensible. Perhaps, however, he was just confused, as claimed by leading institutional economist Geoffrey Hodgson[90] and applied economist Daniel Bromley.[91] Hodgson points out that this position is not only theoretically incoherent; it has not been adopted in practice.[92] Hodgson draws a distinction between different kinds of assumptions: negligibility, domain and heuristic. Negligibility is where some factor will have a negligible impact on the result; domain assumptions specify the domain in which the theory is applicable; and heuristic assumptions are simplifying assumptions made in the early stages of a theory to allow successive approximations. Hodgson argues that Friedman is talking about negligibility assumptions and that it is not true that such assumptions are descriptively false, only that they have a negligible influence on the phenomenon being explained and consequently can safely be disregarded. Clearly, the core assumptions of neoclassical economics are not ones leaving out factors that have a negligible influence, but are truly descriptively false. This does matter.

Of course, Friedman’s claims could be defended on the instrumental ground that the truth of a theory is irrelevant and that all that matters is the accuracy of the resulting prediction. Surely, however, the objective of such studies is not simply to make predictions—desirable though that might be—but rather to provide credible explanations? This is the generally accepted position in the philosophy of science. This instrumental approach would eliminate explanation and falsification from science and that should be the end of the matter. Friedman does not apply this criterion consistently in his own work. Rather, he uses Popper’s falsification criterion in the case of the maximisation hypothesis. He neglects, however, to provide any relevant evidence for his claims and asserts that a failure of critics to develop any coherent, self-consistent alternative provides evidence of the worth of the maximisation idea. In any event, it is extremely doubtful that this approach has, in fact, led to successful prediction. Indeed, given the complex nature of economic systems and the sensitivity of non-linear models to initial conditions, the very possibility of making reliable predictions is being undermined.

German philosopher Wilhelm Dilthey (1833–1911) goes further than Weber, contrasting verstehen with erklären and suggesting that natural and social reality are different kinds of reality requiring different investigative methods—a position more in line with the position taken here.[93] British sociologist Anthony Giddens describes this aspect of the social disciplines as a ‘double hermeneutic’:

The theory-laden character of observation-statements in natural sciences entails that the meaning of scientific concepts is tied to the meaning of other terms in a theoretical network; moving between theories or paradigms involves hermeneutic tasks. The social sciences, however, imply not only this single level of hermeneutic problems, involved in the theoretical meta-language, but also a ‘double hermeneutic’, because social-scientific theories concern a ‘pre-interpreted’ world of lay meanings. There is a two-way connection between the language of social science, and ordinary language. The former cannot ignore the categories used by laymen in the practical organization of social life; but on the other hand, the concepts of social science may be taken over and applied by laymen as elements of their conduct. Rather than treating the latter as something to be avoided or minimised as far as possible, as inimical to the interests of ‘prediction’, we should understand it as integral to the subject–subject relation involved the social sciences.[94]

There has been a broad recognition that a proper understanding of the social disciplines requires an appreciation of the hermeneutical dimension of them. In fact, there is a convergence between the insights of the hermeneutical tradition and the insights derived from the pragmatism that is influencing the philosophy of science outlined above. As Gadamer tells us:

When Aristotle, in the sixth book of the Nichomachean Ethics, distinguishes the manner of practical knowledge…from theoretical and technical knowledge, he expresses, in my opinion, one of the greatest truths by which the Greeks throw light on the ‘scientific’ mystification of modern society of specialisation. In addition, the scientific character of practical philosophy is, as far as I can see, the only methodological model for self-understanding of the human sciences if they are to be liberated from the spurious narrowing imposed by the model of the natural sciences.[95]

For Aristotle, there were three intellectual virtues: epistēmē, phronēsis and techne. As we have already seen, epistēmē is the kind of certain geometric knowledge to which the natural sciences aspire. Phronēsis is the kind of practical wisdom we all use in the expert social practice and moral judgements we make in day-to-day life; this was, for Aristotle, the most important of the intellectual virtues.[96] Techne is technical knowledge or technology. For Aristotle—as for Toulmin, Gadamer, Mary Hesse and Bent Flyvbjerg—it is phronēsis that provides the appropriate methodological model for the social sciences. From this perspective, the natural and social sciences are simply different intellectual ventures. In short, not only have the social sciences—including economics—not achieved practical success in providing certain predictive epistemic theory, they cannot in principle aspire to that certain geometric knowledge of epistēmē. Importantly, it is a view that denies that knowledge of human activity can ever be universal and context-independent in the same way as knowledge in the natural sciences. As Flyvbjerg—following Dreyfus—argues, ‘a theory which makes possible explanation and prediction, requires that the concrete context of everyday human activity be excluded, but this very exclusion of context makes explanation and prediction impossible’.[97]

The actors in a concrete situation will not necessarily conceive of any action in the same way that any attempt at a context-free definition of a social action based on abstract rules or laws might do. Importantly, context-dependence does not imply a more complex form of determinism but an open, contingent relationship between context, action and interpretation. Consequently, it is not meaningful to speak of theory in the natural science sense in the social disciplines. We will return to the ontological consequences of context and openness for neoclassical economics in Chapter 8. This limitation of the disciplines is no real failing, as epistēmē in turn cannot provide the reflective analysis of values that is at the heart of political, economic and cultural life. In this spirit and consistent with Toulmin, Flyvbjerg calls for the social sciences to be restored to their classical position as practical intellectual activities, clarifying the problems, risks and possibilities involved in social and political praxis.

Bernstein and prominent American economist Deirdre McCloskey tell us that part of the problem arises from the English word ‘science’ and the distinctions we English speakers make between the natural sciences, the social sciences and the humanities. In contrast, in German, a distinction is made between the natural sciences and the moral sciences only. The consequence has been that English speakers tend to think of the social sciences as natural sciences concerned with individuals in their social relations, on the assumption that the social sciences differ in degree but not in kind from the natural sciences. In contrast, German speakers have a much greater tendency to think of the social disciplines as moral sciences, sharing essential characteristics with the humanities. McCloskey goes on to advocate the adoption of the word ‘discipline’ to describe these social investigations.

From this perspective, the sciences should be seen as a confederation of enterprises, with methods and patterns of explanation to meet their own distinct problems—not the varied parts of a single, comprehensive, ‘unified science’.[98] The Platonic image of a single, formal type of knowledge is replaced by a picture of enterprises that are always in flux and whose methods of inquiry are adapted to the nature of the case. Importantly, the belief that we can start again by cutting ourselves off from inherited ideas is as illusory as is the hope for a comprehensive system of theories. The hope for certainty and clarity in theory has to be balanced with the impossibility of avoiding uncertainty and ambiguity in practice. We need to reappropriate the reasonable, tolerant, but neglected legacy of humanism more than we need to preserve the systematic, perfectionist legacy of the exact sciences. In particular, formal calculative rationality can no longer be the only measure of intellectual adequacy; one must also evaluate all practical matters by their human ‘reasonableness’. Consequently, for Toulmin:

[T]he charms of logical rigour must now be unlearned. The task is not to build new, more comprehensive systems of theory with universal and timeless relevance, but to limit the scope of even the best-framed theories, and fight the intellectual reductionism that became entrenched during the ascendancy of rationalism. It calls for more subdisciplinary, transdisciplinary, and multidisciplinary reasoning.[99]

In particular, Toulmin believes that biology provides less constricting analogies for thinking about social relations than does physics. In the organic world, diversity and differentiation are the rule and not the exception. The universality of physical theories is rare. In this spirit, Wimsatt recommends that, in studying human behaviour, we should use a variety of models and approaches in the hope that we can thereby detect and correct for biases, special assumptions and the artefacts of any one approach.[100] This perspective sees science as being like other human investigations employing a variety of heuristics that, while not guaranteeing success, is the best we can do. A reductionist, mathematical deductive heuristic is only one possible approach to such modelling. While I am not rejecting the mechanistic modelling of neoclassical economics in its entirety, I am cautioning that it provides only one limited perspective, which could well be mistaken, and there might be more fruitful metaphors. Importantly, it is up to the advocates of reductionism and mathematical deduction to demonstrate its usefulness—particularly as a policy tool—to a justifiably sceptical audience. Furthermore, rather than clinging stubbornly to physics envy and the illusion of certainty provided by what appears to be a degenerating research strategy, economists should learn to embrace pluralism for the richness of the insights it can provide. In particular, economists should open themselves more fully to the possibility of explanation at various levels of organisational complexity throughout the economic system and not stick stubbornly to a reductionist story.

It is within such a framework that American anthropologist Barbara Frankel—drawing on Bateson and Mead—suggests that it might not be forces and objects that are central to human action but rather the information and messages that define the social context and order behaviour within those contexts.[101] She argues, in particular, that there is a danger of confusing biological individuals with social persons—leading to an inability to deal conceptually with contexts and meanings—as opposed to objects and forces. She suggests, therefore, that it might be more appropriate to consider the selves studied by the social disciplines as the sum of an individual’s achieved and ascribed social roles, as nodes in a network of communications, avoiding the distraction of biological boundaries. Consequently, she suggests that we need to take seriously ‘the notion of social persons created by and existing only within systems of interaction, and as bounded, not by the skins of biological individuals, but by contextual boundaries that may be…of indefinite extent’.[102]

One consequence of the positivist approach to the social sciences and the associated attempt to appropriate the prestige associated with the natural sciences has been to suppress political and moral discourse, to confer a privileged position on the status quo and on the professional expert with a capacity for judgement based on the unsustainable claim to technical expertise, neutrality and impartiality. All of this should lead us to be wary—as leading American legal theorist Grant Gilmore (1910–82) advises—of abstract and impersonal values, of universal solutions and of logical imperatives[103] within economics, the law and social life more generally. We should also be wary of grand theory, sacred rules and mystical absolutes that have little connection to reality—especially since we have been taught to be wary of such claims in our spiritual life. As legal historian Morton Horwitz confirms, the belief in the explanatory possibility of general laws capable of making predictive statements in the social sciences has plummeted: ‘The result has been a dramatic turn towards highly specific “thick description” in which narrative and stories purport to substitute for traditional general theories…a complex, multi-factored interdependent world has lost confidence in single-factor “chains of causation” that were embedded in most nineteenth-century explanatory theories.’[104]

In this spirit, English economist Edward Fullbrook recently argued for pluralism among the knowledge narratives with which we organised and interpreted experience, each of which would offer a different view of the object of inquiry.[105] This is because all representations—even the most sophisticated and comprehensive of scientific narratives—involve a radical, stylised and somewhat arbitrary simplification of reality, a choice among an infinite number of possible perspectives or conceptual frameworks. Such a choice rests ultimately on the explanatory usefulness of the narrative and the entities it connects. Fullbrook cites American-born quantum physicist David Bohm in support:

What is called for is not an integration of thought, or a kind of imposed unity, for any such imposed view would itself be merely another fragment. Rather, all our different ways of thinking are to be considered as different ways of looking at the one reality, each with some domain in which it is clear and adequate. One may indeed compare a theory to a particular view of some object. Each view gives an appearance of the object in some aspect. The whole object is not perceived in any one view but, rather, it is grasped only implicitly as that single reality which is shown in all these views.[106]

Any such perspective brings with it a system for classifying the empirical domain, which in turn limits possible descriptions, possible facts, possible questions and possible stories and thus uniquely circumscribes our possible understanding of reality. In particular, the meaning of any concept depends on the framework within which it appears. Viewing a domain from a new perspective brings with it the possibility of new dimensions of understanding. Fullbrook goes on to distinguish between closed narratives such as those of Newtonian mechanics and neoclassical economics and open narratives such as those of evolution, which admit indeterminacy arising from chance, contingency, choice, uncertainty, randomness and spontaneity. In the process, he challenges the hegemony of such closed narratives and the hostility they exhibit to ‘alien’ and open narratives. Rather, he argues that a plurality of narratives enriches our understanding and is essential to the advancement of knowledge.

It might be thought that a rejection of the search for general criteria for judging theories poses problems when it comes to judging economic theories.[107] The undermining of the pretensions of science should not, however, detract us from the task of reasonable judgement in research. It seems that we just have to learn to live with the understanding that all knowledge is a social and linguistic construct, and that this applies with particular force to the social disciplines. Recognition of these difficulties does not justify the proposition that empirical tests are unnecessary. The fact that we are unable to guarantee the truth of a proposition—fulfilling utopian demands of the rationalists—does not absolve us from attempting to develop the best methods we can, even in the absence of an absolute criteria for ‘best’. In particular, it provides no excuse for a failure to take falsification seriously or to subject our theoretical speculative narratives to serious examination. On the contrary, it should provide a good reason to take these tasks and narrative pluralism much more seriously. The awareness of the limitations of one’s tools and how best to use them does not provide an excuse for using them badly—or not using them at all—but rather points to the need to develop the ability to employ them skilfully and honestly. Nor does it license a sloppy use of statistical inference within economic research—a practice that McCloskey has documented.

Of course, it is a standard critique of neoclassical economics that it has abandoned realistic assumptions and has insulated its core beliefs from empirical testing, and does not meet the canons of any reasonable methodology. We will discuss these implications for economics in greater detail in Chapter 8.