Social Limits to Economic Theory

Jonathan D Mulberg

• Publisher: Routledge
• Year: 2005
• Print ISBN: 9780415092982
• eBook ISBN: 9781134804894
• Edition: 1
• Topic: Business
• Subtopic: Business General
• Index: Business

Modern economics makes much of its claim to be impartial, objective and value-free but it is unable to address our most immediate problems such as widespread environmental degradation and persistent poverty. In Social Limits to Economic Theory Jon Mulberg argues for a new progressive political economy, based on notions of community and justice and incorporating environmental and ethical considerations. In doing so he provides the best introduction to date to critical, non-orthodox economics.

1
THE POLITICS OF POSITIVE ECONOMICS

POSITIVE SCIENCE

One of the greatest misnomers in the English language must surely be the term ‘natural science’. It is hard to imagine anything less natural than that which passes for present-day science. To call the study of nuclear fission or chemical weaponry ‘natural science’ is to virtually negate the meaning of the word ‘natural’.

Neither should it be imagined that the term is appropriate because science studies nature. For the most part it does nothing of the sort. In fact, modern science usually hides itself away from nature in laboratories and attempts to keep nature well away from the scientific experiments. Indeed, much modern science is not very good at either understanding or working with nature. Rather the aim appears to be to try and dominate nature, which then means that nature can be ignored.

This becomes particularly significant when we start to discuss the idea of a social ‘science’, and to debate whether or not the social world can be, investigated in the same manner as the physical world. We shall see that the usual view of social science, that it should employ the same value-free, empirically based techniques that supposedly characterise physical science, is often justified precisely by this appeal to domination and control. However, we will also see that this justification for social science leads to a dilemma, because the concepts of value-neutrality and domination are contradictory.

Before we discuss the question of methodological unity, of whether or not social science can adopt the same methods as physical science, we need to be clear what exactly is understood by the term ‘science’. It is far from obvious what being ‘scientific’ actually involves, and we will begin by outlining different conceptions of what constitutes ‘science’, and explain why the popular, classic conception of science is inadequate. Although inadequate, the classic (positivist) view of science is nonetheless still defended by economists, and the second section of this chapter will discuss whether it is possible to square the discipline of economics with the maxims of positivist science. We will see that a value-free social science cannot be sustained, and that such a science has an inbuilt political theory. The third section will discuss the political theory of positivism, and the consequences of this theory for economics.

A useful starting point for our discussion is Keat and Urry's tripartite classification of the conceptions of science. This classification will come under strain later, but will serve our purposes for now. The authors outline three positions—positivism, realism and conventionalism (Keat and Urry 1975 p. 4). Positivism they classify as the belief that science consists of the attempt to obtain ‘predictive and explanatory knowledge’ of the world as it exists. This is achieved by theorising the existence of regularities, which enable prediction and explanation. These theories are objectively tested by observations and experiments (ibid.).

Realists also believe in science as empirical and objective, but posit a difference between explanation and prediction, the former being primary. This involves going ‘beyond’ regularities to the causal mechanisms involved, which may well necessitate positing ‘unobservable entities' (ibid. p. 5).

Conventionalists reject the concept of empirical and objective science (ibid.). As a result, the goals of science become a matter of debate, and conventionalists tend not to share common views on what scientific knowledge consists of.

We can see that a belief in objectivity in science is therefore insufficient to distinguish positive science. In addition, the positivist will put forward claims about the content of explanation and its relation to prediction, the process of observation and the role of theory. Fay cites four essential features of modern positivism: the limited account of explanation, the foundation of knowledge in the language of neutral observation, the ideal of value-free science and the assertion of the methodological unity of all sciences (Fay 1975 p. 13). Our interest for the first part of this chapter will be on the first two of these features.

The positivist view of explanation is that it is limited to the discovery of regularities: to explain something is simply to show that it is the result of a regular pattern. We do not attempt to go ‘behind’ the empirical regularities in order to discover the ‘essence’ of phenomena (Keat and Urry 1975 p. 4). For the positivist, explanation and prediction are congruent: explanations consist only of data which lead to correct predictions, and predictions are the only way of confirming the validity of explanations. In addition, the second feature of positivism is that the explanations and predictions are the result of empirical observation, and that this observation is neutral. Only what can be observed is to be regarded as scientific, and non-observable entities do not form any part of the brief of the scientist.

Before we consider this view of science, it might be as well to enquire as to the status of the philosophies of science themselves. Positivists suggest that science consists of, among other things, explanatory and predictive knowledge of the world. Is this suggestion itself explanatory and predictive? It would not appear to be possible to give a positivist account of what is actually a criterion for science—to attempt to do so is to fall into an infinite regress. Modern positivism has moved away from a theory of scientific meaning towards what Kolakowski (1972) calls a ‘normative attitude’ to knowledge, towards the postulation of criteria for the demarcation of science from nonscience. But if the positivist philosophy consists in fact of normative criteria, these criteria have to be justified normatively. Why have positive science at all? What does science achieve that religion, say, does not?

Habermas suggests that all knowledge has ‘practical intent’, and serves ‘social practice’ in the reproduction of the species. He identifies two areas of social practice: labour, where humans interact with nature in the process of production, and interaction, where social coordination occurs. For Habermas both of these are necessary; all production and existence is social. Each category of social practice has a distinct type of knowledge associated with it. Interaction requires knowledge of appropriate social norms, whereas labour requires ‘purposive-rational’ knowledge (Habermas 1969 pp. 93 ff.) We could therefore suggest that knowledge for labour is of an instrumental nature. It is designed for control over our external environment.

What is occurring in modern capitalist societies, what Habermas refers to as the ‘discourse of modernity’, is that the latter form of instrumental knowledge is dominating the former, so that social norms are increasingly decided by instrumental rationality. Habermas refers to this as the ‘scientization of politics’ (ibid. ch. 6). We will return to this later in the chapter.

The notion of positive science being justified instrumentally appears uncontentious. Positive science is desirable because it gives us knowledge which enables us to improve our lot. This is the point of the exercise. The question is whether this knowledge can be extended to social science. Fay believes that it would be the claim of many social scientists that

just as the natural sciences have provided men with a certain kind of knowledge by which they can control their natural environment, thereby making it more hospitable and productive, so also the knowledge gained from social science will enable men to control their social environment, thereby making it more harmonious and congruent with the needs and wants of its members.

(Fay 1975 p. 14)

Social science is therefore required in order to obtain instruments for social improvement. The positivists' claim is that two elements within positive science make it uniquely placed to provide these instruments. Firstly, they claim that the value-neutrality of positivism means that only positive science can provide objective knowledge of society, and that this is superior to subjective viewpoints. They also claim that the ability to predict gives us instrumental knowledge which would otherwise be lacking (ibid. p. 21).

The normative justification for the demarcation of science from non science along positive lines is that it gives us control over our environment. It is able to do this because it gives us the ability both to predict the future and to understand the present in an objective manner.

We will have more to say on the emphasis on objectivity later in the chapter. Our interest for the present is in the idea of ‘knowledge’. We suggest that the positivist insistence on knowledge as being exclusively based on observation places severe restrictions on the ability to understand and consequently to predict. The justification for positivism in instrumental terms (in terms of instruments for control) tended to suggest that the science would in some manner enhance our predictive ability beyond that of mere chance, which would usually be taken to mean that science could offer us causal knowledge. We would justify a prediction by explaining the causes of the phenomena under investigation.

This requirement of causal explanation creates considerable friction within the positivist philosophy, given the insistence on a limited account of explanation and of observation being the foundation of knowledge. Although the idea of causality is now accepted as implied by explanation (Hempel 1966 p. 52–3), many of what we would usually regard as explanations involve nonobservational entities—viruses, magnetism, electrical resistance, friction or even (from economics) satisfaction. That is to say, we require theories and theoretical terms to develop causal analysis. Of course, the introduction of theories marks a significant weakening of the positivist position, and it is far from clear how they should be treated. Mach believed that positive science should have no interest in theory—theories may be useful heuristic devices for the practising scientist, but are of no consequence for the science itself (Caldwell 1982 p. 23). That is, we should restrict ourselves to the discovery of regularities. This appears to make non-trivial science next-to-impossible though, since we would have no method of deciding which empirical regularities to consider, and which experiments to perform. We also need to be able to interpret our observations; the regularities we observe will require theories to explain them (Hempel 1966 p. 70).

The modern positivists usually accept the centrality of explanation: ‘To explain the phenomena of the physical world is one of the primary objectives of the natural sciences' (ibid. p. 47). In fact, one of the central claims of positivists is the symmetry between explanation and prediction (Keat and Urry 1975 p. 11, Caldwell 1982 p. 29). This is because of the form of explanation now adopted, known as the ‘covering law’ model. Covering laws are conditional statements; given the antecedent condition, we can conclude the explanandum. The explanandum can be predicted given knowledge of the existence of the antecedent conditions, and the covering law can be used as an a posteriori explanation of the explanandum (Keat and Urry 1975 p. 11). The explanations can therefore be tested by their predictions. Hempel refers to this form of explanation as ‘deductive-nomological’ (Benton 1977 p. 55; Fay 1975 p. 32). Nonetheless, some form of theoretical analysis is still required to enable consideration of the covering laws, since we have to distinguish our causal explanations from accidental generalisations.¹ To do this we would require in particular counter-factual conditionals (Hempel 1966 p. 56).² But counter-factual conditionals can seldom be observed without controlled experiment, which is not usually possible in social science.

Perhaps we can get round the problem of conditional statements by placing less emphasis on deductive explanation. The other mode of explanation frequently discussed Hempel calls ‘inductive-statistical’. We can perhaps explain an event inductively by reference to its frequency of past occurrences. We might explain, for example, the fact that inflation rose as unemployment fell by the fact that this has occurred in, say, 90 per cent of all past cases.

It is important to note that this mode of explanation is not a logical probability, which is a question of the confirmation of an hypothesis, but is a statistical probability, a ‘quantitative relation between repeatable kinds of events: a certain kind of outcome…and a certain kind of random process’ (Hempel 1966 p. 63 emphasis on original).

Given this definition, we may question whether a statistical ‘explanation’ is possible (Keat and Urry 1975 p. 13). Indeed, in the above quotation from Hempel the process is described as random. A poker hand which I deal to myself is not ‘explained’ by its probability; if I deal four aces I would not normally be inclined to ‘explain’ this event to the other players. I was lucky, that is all there is to it. In this sense, explanation is precluded.³ What can perhaps be explained is the pattern of repeatable events.⁴ Of course, as the analysis becomes exhaustive, the hypotheses become more like a covering law model: the overall probabilities tend to unity, and the two types of probability outlined by Hempel tend to merge.

The inductive model cannot, then, offer explanations of events. It can enable us to make predictions on the basis of regularities though. This would therefore appear to give two choices: to either break with the requirement of explanation or to adopt a weak version of deductive explanation involving theory. Theories are usually regarded by positivists as containing two types of principles: their own internal tautologies and ‘bridge principles' (sometimes called ‘correspondence rules'). The bridge principles link the theoretical terms to observable entities. It is this that gives a theory its ‘explanatory power’ and its testability (Hempel 1966 p. 74). This is a fairly significant departure from the concept of testability therefore; the theory as a whole is never tested. The only test is of the implications of the theory in terms already (antecendently) available. This makes testing problematic, since we must assume the correctness of the correspondence rules. The explanation as a whole is never tested against observation.

It would appear that the positivist philosophy as a whole cannot be sustained. We must either accept theoretical concepts which violate the principle of observation-based knowledge or abandon the congruency of explanation and prediction.

We have already mentioned the idea of inductive/statistical reasoning that does not offer explanation. In fact we can go either way—predictions that do not explain or explanations that do not predict. Of the latter, the most common form are those known as ‘functional’ explanations, which broadly consist of explaining processes in terms of the function played in achieving an end-state. We shall see how there is a tendency for social explanation to adopt this form. The main question in an examination of functional explanations is the domain of the theory. As special theories these forms of explanation may have some legitimacy, but as general theories, and as social theories, counter-factual conditionals cannot be supported, and functional explanations of this level tend towards bland assertions of the inevitability of the status quo. The whole idea of control, which was one of the reasons for our interest in positivism, is lost when one adopts evolutionary/ functional analysis—indeed, there is a strong tendency towards deterministic theories. We will come across this mode of reasoning later, in the works of Veblen and of the Austrians.

The alternative route is to view theories as simply instruments to obtain predictions: we are no longer concerned...

Table of contents

• Cover
• Half Title
• Full Title
• Copyright
• Dedication
• Contents
• Preface and acknowledgements
• INTRODUCTION AND OVERVIEW
• 1 THE POLITICS OF POSITIVE ECONOMICS
• 2 FROM UTILITY TOW ELFARE THE TRAJECTORY OF ORTHODOX ECONOMICS
• 3 1930s' MARKET SOCIALISM
• 4 AMERICAN INSTITUTIONALISM
• 5 NEW INSTITUTIONALISM
• 6 NEW SOCIAL MOVEMENTS
• Notes
• References
• Index