ECONOMICS AND REALITY
Contemporary academic economics, a project often acclaimed as the most scientific of the social ...
692 downloads
3774 Views
877KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
ECONOMICS AND REALITY
Contemporary academic economics, a project often acclaimed as the most scientific of the social sciences, is in a state of some confusion. Its ‘theoretical models’ are increasingly found to be irrelevant to real world matters, its empirically based forecasting models do not forecast particularly well, and those who provide commentaries on the discipline report a significant disparity between the practices of economists and the ‘methodologies’ they profess to draw upon. Many who reflect seriously on the situation conclude that, despite widespread pretensions to the contrary, there are few grounds for supposing that economics can be a science at all. In an innovative and wide-ranging contribution, Tony Lawson concludes that at the root of these problems is a failure of economists to develop methods that are appropriate to their subject matter. In taking steps to compensate, the author demonstrates that a radical transformation in the basic practices and overall direction of the discipline is warranted. In making his case, however, Tony Lawson parts company with many recent social theorists in contending that an efficacious, and indeed emancipatory, science of economics is entirely feasible. In fact, he concludes that economics can be a science in precisely the sense of natural science. However, he establishes that for this possibility to be realised economists must, at a minimum, show a far greater sensitivity to context and adopt significantly broader, more critical, orientations to theorising and explanation than are currently in evidence. Although informed by philosophy, Economics and Reality is rooted throughout in the problems, practices and writings of economists. Exploring the current state, possibilities and limits of economics, the book will be of interest to researchers, teachers and students of economics alike. Tony Lawson is a member of the Faculty of Economics and Politics at the University of Cambridge. He has published extensively in books and journals including The Review of Economic Studies, the Cambridge Journal of Economics, Oxford Economic Papers and the Economic Journal.
© 1997 Tony Lawson
ECONOMICS AS SOCIAL THEORY Series edited by Tony Lawson University of Cambridge
Social theory is experiencing something of a revival within economics. Critical analyses of the particular nature of the subject matter of social studies and of the types of method, categories and modes of explanation that can legitimately be endorsed for the scientific study of social objects, are re-emerging. Economists are again addressing such issues as the relationship between agency and structure, between the economy and the rest of society, and between inquirer and the object of inquiry. There is renewed interest in elaborating basic categories such as causation, competition, culture, discrimination, evolution, money, need, order, organisation, power, probability, process, rationality, technology, time, truth, uncertainty and value, etc. The objective for this series is to facilitate this revival further. In contemporary economics the label ‘theory’ has been appropriated by a group that confines itself to largely a-social, a-historical, mathematical ‘modelling’. Economics as Social Theory thus reclaims the ‘theory’ label, offering a platform for alternative, rigorous, but broader and more critical conceptions of theorising. Other titles in this series include: ECONOMICS AND LANGUAGE Edited by Willie Henderson RATIONALITY, INSTITUTIONS AND ECONOMIC METHODOLOGY Edited by Uskali Mäki, Bo Gustafsson and Christian Knudson NEW DIRECTIONS IN ECONOMIC METHODOLOGY Edited by Roger Backhouse WHO PAYS FOR THE KIDS? Nancy Folbre RULES AND CHOICE IN ECONOMICS Viktor Vanberg BEYOND RHETORIC AND REALISM IN ECONOMICS Thomas A.Boylan and Paschal F.O’Gorman FEMINISM, OBJECTIVITY AND ECONOMICS Julie A.Nelson ECONOMIC EVOLUTION Jack J.Vromen © 1997 Tony Lawson
ECONOMICS AND REALITY
Tony Lawson
London and New York © 1997 Tony Lawson
First published 1997 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 This edition published in the Taylor and Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” © 1997 Tony Lawson All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book is available from the Library of Congress ISBN 0-203-19539-6 Master e-book ISBN
ISBN 0-203-19542-6 (Adobe eReader Format) ISBN 0-415-15420-0 (hbk) ISBN 0-415-15421-9 (pbk)
© 1997 Tony Lawson
For my parents
© 1997 Tony Lawson
CONTENTS
Preface and acknowledgements Part I Science and economics 1
ENDURING TENSIONS AS POINTS OF DEPARTURE Inconsistency at the level of method Inconsistency at the level of social theory Inconsistency at the level of methodology
2
REALISM, EXPLANATION AND SCIENCE Realism Deductivism Empirical realism Transcendental realism Science, inference and law-statements Knowledge as a produced means of production
3
THE CASE FOR TRANSCENDENTAL REALISM Natural science: Bhaskar’s inference from experiments Social science: inference from human intentionality Ontology and the epistemic fallacy Towards a distinctive science of society and economy
4
THE LEGACY OF POSITIVISM The failure of economics at the level of results Confusion at the level of method Problems in social theory A contradictory orientation to methodology
5
THE NATURE OF THE ARGUMENT The case against normative methodology A question of science Realism and science Transcendental analysis
© 1997 Tony Lawson
Explaining experimental activity The possibility of economics as science Human intentional agency Limits and limitations 6
TOWARDS A RICHER ONTOLOGY A stratified reality Negativity, intentional causality, emergence and internalrelationality Part II Problems of contemporary economics
7
ECONOMETRICS An absence of economic ‘laws’ The ‘Lucas critique’ Reactions to the ‘Lucas critique’ Systematic responses to generalised predictive failure Regularity stochasticism Extrinsic closure Intrinsic closure Aggregation Reinterpreting the ‘Lucas critique’ The fallacies of atomism and isolationism
8
ECONOMIC ‘THEORY’ Nominal features of the orthodox ‘pure theory’ project Changing features Explanatory failure Economic ‘theory’ and deductivism Generalist claims and the question of determinacy A deductivist tradition Regularity determinism Intrinsic closure Extrinsic closure Reinterpreting the ‘theory’ project Solution concepts Economic ‘theory’ as deductivism Competing interpretations Potentials and actualities Prospects
9
ALTERNATIVES AND/OR PRELIMINARIES TO SUCCESSFUL EXPLANATION Theorising as an alternative to explanation Theory as one-sided understanding
© 1997 Tony Lawson
Menger’s defence of ‘theory’ Theoretical science and exact laws Exact laws of theoretical economics The nature of Menger’s argument Unanswered questions Menger’s account of natural science Menger’s inferences for social science Theorising as a preliminary to successful explanation The method of successive approximation The method of isolation 10
SUBJECTIVISM Hayek’s theory of knowledge and natural science Hayek’s theory of knowledge and social science Social structure Hayek’s hermeneuticised social science Social ‘wholes’ Foundationalism and Hayek’s hermeneuticism Hayek’s atomism The material embeddedness and intransitivity of society
11
THE LIMITS OF CONTEMPORARY ECONOMICS Consequences of a misconception of science A typology of responses to failure Part III Human agency and society
12
SOCIETY AND ECONOMY AS REPRODUCED INTER-DEPENDENCIES Critical realism Abandoning social atomism The routinisation of social life The irreducibility of social structure Social rules Social relations and societal practices Systems and collectivities The agency structure relationship Social change The relationship of social rules to positions
13
A SKETCH OF THE ACTING SUBJECT Intentionality Irreducibility and the phenomenon of emergence Tacit knowledge Unconsciousness
© 1997 Tony Lawson
Sameness and continuity Stability in the face of uncertainty Individuality Economic rationality Situated rationality Part IV Economic explanation 14
BROAD OBJECTIVES AND POSSIBLE OBSTACLES Broad aims The centrality of human practice Intrinsic limits to social science
15
ECONOMIC SCIENCE WITHOUT EXPERIMENTATION The hermeneutic moment in science Experimental and non-experimental conditions contrasted Interpreting partial event regularities Explaining the preponderance of demi-regs Contrastive demi-regs Contrastive demi-regs and science The detection of interesting demi-regs Science and scientific interests Inconsistency, surprise and criticism Causal hypotheses Explanatory power Problems in discriminating between theories Responses to explanatory failure Assessing the reality of a hypothesised mechanism The requirements of orthodox economics Pure and applied explanation Economics as an empirical and abstract science The apparent failures of social research including economics
16
ABSTRACTION Abstraction and critical realism The vantage point The level of generality The scope or extension Abstraction and generalisation Abstraction and economic ‘modelling’
17
ON TRUTH IN ECONOMICS The nature of truth Objective truth
© 1997 Tony Lawson
Truth as an expressive-referential dual Criteria of truth Judgemental rationality Falsity and sheer error Truth and economic ‘modelling’ 18
ILLUSTRATION Path-dependence Functionalism and mainstream theory The economy as an open system Britain’s relative productivity performance as a contrastive demi-reg The nature of the explanation Explaining the explanation ‘Origins’ of the explanation Tendencies towards ‘lock-in’ Social reproduction/transformation as an ex posteriori contrastive phenomenon Conditions for social reproduction Relative continuity as an ex posteriori production/achievement The explanation in broader perspective The partiality of all explanation Part V Economic policy and forecasting
19
ECONOMIC POLICY AND INTENDED CHANGE Orthodoxy, economic policy and change Critical realism, economic policy and change Conditions of social emancipation
20
ECONOMIC SCIENCE AND PREDICTION The basic thesis Omissions Consequences The preoccupation with prediction Notes Bibliography
© 1997 Tony Lawson
PREFACE AND ACKNOWLEDGEMENTS
‘No reality, please. We’re economists!’ This heading to a couple of commentaries in a recent edition of The Times Higher Education Supplement (25 March 1994) captures an increasingly widely held perception. According to it, modern economics is largely irrelevant to an understanding of real world matters. Needless to say this view is not accepted by all economists. And it does seem prima facie implausible as a characterisation of an academic discipline that is often represented as a social science. Yet, my assessment set out below is that it is not wholly mistaken; there is a sense in which social reality is neglected in much of modern economics. I shall indicate furthermore how this neglect is debilitating of the subject, with unfortunate repercussions for the multitude of further activities that draw upon it. My purpose here, however, is not merely to identify limitations of the discipline but also to work towards transcending them. My overall aim, if somewhat unfashionably, is to bring reality (or more of it) back into economics. Of course, in formulating my goal in this way, and in acknowledging some validity to the perception of the discipline noted above, I am not wanting to suggest that modern economics is devoid of a conception of reality altogether. Indeed, it could not be. For the adoption of any method, practice or goal, presupposes a world-view, even if it is left implicit and unexamined. The feature of modern economics which is remarkable here is the minimal concern that is shown either for tailoring methods to insights available regarding the nature of the social world, or for explicitly determining the sorts of conditions under which chosen methods would be appropriate. This is despite a dearth of successes according to criteria which the discipline sets itself. Rather, methods and procedures are formulated according to the nature or degree of their technical sophistication, or their conformity with a priori conceptions of proper practice, or some such. The majority of economists seem merely to take it for granted that if ever real world problems or situations are to be studied the chosen methods will necessarily ‘fit’. It is in this sense that reality is neglected in modern economics. And we will see that it is just the presumption that methods of economic analysis can be fashioned © 1997 Tony Lawson
without explicit regard to the nature of social phenomena that accounts for the subject’s numerous failings. In short, contemporary economics is marked by an effective neglect of ontology, by a lack of attention to elaborating the nature of (social) being or existence. And the project reported here is motivated by a desire to compensate for this neglect. The objective is to determine a sustainable account of natural and social being with a view both to explaining and resolving numerous problems which beset the discipline and to gaining an informed vantage-point on the sort of practices that are likely to bear fruit. Theories which address these sorts of concerns and involve commitments to the reality of features elaborated are usually collected under the heading of philosophical (or ontological) realism. The position sustained in this book, then, is an explicitly realist one. The book constitutes a realist theory of and for economics. Its context can be briefly sketched. In contrast to much contemporary work on economic methodology, the project reported here did not emerge from a formal study of philosophy. Having come to economics by way of first studying mathematics I was immediately impressed by, as I saw it, the widespread and rather uncritical application of formalistic methods and systems to conditions for which they were obviously quite unsuited. In consequence, my interests turned fairly quickly to questions of ontology, and specifically to the study of how methods and modes of reasoning might be fashioned to insights concerning the nature of social being. However, I first approached these issues by way of reading economists rather than philosophers. Given my immediate conviction that the conventional economic methods and procedures are rather obviously inappropriate to the successful investigation of social reality, I assumed that many other economists would have reached the same conclusion long before. I thus supposed they would already have investigated and responded to the situation. In any case, it was an understanding of the economics context specifically that was of primary interest to me. It turned out that it is mostly (but not exclusively) economists regarded as non-mainstream who have reflected on these matters to any significant extent. I refer here to the likes of Keynes, Hayek, Marx, Dobb, Veblen, Marshall, Smith, Shackle, Menger, Boulding, and Kaldor. In consequence, it is their writings that provided much of the initial background material for the central argument set out below. In fact, they contribute rather more than this. In one form or another, they already express many of the fundamental tenets of the basic thesis. For example, when Keynes writes that: If we were dealing wit…independent atomic factors and between them completely comprehensive, acting with fluctuating relative strength on © 1997 Tony Lawson
material constant and homogeneous through time, we might be able to use the method of multiple correlation with some confidence for disentangling the laws of their action. (Keynes, 1973b:286) and Marx observes that individuals: make their own history, but they do not make it just as they please; they do not make it under circumstances chosen by themselves, but under circumstances directly encountered, given, and transmitted from the past. (Marx, 1977:300) and Veblen suggests: The growth and mutations of the institutional fabric are an outcome of the conduct of the individual members of the group, since it is out of the experience of the individuals, through the habituation of individuals, that institutions arise; and it is in this same experience that these same institutions act to direct and define the aims and end of conduct. (Veblen, 1909:243) and Hayek adds that: different rules apply to different individuals according to age, sex, status, or some particular state in which each individual finds itself at the moment. (Hayek, 1967:68) each is expressing in condensed form an important aspect of the thesis which follows. But what a lot of fleshing out and unpacking each assessment requires. When I eventually turned to formal philosophy, I naturally sought out the writings of philosophers who have been especially concerned with ontological or metaphysical questions. The influence of this literature on what follows appears to distinguish further my project from most other contributions in economic methodology. For although the latter contributions tend to be heavily rooted in the philosophy of science literature, it is questions of epistemic appraisal (i.e. epistemological questions concerning the rational basis for accepting or rejecting theories) that have dominated the economic methodology discussion. This discussion draws primarily upon the philosophies of Popper, and of others like Lakatos who responded to Popper’s work (see contributions to Backhouse, 1994, and for a useful overview see Boylan and O’Gorman, 1995). Mäki (1993a:5) sums up this situation by © 1997 Tony Lawson
suggesting that ‘A Popperian dominance, a kind of Popperian mainstream in economic methodology has prevailed’. It is not clear to me that this emphasis upon epistemological matters, often more or less detached from explicit ontological theorising, has been overly fruitful, indeed, I feel inclined to support Mäki’s assessment ‘that the Popperian dominance has led to a misallocation of intellectual resources in economic methodology’ resulting in a neglect of, amongst other things, a questioning of ‘the nature of metaphysical commitments involved in actual research practice’ (1993a: 6). This book, then, can also be seen as an attempt to redirect the methodological debate in economics. Even so, any philosophical position, including an ontologically oriented one such as defended below, ultimately bears epistemological implications. In consequence, some readers may still feel some surprise to learn that Popper and/or Popperian (epistemological) themes and categories figure relatively little—significantly less than in comparable texts concerned with economic methodology. This ‘omission’, moreover, may appear especially surprising in the light of the fact that Popper openly proclaims himself a realist, falibilist, and anti-positivist, and recognises the possibility of demarcating science from non-science—where all these are positions openly accepted in what follows. The basic reason for it is that I not only adopt a more explicit focus upon ontological issues but also defend a theory of ontology which is very different to that presupposed by Popper, at least in his most familiar writings on falsificationism (as opposed to his relatively recent (1990) rather general remarks upon propensities). In consequence, the epistemological questions and implications are correspondingly different. In parallel with Popper I am certainly concerned to criticise various results of the (untenable) philosophical perspective of positivism. But although Popper (1992) claims responsibility for ‘killing’ positivism, or at least the logical positivism of the Vienna circle, there never was a single fatal blow. Positivism, including logical positivism, rests upon at least two fundamental assumptions. Not only does it suppose that science is essentially monistic in its development, it also assumes that it is deductive in structure. If Popper’s contributions to the theory of knowledge, and in particular his work on falsificationism or ‘deductivism’ (Popper, 1992:81), have helped undermine the former aspect, these same writings (at least as most have interpreted them) leave the assumed deductivist structure untouched. And it is the presumption of a deductivist structure of science, including its implicit account of being, or ontology, that I am concerned here to criticise before anything else. I should immediately add, however, that if the falsificationist or deductivist programme usually associated with Popper is rejected here as a generalised way of proceeding, Popper’s explicit commitment, in that programme, to the pursuit of objective truth and rationality in science is endorsed. Indeed, at a time when so many economists appear to do little more than flit from fashion © 1997 Tony Lawson
to fashion; when, with one or two notable exceptions (e.g. Mäki, 1988, 1992a, 1992b, 1994), economists rarely even mention (let alone accept) the goal of pursuing truth; and when, despite everything, many mainstream economists persist in claiming the authority of Popper for what they do; it seems worthwhile restating Popper’s views on these matters explicitly: I guess, indeed, that it is the suppressed sense of our own fallibility that is responsible for our despicable tendency to form cliques and to go along with whatever seems to be fashionable: that makes so many of us howl with the wolves. All this is human weakness, which means it ought not to exist. But it does exist, of course; it is even to be found among some scientists. And as it exists, we ought to combat it; first in ourselves, and then, perhaps, in others. For I hold that science ought to strive for objective truth, for truth that depends only on the facts; on truth that is above human authority and above arbitration, and certainly above scientific fashions. Some sociologists fail to understand that this objectivity is a possibility towards which science (and therefore scientists) should aim. Yet science has aimed at truth for at least 2,500 years. (Popper, 1990:34) There are a number of philosophers (Harré, Hesse and Toulmin) who have focused their criticism on the supposition that science is deductive in structure, and whose writings have had an influence on what follows, just as there are philosophers other than Popper (Bachelard, Lakatos, Kuhn and Feyerabend) who have attacked the positivistic view of scientific development as monistic. But in my view it is Bhaskar (1978, 1979, 1986, 1989) who has managed most successfully to combine the insights of both tendencies in constructing a sustainable ontology for science, and whose project has in consequence proved most useful for my own. Although I only came across it in the late 1980s it has had a very significant impact upon what follows. Nevertheless, I should stress that even Bhaskar’s work is not explicitly surveyed here. Influential though Bhaskar’s writing has been, the present book does not constitute an introduction to it, or anything like it; for that the reader can consult not only Bhaskar’s own contributions but also Andrew Collier’s (1994) useful introduction and appraisal. In structure and content the current book is shaped by the situation in economics. As Bhaskar somewhere explicitly acknowledges, his ideas and terminology are not consistent in the sense of being fixed over time. Rather they reveal a developmental consistency: as new problems are resolved, wider concerns are taken on board and earlier conclusions and terminology are revised accordingly. In intent, at
© 1997 Tony Lawson
least, any consistency to be found below with Bhaskar’s contribution is also of this developmental sort. This book, then, is about economics, its current state and real possibilities and limits. Although the thesis which follows may be interpreted as broadly philosophical, it takes its premises from, and so is firmly rooted in, problems, practices and writings of economists. The intended readership is, above all, economists, especially those prepared at least to question received doctrines. The central argument which follows does have a wider relevance, but the book was written with the economics community in mind. It is the product of a continuing dissatisfaction with the dominant orientation of the economics discipline combined with a conviction that it is possible to do significantly better. A brief overview of the argument can be provided by summarising each part of the book. Part I In Chapter 1 I initiate the discussion through drawing attention to various problems and tensions which pervade economics. In Chapter 2 the widespread acceptance of a particular ‘deductivist’ conception of science and explanation ultimately rooted in, or at least parasitic upon, positivist results is indicated, and an alternative ‘transcendental realist’ conception is outlined. The relative superiority of the latter over the former is demonstrated in Chapter 3. In Chapter 4 the problems and tensions of contemporary economics which are noted and described in Chapter 1 are shown to derive from the widespread reliance upon the deductivist conception of science and explanation, while the superior transcendental realist account is found to be free of such problems. In Chapter 5 the reasoning used to establish these results is examined in some detail. Further implications of the perspective supported are elaborated in Chapter 6. Part II With the central argument broadly laid out in Part I, the object of Part II is to examine in greater detail the various ways in which mainstream economics turns on the deductivist conception of science and explanation and to indicate the sorts of problems that arise in each case. The focus is upon econometrics in Chapter 7; upon ‘economic theory’ in Chapter 8; and, in Chapter 9, upon those variations on the latter which attempt to play down the need to explain (yet) any actual phenomena. In Chapter 10 I show that the lingering effects of positivism even shape Hayek’s early, and currently influential, attempt to develop a ‘subjectivist’ alternative to the mainstream approach. The essential characteristics of the various projects considered in Chapters 7 to 10 are summarised in Chapter 11, along with conceivable alternatives which also fail to abandon positivist restrictions.
© 1997 Tony Lawson
Part III A basic position sustained in Part I is that there are benefits to continually modifying methods of analysis in the light of insights gained into the nature of any given object(s) of enquiry. With this in mind, the focus of Part III is upon social being, and specifically upon the nature of human agency, social structure and their irreducible mutual inter-dependency. While Chapter 12 is concerned with social structure Chapter 13 concentrates on human agency, including aspects of subjectivity and the nature of human rationality. Part IV With a conception of social being set out in Parts I and III, the various consequences for economic scientific practice are elaborated in Part IV. In Chapter 14 I draw out implications concerning the basic orientation of economic explanation, and in Chapter 15 I discuss how competing explanatory theories of social phenomena can be assessed. In Chapter 16 I focus upon the much neglected yet fundamental issue of scientific abstraction. The epistemic status of explanatory results is considered in Chapter 17 and the possibility of objective truth is assessed. Chapter 18 constitutes an illustrative example of the explanatory approach defended. Part V The final part of the book rounds off the argument. In Chapter 19 I trace out the implications for questions of policy and intended change and in Chapter 20 I draw together the various threads of the argument by way of examining the consequences for the enterprise of economic prediction. The basic overall conclusion is that while the academic discipline stands in need of radical transformation, involving amongst other things the abandonment of many of its most cherished ways of proceeding, the possibility of realising an efficacious and emancipatory science of economics is far from being out of the question. Such an achievement, though, will inevitably involve a greater sensitivity to context and a rather more reflective and critical orientation than are currently in evidence.
ACKNOWLEDGEMENTS The final task that awaits me at this point is the pleasant one of acknowledging at least some of those who have contributed to the preparation of this book over its numerous stages. First of all I must thank those who have helped me to type it. I am ashamed to admit that it was only very recently that I learnt to use a ‘word processor’. Thus although I actually managed to type recent versions myself, the project has been under way for sufficiently long that various people have typed early drafts of chapters. In particular I thank Louise
© 1997 Tony Lawson
Cross, Nacirah Lutton, Maxine Maresz, Aliye Seif, Jackie Walpole and Ann Widdop for their skilful work, wonderful patience and even encouragement. Next I must thank the many participants of the Cambridge Workshop On Realism And Economics. This workshop has been running for more than five years now (since October 1990) and most of the arguments which follow stem from ideas that have been presented, scrutinised, criticised and transformed at this forum. Although the list of participants is too long to include here I must especially thank Steve Fleetwood, Geoff Harcourt, Geoff Hodgson, Mark Peacock, Steve Pratten and Jochen Runde who, along with Roger Backhouse, Donald Gillies, Wade Hands and various anonymous referees, read and commented upon an early version of the manuscript in its entirety. Geoff Hodgson and Jochen Runde even read and commented upon parts of a revised version as well. I am truly grateful. Various chapters, or papers which formed the bases of certain chapters, were also read and criticised by Bruce Caldwell, Paul David, John Davis, Simon Deakin, Paul Dalziel, Sheila Dow, Johannes Fedderke, Aly Fischer, Prue Kerr, Clive Lawson, Uskali Mäki, Pat Northover, Gabriele Pastrello, Jonathan Perraton, Roy Rotheim, Warren Samuels, Leslie Turano and Arnis Vilks, and I have had occasion to discuss various aspects of the argument with Roy Bhaskar. To all of these I extend my gratitude. Here I also owe a special mention to Evie Sofianou, a PhD student working with me, who during the final stages of a terminal illness still took the time to provide me with her very constructive criticisms of numerous chapters. Words cannot adequately express my feelings about this. Others have helped through their encouragement and general support. Although this is once more a category in which there are far too many people for all to be mentioned, I must single out my family who have put up with the whole project for so long. Here I thank in particular my companion Joëlle Patient who provided not only generalised support and encouragement but also a significant amount of detailed advice and criticism on all aspects of the book. This latter contribution, which far surpassed what might reasonably be expected of a partner, proved invaluable to the book’s development. Chloë and Julien too have given support in their own ways; I am especially indebted to Julien for double-checking the accuracy of quotations used and references made. And I must not forget Heather who not only allowed me some use of the family computer but also, at the age of seven, on noticing the enduring nature of the project with which I was involved, informed me of her readiness to finish it for me when she was ‘grown up’, should the book remain uncompleted. In the face of such selflessness I had no option but to complete the manuscript as quickly as I was able. I gratefully thank all these people and others who have helped me in so many ways.
© 1997 Tony Lawson
Finally, I thank the editors and publishers of the following journals and collections of essays for permission to reproduce here revised versions of material that earlier appeared in their pages: ‘Abstraction, Tendencies and Stylized Facts: A Realist Approach to Economic Analysis’, Cambridge Journal of Economics, 1989, Vol. 13, No. 1, pp. 59–78. Reprinted in T.Lawson, G.Palma and J.Sender (eds), Kaldor’s Political Economy, 1989, London: Academic Press. Also reprinted in P.Ekins and M.Max-Neef (eds), Real Life Economics: Understanding Wealth Creation, London and New York: Routledge, 1992, pp. 21–38. ‘Realism and Hayek: A Case of Continuous Transformation’, in M. Colonna, H.Haggemann and O.F.Hamouda (eds), Capitalism, Socialism and Knowledge: The Economics of F.A.Hayek, Vol. II, 1994, Cheltenham: Edward Elgar. ‘Critical Realism and the Analysis of Choice, Explanation and Change’, Advances in Austrian Economics, 1994, Vol. 1, No. 1, pp. 3–30. ‘The Nature of Post Keynesianism and its Links to other Traditions’, Journal of Post Keynesian Economics, 1994, Vol. 16, No. 4, pp. 503– 538. Reprinted in D.L.Prychitko (ed.), 1996, Why Economists Disagree: An Introduction to the Contemporary Schools of Thought, New York: State University of New York Press. ‘A Realist Theory for Economics’, in R.E.Backhouse (ed.), New Directions in Economic Methodology, 1994, London and New York: Routledge. ‘Why Are So Many Economists So Opposed to Methodology?’, Journal of Economic Methodology, 1994, Vol. 1, No. 1, pp. 105–133. ‘The “Lucas Critique”: A Generalisation’, Cambridge Journal of Economics, 1995, Vol. 19, No. 2, pp. 257–277. ‘A Realist Perspective on Contemporary “Economic Theory”’, Journal of Economic Issues, 1995, Vol. 29, No. 1, pp. 1–32. Tony Lawson Cambridge January 1996
© 1997 Tony Lawson
Part I
SCIENCE AND ECONOMICS
© 1997 Tony Lawson
1 ENDURING TENSIONS AS POINTS OF DEPARTURE
Contemporary academic economics is not in a healthy state. Over many years now problems have regularly come to light which throw considerable doubt on the capacity of many of its strands to explain, or even always to address, real world events or to facilitate policy evaluation. Such problems especially beset the rather dominant ‘mainstream’ or ‘orthodox’ project,1 centring on econometrics and formalistic ‘economic theory’, which is my main concern here. This unhappy situation, moreover, appears to be increasingly recognised both inside and outside of the academy. On the outside, for example, the (UK) Observer Magazine (20 September 1992:7) concludes that ‘there’s no such thing as economics. It’s all voodoo, bluff and pseudo-science’. New Scientist (31 October 1992:26–31) even carried sketches of economists forecasting economic variables by reading lines on the palms of their hands and of econometricians pushing numbers around while blindfolded. Donaldson (1984) sums up this sentiment in reporting ‘a growing suspicion that in fact economists may not be very good at dealing with the real problems which face us’ (Donaldson, 1984:11). It seems that the ‘charge against economists is, above all, that of irrelevance’ (ibid.: 13). Within the economics’ academy, the unsatisfactory nature of the prevailing state of affairs appears increasingly to be acknowledged by non-orthodox and orthodox economists alike. In truth, non-orthodox economists have long recognised and emphasised it (see, for example, references in Hodgson, 1988). It is noteworthy, however, that those mainstream economists who reflect upon the nature and state of the discipline are increasingly also found to be airing some concern. The field of econometrics, for example, has witnessed a stream of contributions with such titles as ‘Econometrics—Alchemy or Science?’ (Hendry, 1980) or ‘Lets take the con out of Econometrics’ (Leamer, 1983). In the latter Leamer observes that ‘after three decades of churning out estimates, the econometrics club finds itself under critical scrutiny and faces incredulity as never before’ (p. 42), and goes so far as to suggest that ‘hardly anyone © 1997 Tony Lawson
takes anyone else’s data analyses seriously’ (p. 37). At the same time ‘economic theory’ or ‘pure theory’, that most prestigious strand of contemporary economics, including its traditional exemplar general equilibrium theory, recently seems also to have been recognised as a project in dire straights. The upshot here is a bout of articles with such titles as ‘The Intrinsic Limits of Modern Economic Theory: The Emperor has no clothes’ (Kirman, 1989), or books with such titles as Economics in Disarray (Wiles and Routh (eds), 1984), The Crisis in Economic Theory (Bell and Kristol, 1981) or ‘The Death of Economics’ (Ormerod, 1994), and commentaries by ‘pure theorists’ such as Hahn (1991) effectively acknowledging the demise of his long term hopes of successes for the project. One Nobel Memorial Prize winner in economic science despairs of the whole contemporary project, viewing it as basically an arbitrary activity which singularly fails to advance understanding: Page after page of professional economic journals are filled with mathematical formulas leading the reader from sets of more or less plausible but entirely arbitrary assumptions to precisely stated but irrelevant theoretical conclusions... Year after year economic theorists continue to produce scores of mathematical models and to explore in great detail their formal properties; and the econometricians fit algebraic functions of all possible shapes to essentially the same sets of data without being able to advance, in any perceptible way, a systematic understanding of the structure and the operations of a real economic system. (Leontief, 1982:104) Needless to say some diagnoses of this situation have been produced by those working within the mainstream project and certain ‘solutions’ proposed. But none have proven to be successful. I shall argue that this failure reflects an inadequate analysis of the nature of the disorder. The ‘solutions’ so far tried, indeed, have achieved little more than a modification in the form of the symptoms, and in certain cases might themselves be interpreted merely as novel manifestations of the underlying problems. An aim of this book then is to attempt to elaborate the real nature of the disorder afflicting contemporary economics with a view to identifying an appropriate course of action. Now when any project falls into a state of disarray, signs of disorder tend to be manifest in additional ways to failure in terms of the criteria which the project sets itself. This is so with modern economics, and in particular with the extraordinarily influential mainstream project which is my primary concern here. The most widely noted cause for concern is that while economic forecasters do not forecast sufficiently accurately, ‘theorists’ fail to provide non-arbitrary explanations of any real economic phenomenon of interest. But © 1997 Tony Lawson
at least as significant are widespread and persistent tensions and incongruities that are best described as theory/practice inconsistencies. In contemporary economics, there are many situations in which economists claim, in their theorising, to be doing one thing, when it is quite evident from their practices, and acknowledged by the more reflective members of the community, that they are actually doing something quite different. In the light of the confusions and uncertainties which such theory/practice inconsistencies can generate, resulting in unhappy compromises, ad hoc accommodations, the explicit suppression of criticism and other pathologies of action, the existence of these inconsistencies is at least as significant as the shortcomings in actual results. Although the inconsistencies between theoretical perspectives and actual practices that I have in mind are numerous and span all aspects of academic economic research, it is sufficient for my purposes to identify but a few, albeit widely prevalent and enduring, forms. Specifically, I shall briefly sketch three types, relating, in turn, to the implementation of method, social theory and methodology. These examples should be sufficient to indicate the scope and nature of the problems. They also make it apparent that any programme of transforming the discipline into something rather more useful than it has been over the last fifty years or so will need to transcend (and explain) enduring contemporary incoherences and problems that are somewhat more wide ranging than the inability of the contemporary project to forecast or explain any actual economic event or situation.
INCONSISTENCY AT THE LEVEL OF METHOD A first disparity that I want to consider, one that is often commented upon (e.g. Blaug, 1980; McCloskey, 1983, 1986), is that economists frequently employ methods, practices and techniques of enquiry and modes of inference, that are inconsistent with the theoretical perspectives on method which they claim to draw upon. As McCloskey (1983) interprets the situation: Economists do not follow the laws of enquiry their methodologies lay down. A good thing, too. If they did they would stand silent on...[most of the]...matters about which they commonly speak. In view of the volubility of economists the many official methodologies are apparently not the grounds for their scientific conviction. (McCloskey, 1983:482) In other words, according to McCloskey, economists ‘have two attitudes towards discourse, the official and unofficial, the explicit and the implicit. The official rhetoric, to which they subscribe in the abstract and in © 1997 Tony Lawson
methodological ruminations, declares them to be scientists in the modern mode’ (ibid.: 484). But it is not a rhetoric, in McCloskey’s assessment, that fits their actual practices. Econometrics is the paradigm example here. It is widely observed, for example, that when econometricians formulate and (often simultaneously) estimate their ‘models’ they frequently ‘run’ hundreds if not thousands of regressions, and in doing so contradict the classical theory of inference they explicitly acknowledge. Moreover, when their models are used to forecast unobserved (typically future) states of the economy, econometricians repeatedly make ad hoc revisions to estimated parameter values, or introduce ‘add on’ factors, in order to generate results that are ‘sensible’ or ‘believable’, thereby contravening what Lucas designates the ‘theory of economic policy’ (Lucas, 1976). Thus although economists do pay ‘lip-service’ to this latter theory Lucas identifies ‘many signs that practicing econometricians pay little more than lip-service to the theory’ (ibid.: 22). It is this observed theory/practice inconsistency, indeed, which motivated the extraordinarily influential ‘Lucas critique’ of economic policy evaluation (ibid.). In this critique (which we shall cover in Chapter 7) Lucas focuses specifically upon the ‘indifference of econometric forecasters to [certain sub-sets of available] data series’; ‘the frequent and frequently important refitting of econometric relationships’; and ‘most suggestively... the practice of using patterns in recent residuals to revise intercept estimates for forecasting purposes’ (ibid.: 22, 23), where all such practices are inconsistent with accepted theory. It is not only outsiders to the econometrics community who draw attention to these and related matters. Indeed, one of the most telling assessments is provided by Learner (1978) who is himself a respected theoretical and practising econometrician. Learner identifies a clear gap between econometric theory and practice. Moreover, in doing so he draws attention to the calmness with which most econometricians accept this situation: The opinion that econometric theory is largely irrelevant is held by an embarrassingly large share of the economics profession. The wide gap between econometric theory and econometric practice might be expected to cause professional tension. In fact, a calm equilibrium permeates our journals and our meetings. We comfortably divide ourselves into a celibate priesthood of statistical theorists, on the one hand, and a legion of inveterate sinner-data analysts, on the other. The priests are empowered to draw up lists of sins and are revered for the special talents they display. Sinners are not expected to avoid sins; they need only confess their errors openly. (Leamer, 1978:vi) © 1997 Tony Lawson
In fact the situation is further puzzling. For those who teach econometric theory and those who practice econometric modelling not only employ identical rhetoric (the rhetoric of the theory that supposedly grounds the officially sanctioned methods of enquiry), they also usually constitute the same individuals flitting between the noted incompatible activities. Consider Leamer’s experiences once more: I began thinking about these problems when I was a graduate student in economics at the University of Michigan, 1966–1970. At that time there was a very active group building an econometric model of the United States. As it happens, the econometric modelling was done in the basement of the building and the econometric theory courses were taught on the top floor (the third). I was perplexed by the fact that the same language was used in both places. Even more amazing was the transmogrification of particular individuals who wantonly sinned in the basement and metamorphosed into the highest of high priests as they ascended to the third floor. (ibid.: vi) What can be made of all this? We can allow that theory/practice inconsistencies per se are characteristic of all formative, learning and/or developmental processes. They are a pointer that, for (further) progress to be achieved, the existing framework of analysis must be transformed in some way. If, for example, theory cannot be followed in practice, or if practice achieves something which in theory is impossible, then the theory, at least, stands in need of revision. As any research project progresses the only sort of consistency over time that can conceivably be sustained, or reasonably aspired to, or expected of it, is of a developmental sort, much like a tadpole evolving into a frog or initial thoughts and jottings being transformed into a thesis. The real failing of contemporary econometrics that Leamer is observing, however, is the preparedness of so many econometricians to live with the observed incongruities, and thereby to sanction the proliferation of compromise formulations, ad hoc stratagems and the like, which inevitably result. Leamer himself, it should be emphasised, is not guilty of this. Indeed, his book, which is premised on the belief that it ‘is a sin to not know why you are sinning’, represents a desire to ‘discard the elitism of the statistical priesthood and proceed under the assumption that unavoidable sins cannot be sins at all’ (ibid.: vi). Ultimately, however, he recognises that his own proposed solution, of replacing the classical theory with the Bayesian theory of inference, can achieve only limited success, with the result that Leamer is rather pessimistic about the prospects: © 1997 Tony Lawson
Nor do I foresee developments on the horizon that will make any mathematical theory of inference fully applicable. For better or for worse, real inference will remain a highly complicated, poorly understood phenomenon. (ibid.: vi) But if a mathematical theory of inference is inapplicable to economic phenomena it does not follow that a satisfactory and comprehensible alternative is beyond reach. Certainly a feasible alternative hypothesis worth contemplating is that a somewhat more radical response than a shift from the classical to the Bayesian approach is required. Moreover, given that econometricians appear to decide which equation specifications including parameter estimates and/or model extrapolations are acceptable prior to any formal econometric exercise, the question naturally arises as to whether a formal method of statistical inference is really required at all. In any case, I shall demonstrate in due course that a radical transformation in the over-all orientation of the subject can indeed succeed in transcending problems of the sort in question. Before turning to such matters explicitly, though, some further examples of practices which are inconsistent with the theories which supposedly ground them, can be noted. For it is not only in the labours of econometricians that we find theory and practice to be persistently out of step.2
INCONSISTENCY AT THE LEVEL OF SOCIAL THEORY Other such tensions or inconsistencies turn on, or, more accurately, are manifest at the level of, the substantive content of economic theories. Specifically, in their day-to-day discussions, generalised rhetorics, public and/ or programmatic stances, mainstream economists refer to many substantive items which their official ways of proceeding cannot accommodate without severe strain. Such items even include the basic economic categories of choice, institutions, markets, money, social relationships, uncertainty and change (see e.g. Farmer and Matthews, 1991:106). I focus here on just one of these. Most economists appear to acknowledge an intuition that human beings possess the capacity of exercising real choice. Not only does the language of choice come into all the day-to-day musings of economists and figure in their programmatic stances, but mainstream economists frequently even refer to their project, or to a prominent strand of it at least, as choice theory. So prevalent is the rhetoric of choice, indeed, that Duesenberry (1960:233) famously concludes that what distinguishes the subject (from sociology) is that ‘[e]conomics is all about how people make choices’.3 © 1997 Tony Lawson
Yet it isn’t. In the formal ‘models’ found in mainstream journals and books, human choice is ultimately denied. For if real choice means anything it is that any individual could always have acted otherwise. And this is precisely what contemporary ‘theorists’ are unable to allow in their formalistic modelling. That is, although the reality of choice appears to be widely acknowledged by economists in their more informal discussions and public pronouncements, the exercise of choice is a phenomenon that is always absent from the formal substantive analyses that are conventionally reported. Instead, individuals are represented in such a way that, relative to their situations, there is almost always but one preferred or rational course of action and this is always followed. Despite some suggestive rhetoric, human doings, as modelled, could not have been otherwise. (On this also see Shackle, 1972, 1989; Loasby, 1976; Farmer, 1982; Hodgson, 1988, 1993a; Runde, 1997.) The disparity in question here, as with the inconsistencies of econometric theory and practice discussed above, is widely recognised in mainstream contributions. In this regard attention is often drawn to ‘dilemmas’, ‘conundrums’, and the like. But few resources are devoted to resolving the matter. The usual resort, it seems, is to hope that the noted tension will somehow disappear of its own accord. In practice, ways of coping are usually arbitrary. A familiar enough example of this is the partitioning of postulated econometric relationships into endogenous and exogenous components, with only the latter being regarded as subject to choice, albeit in the exercise solely of a privileged policy-making elite. But, no sooner is this latter assignment of the possibility of choice made than it is undermined by the remorseless tendency of that same project to endogenise ever more variables. A recent example of this lies in the efforts of proponents of the rational expectations hypothesis to convince other mainstream economists to accept, and to produce models consistent with, the claim that all agents’ beliefs are (in the sense of that hypothesis) rational. A version of the dilemma or ‘conundrum’ which thereby arises is noted by Sargent and Wallace (1976): The conundrum facing the economist can be put as follows. In order for a model to have normative implications, it must contain some parameters whose values can be chosen by the policy maker. But if these can be chosen, rational agents will not view them as fixed and will make use of schemes for predicting their values. If the economist models the economy taking these schemes into account, then those parameters become endogenous variables and no longer appear in the reduced form equations for the other endogenous variables. If he models the economy without taking the schemes into account, he is not imposing rationality. (Sargent and Wallace, 1976:183) © 1997 Tony Lawson
This puzzle for orthodox economics is alternatively put by Reder (1982): Associated with the assumption of stable preferences, but logically distinct, is the ‘thrust for endogenisation’. A leading manifestation of this tendency is Stigler’s attempt to explain—and constrain—the behaviour of political decision makers, but it is not the only one... Successfully to endogenize a new variable is to enhance the explanatory power of economics, and there is much interest in such achievements. However, it must be noted that where variables are made ‘endogenous’, they can no longer serve as objects of social choice. To the extent that variables are endogenized—choice is explained— ‘society’s’ freedom of choice is seen as illusory. Freedom appears to consist not in power of choice, but (pace Hegel) in recognition of necessity. This is not a likely conclusion for followers of Adam Smith, and surely not one they desire, but one from which they can be saved only by failure of this direction of research. (Reder, 1982:34–5) We might notice, too, that the knowledge process as represented in mainstream formulations, is always a deterministic and in fact quite individualistic affair, wherein each agent unproblematically records economic ‘signals’ or other atomistic items of information. Usually, the latter signals take the form of market prices. But, in general, any measurable event or state of affairs is a potential candidate. All that appears to be essential in this process is that agents be passive sensors, that is, unreflexive receivers and recorders of whatever ‘signals’ are deemed to be available. The overall result is a conception wherein any contribution of human beings to the active making of their own history is denied. Indeed, the continuing tendency of mainstream economics to characterise human agents as passive automata, propelled along under the influence only of external forces, invites here a restatement of Veblen’s so colourful and apt enduring assessment of this, usually ‘hedonistic’, person: The hedonistic conception of man is that of a lightning calculator of pleasures and pains, who oscillates like a homogeneous globule of desire of happiness under the impulse of stimuli that shift him about the area, but leave him intact. He has neither antecedent nor consequent. He is an isolated, definitive human datum, in stable equilibrium except for the buffets of the impinging forces that displace him in one direction or another. Self-imposed in elemental space, he spins symmetrically about his own spiritual axis until the parallelogram of forces bears down upon him, where upon he follows the line of the resultant. When the force of the impact is spent, he comes to rest, a self-contained globule of desire © 1997 Tony Lawson
as before. Spiritually, the hedonistic man is not a prime mover. He is not the seat of a process of living, except in the sense that he is subject to a series of permutations enforced upon him by circumstances external and alien to him. (Veblen, 1919:73, 74) The dilemma for mainstream economists, then, is whether to abandon the widely felt intuition that human choice is real or to stick with it and forego methods and associated theories which ultimately cannot accommodate it. Formally, in contriving formalistic ‘models’ and insisting that this is the only ‘proper’ way to proceed (even if ‘unofficial practices’ may be rather different) mainstream theorists settle for the former option. Quite clearly, though, if choice is a real feature of human agency that is continually exercised we ought to be doing something rather better than sanctioning practices or theories (of explanation as well as human doings and situations) in which it ultimately cannot be sustained.
INCONSISTENCY AT THE LEVEL OF METHODOLOGY A third tension that can be mentioned here concerns the prevalent orientation to methodology. While mainstream economists frequently conclude that ‘methodology’ or ‘philosophy’ is irrelevant to (progress in) economic science, and actively discourage it, these very same economists appear quite unable to refrain from explicit methodological discussion and enquiry themselves (see e.g. Backhouse 1992; Hahn, 1992a, 1992b). Methodology, it seems, is regarded as being useless to, and a distraction from, economic science in theory and yet indispensable for it, or anyway unavoidable, in practice. Part of the puzzle here is why it should ever be thought that methodology could be dispensable in theory. Indeed, it will be necessary to account for this theoretical perspective in due course. For the time being, however, it can be noted that such a theoretical orientation is pervasive, being manifested in many ways. Perhaps the most immediate form is the perpetual repetition of such quips as ‘don’t think about it just do it’; or ‘methodologists are crazy’ or ‘those who can, do economics while those who cannot, do methodology’.4 An effective restraint on methodology, moreover, is the clear reluctance of mainstream economic journals to publish much of it. Also significant is the apparent refusal of many central research funding authorities to promote it. In the UK, for example, the training currently provided and recommended for economics students tends to be more or less devoid of any explicit methodological content. A quick run through a recent Economic and Social Research Council5 Guidelines for Post Graduate Training reveals that © 1997 Tony Lawson
economics is one of only two out of the twenty subject guidelines provided (the other being for ‘town planning’) that do not include an explicit section detailing the need for some form of formal training in the ‘philosophy of the social sciences’ or something similar. But this opposition to methodology is not (and could hardly be) a coherent or consistent one.6 Not only do those who most vociferously oppose methodology repeatedly fall back upon it in their own substantive contributions, but every now and again, if mainly in coffee-room discussions or local seminars, they can be found putting forward arguments and defending positions that can only be described as overtly methodological/philosophical. For example, Frank Hahn has both published on methodology (see for example Hahn, 1984, 1985) and, in my own institution, frequently entered into local methodological debate. Yet, his official position has long been one of opposition to training or study in such matters. Thus when in the July 1992 issue of the Royal Economic Society Newsletter, Backhouse puts the question: ‘Should we ignore methodology?’, the heading of a response by Hahn is ‘Answer to Backhouse: Yes’. The interchange which follows is itself a response to a contribution by Hahn in the earlier, April 1992, Newsletter whereby, on the occasion of his retirement from Cambridge, Hahn offers various ‘reflections’ which take the form of advice to young economists. Notably these include the recommendation to ‘avoid discussion of “mathematics in economics” like the plague’, and [as if unrelated] to ‘give no thought at all to methodology’. When Backhouse, in the later July issue of the Newsletter points out that Hahn too practices methodology, Hahn’s grudging acknowledgement of this observation is still tempered by an explicit astonishment that anyone could think that economists should learn philosophy, or even spend much time on philosophy/methodology at all: Methodology like original sin won’t go away, and Backhouse is right in saying that I myself have sinned. Perhaps it would have been better if I had not.... What I really wanted to advise the young to do was to avoid spending much time and thought on it. As for them learning philosophy, whatever next? (Hahn 1992b) It perhaps warrants some emphasis here that orthodox economists do not just formally discourage methodology, they do so explicitly and boldly. They do not so much neglect the topic (as, say, they often neglect real world structures and abuses of power, widespread wars, famines and other miseries, social decline) they boldly assert, as Hahn does, that any engagement in methodology is a waste of time. Moreover, they do so, by and large, without much explicit or cogent argument. The most that tends to be offered by way of persuasion are rather bland assertions. It is as if the antagonists hardly feel © 1997 Tony Lawson
it is worth bothering to think about the question anyway. Caldwell (1990:64; see also 1993:45) records a similar observation. At the 1989 History of Economics Society meetings in Richmond, Virginia there was a session entitled, ‘Should Methodology Matter to the Economist or to the Historian of Economics?’ Some of the participants answered in the negative. As an observer I was disappointed in the session, not because the study of methodology was attacked, but because the attack was such an anaemic one. The major worry seemed to be that many economists think that methodological study is a waste of time. One panellist even suggested that it would be alright to keep doing methodological investigations as long as we called them by another name so as not to offend our fellow economists. Lest there be any doubt, it should be stated at the outset that, at least in the US, most economists are indifferent towards methodology, and many of the rest are openly hostile to it. Indeed, explaining these attitudes is a methodological topic worthy of further study. But their existence, at least for now, should be taken as a given. (Caldwell, 1990:64) Now there may be some validity to Hahn’s impression that methodological discussion within economics has often proven ex post to be largely unfruitful. But, if this is so, the widespread a priori assessment that methodology cannot bear fruit, as we shall see below, is not one that thereby follows, nor one that has any validity at all. It is merely a claim, to repeat, that is never really argued but asserted. However, in their failure to articulate cogent arguments mainstream economists, I suspect, are not choosing to be arbitrary or to discriminate randomly. Rather, to such people the dismissal of methodology appears to be a matter of common sense, an intuition that lies naturally with more generally held belief systems and mentalities. In consequence, the dismissal takes on the appearance of being philosophically neutral. It is put forward as the obviously sound attitude for any economist to take, at one with the project of economics, and well captured in such remarks as ‘don’t think about it just get on with it’. To repeat, however, it represents a stipulation that economists are unable to live up to. And the point here is that the coincidence of a widespread acceptance of this stipulation in economics with an obvious inability to act consistently with it, is a further phenomenon that warrants some explanation, a situation which once more suggests an undesirable level of confusion and inconsistency in the contemporary economics discipline. In summary, contemporary economics is not in a fit state. Most obviously, it fares poorly on its own terms; it neither provides particularly accurate forecasts of events nor illuminates the world in which we live. But of equal © 1997 Tony Lawson
significance, the whole project is riddled with confusion and incoherence. Rarely, though, is the significance of this wider situation acknowledged. Certainly the persistent theory/practice inconsistencies focused upon here will be recognisable to most practising economists. But of those who work in, or who address, the mainstream, only a few (Blaug, Hendry, Leamer, Lucas, McCloskey, Reder and the like) seem prepared to treat these inconsistencies with anything like the sort of seriousness they warrant. Rather the dominant tendency is to regard the noted puzzles or tensions in the manner of mild irritants that can be lived with, if necessary, but which are expected, in due course, somehow to wither away of their own accord. Thus, the failure of the contemporary project to explain, predict or otherwise illuminate is met with only a restricted set of responses: with continuous revisions to certain parts of theories, the collection of ever more data, the development of increasingly sophisticated forms of computer software, and so forth. While such developments are usually welcome ones, in the main they do nothing to address, let alone resolve, problems, tensions, confusions and consequent pathologies of action, which turn on the acceptance of enduring discrepancies between practices adopted and the theories which are supposed to inform them. All such responses, rather, share the feature of allowing, in their pursuit, that the existing ways of doing economics by and large remain untouched. Once, however, we face up to the depth and extent of the problems that pervade the discipline it is apparent that some more thorough-going response is warranted. And, amongst other things, it is evident that a rather fundamental, necessarily broadly methodological, enquiry, and a sustained one at that, is called for. In what follows, in fact, it will be seen that a resolution of the noted inconsistencies, along with other more general limitations of the contemporary discipline, necessitates a quite radical transformation in the very nature of contemporary economics.
© 1997 Tony Lawson
2 REALISM, EXPLANATION AND SCIENCE
How are the problems and failings of modern economics to be explained and resolved? The answer developed here is that these problems: (1) result ultimately from a widespread, rather uncritical, reliance by economists upon a questionable conception of science and explanation; and (2) can be resolved through replacing this conception with a more adequate one, derived by way of adopting an explicitly realist orientation. The next few chapters are concerned with elaborating upon this ‘answer’. In the process I indicate just how adopting a realist orientation (in the sense understood here) can facilitate a more relevant economics than is currently available or is even possible within the confines of the explanatory norms and criteria of the contemporary economics discipline. It is not until Chapter 4, however, that I deal fully with the problems and/or inconsistencies just noted.
REALISM I start by describing the sort of position that I am designating realist. In fact any position might be designated a realism (in the philosophical sense of the term) that asserts the existence of some disputed kind of entity (such as black holes, class relations, economic equilibria, gravitons, tanon-neutrinos, utilities). Clearly on this definition we are all realists of a kind, and there are very many conceivable realisms. In science, a realist position, i.e. a scientific realism, asserts that the ultimate objects of scientific investigation exist for the most part quite independent of, or at least prior to, their investigation. Now the conception of realism I want to argue for is closely and explicitly bound up with ontology or ‘metaphysics’, i.e. with enquiry into the nature of being, of existence, including the nature, constitution and structure of the objects of study. Indeed, it is a forthright concern with ontology, and in particular with elaborating the broad nature of aspects of natural and social © 1997 Tony Lawson
reality, that explains, in what follows, the term realism being used in labelling perspectives distinguished. First and foremost, then, the term realism is utilised here to denote specific accounts of the nature of reality, natural and/or social. But why bother with realism so interpreted? Surely economists can and should avoid wasting their time reflecting upon such matters as ‘realism’ and ‘ontology’, and, accepting the goals, criteria, methods and procedures as laid down in standard economic texts books and recent articles, just get on with the job? Even accepting that contemporary economics has proven to be notoriously unsuccessful over the last fifty years or so, is not the proper solution just to continue as before, albeit trying that little bit harder? The latter has been the usual and recommended response, of course, at least within the mainstream. The problem with it, we shall see, is that the consequent disregard for the nature of social material has in practice resulted in certain methods and procedures being accepted as fundamental to ‘proper’ economics, and even held as more or less universally valid, that are actually quite inappropriate to the analysis of social objects.1 In propagating their project contemporary orthodox economists seem merely to assume that whatever the methods employed, social reality must, as it were, ‘fit’. We will find that it is this presumption that ultimately lies behind the problems of the discipline noted in the previous chapter. It is thus with the intention of avoiding such an epistemological reductionism, and in particular of facilitating ways of proceeding that are adequate to social objects, that ontology is an explicit concern here. In other words, a realist orientation of the sort I am intending to defend insists that methods of social science can, and indeed should, be designed to take account of available insights concerning the nature of social material. To repeat, it will be seen that it is by way of adopting such a strategy that the noted problems of the discipline can be resolved.
DEDUCTIVISM But is there really a set of methods or ways of proceeding in economics that can be shown to be both as fundamental to the mainstream project as I am suggesting, and yet largely irrelevant to the analysis of social phenomena? Even accepting the often noted incapacity of orthodox economics to explain or otherwise illuminate real world phenomena, surely it is impossible to lay the ‘blame’ for this at the door of any erroneously universalised set of methods or modes of analysis? Actually, I think it is possible. The main ‘culprit’, I shall argue, is a mode of explanation that can be referred to as deductivist, or, more particularly, it is the conception of ‘laws’ (or ‘significant results’ or ‘theoretical formulations’) upon which deductivist explanation ultimately depends. © 1997 Tony Lawson
This conception of laws2 is formulated in terms of constant conjunctions of events or states of affairs. It is an interpretation of laws as, or as dependent upon, constant relations connecting outcomes at the level of the actual course of events or states of affairs. On this view, laws, which are often referred to as ‘covering laws’, express regularities of the form ‘whenever event x then event y’. This formulation should be interpreted quite generally. Thus, ‘event x’ can be a composite of many events, for example, and the suggested relationship between events can be probabilistic (so that y can be interpreted as the average or limit of a series) or deterministic. In practice, the noted perspective on laws is usually associated with the principle that such laws are to be assessed (confirmed, corroborated, falsified, tested) by their instances. By deductivism I simply mean the collection of theories (of science, explanation, scientific progress, and so forth) that is erected upon the event regularity conception of laws in conjunction with the just noted principle of theory assessment. Now, generally speaking, to explain some event, thing, or phenomenon, (i.e. the ‘explanandum’) is to provide an account (the ‘explanans’) whereby the initial phenomenon is rendered intelligible. According to deductivist explanation, then, the explanandum must be deduced from a set of initial and boundary conditions plus universal laws of the form ‘whenever event x then event y’. Clearly on this deductivist conception, explanation and prediction amount to much the same thing (this is the so-called symmetry thesis). The former entails the deduction of an event after it has (or is known to have) occurred, the latter prior to (knowledge of) its occurrence.3 According to deductivism, the explanation of laws, theories and sciences similarly proceeds by deductive subsumption. We can note, parenthetically, that this theory of explanation is also variously known as the covering law model, the Popper-Hempel theory of explanation (Popper first restated it in recent times, Hempel has been its most ardent defender), the deductive-nomological model (nomos being Greek for law) or D-N model for short, amongst other things. In the case of ‘statistical laws’ the D-N description is often replaced by the ‘inductive-probabilistic’ or I-P label, though many include both under the former deductive-nomological head. The deductivist model of explanation is easily illustrated. Suppose I want to use the D-N model to explain the event that my car radiator contained ice this morning. I need a set of initial conditions and at least one universal ‘covering law’ which logically entails this event. A possible contender for an explanation comprises the following set of initial conditions: (1) my car radiator contained water yesterday, (2) my radiator does not leak, (3) the temperature fell below 0°C last night, along with the empirical law: water freezes at 0°C.4 In practice this sort of explanation is frequently laid out as follows: © 1997 Tony Lawson
1 Initial conditions 2 Laws or theories 3 Phenomenon explained and/or predicted Thus, the example just given can be structured in the following manner: I1 I2 I3 L1
My car radiator contained water yesterday My car radiator does not leak The temperature fell below 0°C last night Water freezes at 0°C
P1 My radiator contained ice this morning I do not think that it is contentious to observe that deductivism so understood, and in particular the conception of laws which underpins it (however the event regularities might be labelled or interpreted), characterises contemporary mainstream economics. Just as the persistent search for event regularities of a probabilistic kind characterises econometrics for example (see Chapter 7), so the positing of strict constant event conjunctions, interpreted usually as ‘axioms’ or ‘assumptions’, is a condition of modern ‘economic theorising’5 (see Chapters 8 and 9). Rather, it is implicitly taken for granted that the deductivist model of explanation is universally valid. Those few, typically non-orthodox, economists who occasionally voice a degree of concern at the extent to which mainstream economists rely upon the deductivist conception are usually (mis)interpreted as not agreeing with the particular substantive premises, the axioms or assumptions, that are proposed. Criticism of the mode of reasoning is something that appears rarely even to be contemplated within the mainstream; the relevance of deductivism is never questioned. As Hahn (1985) writes:6 Opponents of [economic] theory often argue that it is tautologous because it consists of logical deductions from axioms and assumptions. If one is kind to such critics one interprets them as signalling that they do not care for these axioms and these assumptions. In any case all theory in all subjects proceeds in this manner. (Hahn, 1985:9) Here, though, I do want to question the presumption of the generalised relevance of deductivism. In particular, I want to determine the conception of reality presupposed by it, and specifically by the theory of laws or law-like statements on which it depends, and to assess the relevance of this conception as a characterisation of social phenomena. My aim, in other words, is to © 1997 Tony Lawson
determine the character of social material which would legitimise the wielding of methods and techniques premised upon the deductivist theory of explanation and the associated conception of scientific laws. Once this is achieved we can assess the extent of relevance of the ontological conception uncovered and so of deductivism. More specifically, we can determine the actual bearing of the latter for the social domain.
EMPIRICAL REALISM It is clear, in fact, that if the theory of explanation and science in question turns upon identifying or positing regularities of the form ‘whenever event x then event y’—let us refer to systems in which such constant conjunctions of events arise as closed—then a precondition of the universality, or wide applicability, of deductivism is simply that reality is characterised by a ubiquity of such closures. Are there, then, grounds for supposing that event regularities or closed systems of the kind in question are pervasive? Historically, encouragement for the conception of science and explanation in question stems from a version of positivism that is rooted in Hume’s analysis of causality, or at least from Hume’s arguments as they are usually interpreted. Now positivism as I interpret the term7 is first of all a theory of knowledge, its nature and limits. Specifically, it is a claim that human knowledge takes the form of senseexperience or impression. But any theory of knowledge presupposes some ontology. For it must be supposed, even if only implicitly, that the nature of reality is such that it could be the object of knowledge of the required or specified sort. The positivistic conception in question entails an account of reality as consisting of the objects of experience or impression constituting atomistic events. Indeed, because reality is essentially defined in positivism as that which is given in experience, I follow Bhaskar and others in referring to the perspective on reality sustained as empirical realism. From this perspective it follows that if science is to be at all possible, it must take the form of elaborating regularities of the type ‘whenever event x then event y’. For if particular knowledge is restricted to atomistic events given in experience, the only possibility for general, including scientific, knowledge is the elaboration of patterns of association of these events. It is thus such constant event patterns, or regularities of the form ‘whenever event x then event y’, that constitute the Humean or positivist account of causal laws. If, moreover, the successes of science are perceived to be widespread it also follows that Humean causal laws must be ubiquitous. It follows that if I am to argue against the presumed universal applicability of the deductivist conception of explanation (and, to repeat, I am suggesting © 1997 Tony Lawson
that its usefulness in the social context is highly circumscribed), it is essential that I formulate an alternative theory of ontology to empirical realism (i.e. to the claim that reality is exhausted by atomistic events and their constant conjunctions). That is, it is necessary that I now set out a theory of ontology which at least allows the possibility of a different conception of science to the search for the noted event regularities. And of course, given the intention of my project, it is equally essential that I establish that this alternative conception is, while deductivism is not, appropriate to the analysis of (features of) the social realm. However, before considering peculiarities of social phenomena explicitly, that is before elaborating a theory of social ontology specifically, it is necessary to give a more adequate characterisation of the material of the natural realm as well. For we will find that even in this realm, spontaneous event regularities of the form ‘whenever event x then event y’ are hardly ubiquitous but arise only in rather special circumstances. In fact, at a relatively high level of abstraction it is possible to set out and defend a philosophical ontology that, as it turns out, both characterises natural and social phenomena alike, and is also seen to support an alternative account of science and explanation that is broader than deductivism. In the rest of this chapter I will be occupied with describing this alternative conception and contrasting it with the deductivist framework already described. Only in the following chapter will the superiority of this alternative conception over deductivism be demonstrated. Because at this point I am concerned mainly with conveying the nature of the conception I am wanting to defend, and because at the level of abstraction in question this conception will be seen in due course to apply to natural and social phenomena alike, I shall illustrate with examples drawn intermittently from natural and social domains, as suits my purposes.8 However, when I come, in Chapter 3, to argue for the relative adequacy of the conception in question, I shall, of course, consider the cases of the natural and social domains separately, each on its own merits.
TRANSCENDENTAL REALISM The alternative conception to empirical realism that I want to argue for, one which, to repeat, will be found in due course to apply to the natural and social realms alike (albeit, it is worth stressing once more, at a relatively high level of abstraction), is a perspective that has recently been explicitly systematised under the heading of transcendental realism. Here I draw extensively on the writings of Bhaskar. Now there are two very significant ways in which transcendental realism and empirical realism differ. The first is that according to transcendental realism (the reason for this particular © 1997 Tony Lawson
label will become clear in due course) the world is composed not only of events and states of affairs and our experiences or impressions, but also of underlying structures, powers, mechanisms and tendencies that exist, whether or not detected, and govern or facilitate actual events. The second difference is that, on the transcendental realist conception, the different levels of reality are out of phase with each other.9 Let me expand upon these differences and briefly set out my understanding of the categories stated. The conception I am proposing to defend is of a world composed in part of complex things (including systems and complexly structured situations) which, by virtue of their structures, possess certain powers—potentials, capacities, or abilities to act in certain ways and/or to facilitate various activities and developments. A bicycle, in virtue of its constitution or structure has the capability of facilitating a ride; gunpowder of causing an explosion; a language system of facilitating speech-acts. Such powers exist whether or not they are exercised. The bike can facilitate a ride even though it always sits in the back of the shed; the gunpowder has the power to cause damage even if it is never ignited; the language system makes a conversation possible even where people choose not to speak. In many cases we can infer something of a thing’s potential from a knowledge of its structure. Certainly a good deal about the powers or capabilities of rockets, planes, bridges and parachutes are inferred before any particular one is built and subsequently ‘tried out’. Complex things, then, have powers in virtue of their structures, and we can investigate their structures and in some instances thereby infer something of their powers. A mechanism is basically a way of acting or working of a structured thing. Bicycles and rockets work in certain ways. Of course they cannot work or act in the ways they do without possessing the power to do so. Mechanisms then exist as causal powers of things. Powers of structured things are usually exercised only as a result of some input: the striking of a match, the lifting and wielding of the hammer, the switching on of a computer or interacting with it, the flexing of vocal chords, the arrival of teachers and children at school or employees at the workplace. And mechanisms when triggered (where relevant) have effects. Structured things, then, possess causal powers which, when triggered or released, act as generative mechanisms to determine the actual phenomena of the world.10 I will expand upon the notion of a tendency below. The world on this view thus consists in more than the actual course of events and our experiences. Rather, three domains of reality are distinguished, namely the empirical (experience and impression), the actual (actual events and states of affairs in addition to the empirical) and the real (structures, powers, mechanisms and tendencies, in addition to actual events and experiences). Not only does the autumn leaf pass to the ground and not only do we experience it as falling but, according to the perspective in question, © 1997 Tony Lawson
underlying such movement and governing11 it are real structures and mechanisms such as gravity (or curved space). Similarly the world is composed not only of such ‘surface phenomena’ as skin spots, puppies turning into dogs, and relatively slow productivity growth in the UK, but also of underlying and governing structures or mechanisms such as are entailed in the workings of, respectively, viruses, genetic codes and the British system of industrial relations. In the UK, at present, not only are cows prematurely dying, or showing signs of losing control of their bodily functions, and not only is their behaviour perceived by observers as an illness (the condition is widely known in the UK as ‘mad cow disease’), but, according to scientists implicitly acting on the perspective here in question, this phenomenon is governed by some, as yet inadequately understood, agent or causal mechanism, and the search is on to uncover more about it.12 Now according to the conception in question, not only are the noted three domains ontologically distinct and irreducible (the real cannot be reduced to the actual nor the latter identified with the empirical) but also, and crucially, their characteristic components (mechanisms, events and experiences) are unsynchronised or out of phase with one another. This non-isomorphism of ontologically distinct items is the second noted contrast with empirical realism. Thus, just as experience is out of phase with events, allowing the possibility of contrasting experiences of a given event, so events are typically unsynchronised with the mechanisms that govern them. As an example of the independence of events and experience we can note that different people watching a football or hockey game frequently experience the same event (goal or foul) somewhat differently, just as a given person may experience an already observed event differently when later it figures on a television ‘replay’. The independence of mechanisms from the events upon which they bear is illustrated by the example of autumn leaves which are not in phase with the action of gravity for the reason that they are also subject to aerodynamic, thermal and other causal factors. Events, in other words, are conjointly determined by various, perhaps countervailing, influences so that the governing causes, though necessarily ‘appearing’ through, or in, events can rarely be read straight off.13 It is with regard to this latter situation that the category of a tendency is primed. Because actual events or states of affairs may be co-determined by numerous, often countervailing, mechanisms the action of any one mechanism, though real and perhaps expressing necessity in nature, may not be directly manifest or ‘actualised’. Characteristic ways of acting or effects of mechanisms which may not be actualised because of the openness of the relevant system are conceptualised here as tendencies. It is the idea of continuing activity (as distinct from enduring power per se) that the notion of tendency is designed to capture. Tendencies, in short, are potentialities which © 1997 Tony Lawson
may be exercised or in play without being directly realised or manifest in any particular outcome. It should be clear, then, that the way in which I am employing the term tendency differs from many of its interpretations in the economics literature (see Pratten, 1994). A statement of a tendency, according to its primary usage here, is not about long-run, ‘normal’, usual, or average outcomes at the level of events. Nor is it reducible to a counterfactual claim about events or states of affairs that would occur if the world were different. Indeed, it is not a claim about anything at the level of the actual course of events at all. Rather it is a transfactual statement about the typically non-empirical activity of a structured thing or agent; here transfactuals are not counter-factuals but take us to the level at which things are going on irrespective of the actual outcome. A statement of a tendency, in other words, is not a conditional statement about something actual or empirical but an unconditional statement about something non-actual and non-empirical. It is not a statement of logical necessity subject to ceteris paribus restrictions, but a statement of natural necessity without qualifications attached. It is not about events that would occur if things were different but about a power that is being exercised whatever events ensue. It is for example about the gravitational field which acts on the pen in my hand and continues to do so irrespective of whether I toss the pen in the air, continue to write with it, or drop it in a vacuum.
SCIENCE, INFERENCE AND LAW-STATEMENTS It can now be seen that transcendental realism supports a perspective on science and explanation that is quite different from that stylised above as deductivism. Science, on this transcendental realist view, is no longer confined to, or even dependent upon, the seeking out of constant event conjunctions, but aims at identifying and illuminating the structures and mechanisms, powers and tendencies, that govern or facilitate the course of events. The scientific objective is to identify relatively enduring structures and to understand their characteristic ways of acting. Explanation (in a manner to be described at some length in due course) entails providing an account of those structures, powers and tendencies that have contributed to the production of, or facilitated, some already identified phenomenon of interest. It is by reference to enduring powers, mechanisms and associated tendencies, that the phenomena of the world are explained. Tendencies, in this framework, are thus seen to be of fundamental significance in science. For it is because tendencies always have effects, and so, despite being non-empirical, are revealed in some way in whatever © 1997 Tony Lawson
outcome ensues, that science, on the conception I am proposing, is feasible. And it is because tendencies frequently act in conjunction with (a possibly shifting mix of) other, often countervailing tendencies (so that it will rarely be possible for any one tendency to be read straight off), that science will usually be non-trivial. We are also now in a position to see that the conception of science and explanation being proposed entails a rather distinct mode of inference to any sponsored by positivism. A concern with generalisations about conjunctions of events encourages a concentration of methodological debate upon questioning the relative advantages and limitations of methods of induction versus those of deduction, including its ‘falsificationist’ variant. Indeed, it is doubtless a predominance of this concern in much economic methodological discussion that has prompted the assessment of Hahn and others that methodological analysis has achieved very little. It is important to recognise, therefore, that the essential mode of inference sponsored by transcendental realism is neither induction nor deduction but one that can be styled retroduction or abduction or ‘as if’ reasoning. It consists in the movement, on the basis of analogy and metaphor amongst other things, from a conception of some phenomenon of interest to a conception of some totally different type of thing, mechanism, structure or condition that, at least in part, is responsible for the given phenomenon. If deduction is illustrated by the move from the general claim that ‘all ravens are black’ to the particular inference that the next one seen will be black, and induction by the move from the particular observation of numerous black ravens to the general claim that ‘all ravens are black’, retroductive or abductive reasoning is indicated by a move from the observation of numerous black ravens to a theory of a mechanism intrinsic (and perhaps also extrinsic) to ravens which disposes them to be black. It is a movement, paradigmatically, from a ‘surface phenomenon’ to some ‘deeper’ causal thing.14 According to transcendental realism, then, science aims at uncovering causal factors, that is, it is concerned with identifying structures, mechanisms and the tendencies they ground, which produce, govern or facilitate phenomena at a different level. And if the aim of science is to illuminate structures that govern surface phenomena then laws or law-statements are neither empirical statements (statements about experiences) nor statements about events or their regularities (whether unqualified or subject to ceteris paribus restrictions), but precisely statements elucidating structures and their characteristic modes of activity.15 From this perspective, moreover, there can be no a priori presumption that in the scientific explanatory endeavour there is any end to the process of revealing deeper levels of reality. If, for example, the phenomenon to be explained is copper’s ability to conduct electricity well, and we locate the reason in copper’s atomic structure, the latter then marks the next phenomenon to be explained, and so on. Of course, as deeper layers © 1997 Tony Lawson
of reality are revealed and understood this acquired knowledge may lead us to revise our previous conceptions of any phenomenon which the deeper level explains. Thus, science, in this framework, is seen to proceed via a continuing spiral of discovery and understanding, further discovery and revised, more adequate, understanding.
KNOWLEDGE AS A PRODUCED MEANS OF PRODUCTION Transcendental realism differs from empirical realism then in viewing the world as in part composed of objects, including causal laws, which are structured and, to adopt Bhaskar’s term, intransitive. They are structured in the sense of being irreducible to events and their patterns; it is mechanisms, tendencies and structures that are designated in causal laws. They are intransitive in the sense of existing, enduring and acting independently of the process of their identification; the mechanisms, tendencies and causal structures that are designated in causal laws are irreducible to our knowledge of them, and, in some part at least, endure and act independently of our knowledge of them (even if on occasion the identification and comprehension of social structures and processes leads in due course to their being transformed—see Part IV below). But if, according to transcendental realism, the primary aim of science is to identify and understand such objects where might our knowledge come from and how does it come about? Clearly, with an intransitive dimension acknowledged, we can understand how a changing knowledge of (possibly) unchanging objects is feasible. But if the objects of the world are not merely given in (and so effectively constituted through) sense experience, and specifically if structures and mechanisms act independently of their being identified, how does our knowledge of them arise? If knowledge is not merely given in experience, it is hardly intelligible that it is created out of nothing. It must, then, come about through a transformation of pre-existing knowledgelike materials. In other words, it is necessary (employing Bhaskar’s terminology once more) to recognise a transitive dimension to knowledge, or epistemology, to complement the intransitive dimension,16 or ontology, already established. That is, it is necessary to recognise a dimension of transitive objects of knowledge, including facts, observations, theories, hypotheses, guesses, hunches, intuitions, speculations, anomalies, etc., which condition all further knowledge,17 and in particular, facilitate, and come to be actively transformed through, the laborious social practice of science. In short, knowledge must be recognised as a produced means of production (of further knowledge) and science as an ongoing transformative social activity.
© 1997 Tony Lawson
Knowledge is a social product, actively produced by means of antecedent social products. Of course, if this must be our view of the nature of knowledge and its mode of production (accepting the perspective in question), our conception of the aim of science remains the production of knowledge of objects which mostly exist independently of our, or at least of any individual, knowledge of them. Specifically, if to repeat, the primary aim of science is the production of knowledge of mechanisms that, singularly or in combination, produce the phenomena that are actually manifest. To sum up, two broad perspectives on science and explanation have been set out and contrasted. The first, stylised as deductivism, depends upon, and in its claim to universality ultimately presupposes the ubiquity of, spontaneous event regularities; it accepts as an essential aspect of science and explanation the need to formulate regularities at the level of actual events or states of affairs. The perspective systematised as transcendental realism, in contrast, rejects the need to elaborate (strict) event regularities of the sort in question. Rather it construes science as a fallible social process which is primarily concerned to identify and understand structures, powers, mechanisms and their tendencies that have produced, or contributed in a significant way to the production of, some identified phenomenon of interest— mechanisms, etc., which, if triggered, are operative in open and closed systems alike. It is a conception in which science is characterised by its retroductwe mode of inference, by the move from knowledge of some phenomenon existing at any one level of reality, to a knowledge of mechanisms, at a deeper level or strata of reality, which contributed to the generation of the original phenomenon of interest.18,19 In setting out the different perspectives here I undoubtedly have adopted a rather positive presentational orientation to the transcendental realist position. However, the case for it in preference to deductivism, whether in regard to the analysis of the natural or the social realms, remains to be made, and is the concern of the following chapter.
© 1997 Tony Lawson
3 THE CASE FOR TRANSCENDENTAL REALISM
How is it possible to decide between the noted contrasting perspectives? In fact the inadequacy of the positivistic or Humean conception of the structure of scientific results, and so of deductivism as a generalised form of explanation of real phenomena, is immediately apparent once we examine the nature of the situations in which the sought-after event regularities actually obtain.
NATURAL SCIENCE: BHASKAR’S INFERENCE FROM EXPERIMENTS I first focus upon the situation in the natural sciences, turning to the social domain in due course. Specifically I want at this point to consider two observations whose relevance has been brought out by Bhaskar. The first of these is that, outside astronomy at least, most of the constant-event conjunctions which are held to be significant in science, in fact occur only under the restricted conditions of experimental control. In other words, they are not spontaneous in nature but a product of human intervention. The second observation is that results or insights obtained through controlled experimental activity are, along with other forms of knowledge, nevertheless successfully applied outside of the experimental situation.1 Put differently, constant conjunctions of events are in fact extremely rare, spatio-temporally restricted and usually artificially produced, while law-like knowledge appears to be generally available and useful, and some (but only some) of it experimentally corroborated. It is easy enough to see that these observations raise severe problems for any conception which ties laws or ‘significant results’ to constant conjunctions of events, i.e. to regularities of the form ‘whenever event x then event y’. For, if such regularities occur only in such restricted conditions as experimental set-ups, this bears the rather constricting implication that science and its © 1997 Tony Lawson
results, far from being universal, are effectively fenced-off from most of the goings-on in the world. In other words, most of the accepted results of science are not of the form ‘whenever event x then event y always follows’ after all, but of the form ‘whenever event x then event y always follows, as long as conditions e hold’, where conditions e typically amount to a specification of the experimental situation. A counter-intuitive implication of this situation, moreover, is that any event regularity that a law of nature supposedly denotes depends upon human intervention. For scientists and laboratory technicians are instrumental both in constructing and assembling any necessary apparatus and in conducting the experiment, in the absence of which the observed regularity, the supposed law of nature, would not occur. In addition to such problems, and at least as significant, the constant conjunction view of laws leaves the question of what governs events outside of experimental situations not only unanswered but completely unaddressed. In doing so, it also leaves the observation that experimentally obtained results are successfully applied outside experimental situations without any valid explanation. How though are we to make sense of the noted observations? How is it that scientists, in their experimental activities, can (frequently) co-determine a particular pattern of events which would not have come about but for their intervention? And how can we make sense of the successful application of science outside of the experimental laboratory, and specifically in conditions where event regularities do not necessarily occur? What must the world be like for such experimental practices, results and their successful nonexperimental application to be possible? In order to provide a satisfactory set of answers to such questions it is necessary to abandon the view that the generalisations of nature consist of event regularities. Instead, we must accept something very much like the transcendental realist account whereby the objects of science are structured (irreducible to events) and intransitive (existing and acting independently of their being identified). That is, experimental activity and results, and the application of experimentally determined knowledge outside of experimental situations, can be accommodated only through invoking something like the transcendental realist ontology of structures, powers, generative mechanisms and their tendencies that lie behind and govern the flux of events in an essentially open world. The fall of an autumn leaf, for example, does not conform to an empirical regularity, and precisely because it is governed in complex ways by the actions of different juxtaposed and counteracting mechanisms. Not only is the path of the leaf governed by gravitational pull, but also by aerodynamic, thermal, inertial and other mechanisms. According to this conception, then, experimental activity can be understood as an attempt to intervene in order to insulate a particular mechanism of interest by holding off all other potentially counteracting forces. The aim is to engineer a system in which the actions of any mechanism being investigated are more readily © 1997 Tony Lawson
identifiable. Thus, experimental activity is rendered intelligible not as the production of a rare situation in which an empirical law is put into effect, but as an intervention designed to bring about those special circumstances under which a non-empirical law, a mechanism or tendency, can be identified empirically. The law itself is always operative; if the triggering conditions hold, the mechanism is activated and in play whatever else is going on. On this transcendental realist understanding, for example, a leaf is subject to the gravitational tendency even as I hold it in the palm of my hand or as it ‘flies’ over roof tops and chimneys. Through this sort of reasoning we can render intelligible the application of scientific knowledge outside experimental situations. The context in which a mechanism is operative is irrelevant to the law’s specification. In short, although the traditional post-Humean conception of science necessitates the detection of constant conjunctions of events, such event regularities as have been reported are mainly restricted to situations of experimental control. Transcendental realism, unlike empirical realism, can make sense of this situation. And it follows from the transcendental realist perspective that the familiar post-Humean conception rests upon an inadequate analysis, and illegitimate generalisation, of what emerges as a special case: wherein a single and stable (set of) aspect(s) or mechanism(s) is physically isolated and thereby empirically identified. A premise of the above argument is the observation that outside astronomy strict event regularities are usually restricted to situations of experimental control. But what are we to make of astronomy, and specifically the celestial closure so successfully utilised by Newton? Is not the fact, as well as the spectacular nature, of this particular closure sufficient to justify the Humean insistence on the actuality of causal laws? The answer is no. First of all, it is possible from the transcendental realist perspective to provide an explanation of the celestial closure. It arises because of rather peculiar conditions that hold in the case of the planets, in that their intrinsic states as well as the extrinsic forces acting upon them are, in relevant respects, sufficiently stable, at least over the time period with which most people are usually concerned, i.e. over human life-spans. Properly interpreted, Newtonian mechanics posits theories of how bodies (tend to) act; celestial phenomena function merely as evidence of the postulated tendencies. Thus, if the intrinsic or extrinsic states of the planets in our solar system were not so stable but were to change in some way, perhaps a massive meteoroid were to pass through the solar system, then such a mechanics would entail a consequent disruption of the familiar celestial phenomenal patterns. There is no problem in this for transcendental realism. Second, and just as much to the point, it must be recognised that although the celestial closure is spectacular in nature, it represents a relative rarity in constituting a spontaneous closure. No doubt it is precisely its spectacular nature that accounts for some part of the general failure from © 1997 Tony Lawson
Laplace onwards to realise that the situation is relatively uncommon, to appreciate that the celestial closure is far from being indicative of the phenomenal situation that can be expected to prevail more or less everywhere. This failure, in turn, appears to be largely responsible for the widespread, if tacit, acceptance, formerly in philosophy, and currently in the social sciences in particular, of a ubiquity of constant conjunctions of events in nature, and thus of the doctrine of the actuality of ‘causal’ laws.
SOCIAL SCIENCE: INFERENCE FROM HUMAN INTENTIONALITY It will not have gone unnoticed that the argument in support of the transcendental realist account of science and explanation has so far focused only on the situation in the ‘natural’ sciences. However, it is not difficult to reason that the transcendental realist perspective also bears relevance for the social realm. Consider two often noted problematic features of contemporary economics. The first is the ex posteriori result that significant event regularities of the sought-after kind have yet to be turned up in the social realm (see Chapter 7). The second is the situation, already noted in Chapter 1, and especially celebrated in the programmatic headings (if not actual analyses) of mainstream economists, that human beings make choices. This latter situation is only problematic, of course, in that these same economists are seen to be unable to reconcile real choice with their project of economic ‘modelling’. An interesting question to pursue, then, is what is implied by the reality of people making choices? Now if choice is real any agent could always have done otherwise; each agent could always have acted differently than he or she in fact did. As Keynes insists: ‘A decision to consume or not to consume truly lies within the power of the individual; so does a decision to invest or not to invest’ (1973a:65). And a necessary condition for this is that the world, social as well as natural, is open in the sense that events really could have been different. Put differently, if under conditions x an agent in fact chose to do y, it is the case that this same agent could really instead have not done y. Choice, to repeat, presupposes that the world is open and actual events need not have been. But the possibility of choice presupposes not only that events could have been different, but also that agents have some conception of what they are doing and wanting to achieve in their activity. That is, if choice is real then human actions must be intentional under some description. And intentionality, in turn, is bound up with knowledgeability. For human beings must have some knowledge at least of the conditions that render their intended acts feasible. In turn again, knowledge presupposes a degree of endurability in the © 1997 Tony Lawson
objects of knowledge sufficient to facilitate their coming to be known. Now if, as widely reported, scientifically significant event regularities do not often occur in the social realm (or at least are yet to be uncovered), the enduring objects of knowledge that condition actual human practices must lie at a different level, at that of the structures which govern, but are irreducible to events, including human activities. The ex posteriori fact of human intentionality and choice indicates that there are real material causes or structures which facilitate intentional action. But it does not yet follow that there are structures which can be said to be clearly social. Now if the term social is to designate anything specific here, it must be a dependency on human intentional agency. This is a standard interpretation of the term, and does not seem contentious. The question, then, is whether such social structures exist. If they do, and if, like many features of the natural realm (powers, tendencies), they cannot (or cannot always) be perceived directly, the possibility of their detection will turn on the perception of their effects. In other words, in determining the real possibility of social science we must acknowledge that science employs not only a perceptual, but also a causal, criterion for the ascription of reality to a posited object. The latter turns on the capacity of the entity whose existence is in question to facilitate changes in material things. Entities which cannot be observed directly can be known to exist through the perception of their consequences at the level of actual events and states of affairs. This we have seen is a basic aspect of experimental activity. And it is only on the basis of the causal criterion that many familiar features of the physical realm are known. Thus, radioactive materials, electricity, magnetic and gravitational fields satisfy this criterion but not that of perceivability; their detection rests respectively upon the geiger counter, the electroscope, the compass, and perhaps even falling leaves and apples. In this way the reality of hypothesised entities can be assessed quite empirically, albeit indirectly. Once we accept the property of depending upon human agency as criterial for the social, and acknowledge the causal criterion for ascribing reality, it is easy enough to see that identifiable social structures do exist. Items such as (societal) rules, relations and positions clearly depend on human agency as well as condition our every day (physical) activities. The human (intentional) activities of speaking, writing, driving on public roadways, cashing cheques, playing games, giving lectures, and so forth, would be impossible without such social material conditions as rules of grammar, the highway code, banking systems, rules of play, teacher-student relationships, etc. All are structures which pre-exist and make a difference to (facilitate as well as constrain) related human activities. Commons (1934), for example, captures this property of social structures as follows:
© 1997 Tony Lawson
The business man who declines to use the banking system which has grown up in the past, the labourer who refuses to come to work when others come, may be industrious, but he cannot live in industrial society. This is familiar enough.... But when customs change, or when judges and arbitrators enforce a custom by deciding a dispute, or when labourers or farmers strike in order to modify a custom of business, or when a revolution confiscates slaves or other property of capitalists, or when a statute prohibits a customary mode of living, or when a holding company extends an old custom into new fields—then it is realised that the compulsion of custom has been there all along, but unquestioned and undisturbed. (Commons, 1934:701) To repeat, if it is the dependency of such structures upon human agency that marks them out as being social, it is their ability, in turn, to make a difference to (to enable as well as to constrain) physical states, or actions, that (just as with non-perceivable objects of the natural realm such as gravitational and magnetic fields) establishes that they are real. At a very general level, then, the transcendental realist conception carries over to the social realm. In the natural realm, strict or constant event conjunctions mainly occur (at least outside astronomy) in conditions of experimental control. Of course the feasibility of experimental control in the social realm is rather limited. Are there any grounds for expecting scientifically significant constant event conjunctions to be a feature of the social realm? If natural and social realms are similar in that both are characterised by structures underlying the course of events, they are of course dissimilar in that social structures in turn (in order to qualify for the designation of social) depend for their existence on human agency. Thus, although a language system is like gravity in that it facilitates human action it is unlike gravity in depending in turn on human action. Human agency and social structure then presuppose each other. Neither can be reduced to, identified with, or explained completely in terms of the other, for each requires the other. Now the significant point here is that because social structure is human-agent dependent it is only ever manifest in human activity. Thus, given the open nature of human action—the fact that each person could always have acted otherwise—it follows that social structure can only ever be present in an open system. In consequence, any economic laws must be interpreted as tendencies that are manifested as strict event regularities only very rarely, usually in conditions where they are consciously brought about (such as the occurrence of some annual holidays), so that the post-Humean project in its economic guise is, as a general approach, seen to be misguided. © 1997 Tony Lawson
ONTOLOGY AND THE EPISTEMIC FALLACY I have argued that the deductivist mode of explanation with its associated conception of the structure of scientific laws is limited in its legitimate scope of application, and I have attempted to situate it, to determine the conditions of its relevance. It turns out that, outside astronomy at least, the relevance of this mode of explanation is mainly restricted to situations of experimental control, giving little reason to expect any significant scientific applications in the social realm. What though follows for the conduct of economics? Have I not fallen into the same mistaken epistemological reductionism that I am criticising by arguing that the transcendental realist ontology and (associated) mode of explanation apply to the natural and social realms alike? I think not, but two aspects of the analysis may warrant further emphasis here. First, the argument for transcendental realism has not been an a priori one either in the case of the natural or the social realm. Certainly I have not asserted the relevance of the transcendental realist conception for the social realm on the basis of its perceived adequacy for the natural domain, or some such. Rather, the case for the transcendental realist ontology and its associated conceptions of science and explanation has been made separately for the natural and social realms alike. In the former instance the premise of the argument has been the experimental experience; in the latter the argument has turned upon rendering intelligible the fact of human choice and intentionality and the more or less total absence of any scientifically significant social event regularities so far.2 As it happens the transcendental realist ontology and general mode of explanation and conception of science receive support in both instances. But this outcome is a conclusion of ontological enquiry, not a premise. It is worth labouring this point, for a central aim of my project is to indicate the significance of ontological enquiry, of facing up to ontological issues explicitly. And we are in a position to see that, ultimately, the basic error underpinning the now discredited positivist conception of science is precisely the abandonment of explicit ontological reasoning. With Hume this error turns on the use of the category of experience to define the world. This entails giving an epistemological category an ontological task and constitutes a specific example of a general mistake which, following Bhaskar once more, can be labelled the epistemic fallacy. This fallacy consists in the view that statements about being can always be reduced to, or analysed solely in terms of, statements about knowledge, that matters of ontology can always be translated into epistemological terms. As with Hume, such a stance ultimately cannot avoid ontology but succeeds only in glossing an implicitly generated one (in Hume’s case of atomistic events and closed systems) and so an implicit realism (here empirical realism). It also conceals a deep-seated © 1997 Tony Lawson
anthropocentric bias, which might be termed the anthropic fallacy, that all being can be analysed in terms of some attribute of human being. Hume’s position encourages the view that being experienced, or being open to possible experience, is an essential feature of reality, as opposed to an accidental property of some things or phenomena. And a significant consequence is an inevitable inability to distinguish those conditions, such as the experimental situation examined above, under which experience is in fact significant in science.
TOWARDS A DISTINCTIVE SCIENCE OF SOCIETY AND ECONOMY The second aspect of the analysis that warrants (re)emphasis is that the transcendental realist perspective that I have been defending has been formulated at an obviously rather high level of generality or abstraction. Although analysis at this level has been found to be sufficient to undermine the deductivist project that characterises contemporary economics, the perspective set out is insufficiently concrete to illuminate very much about the character of potentially fruitful research in the social realm. In order to say significantly more the specific nature of the fabric of society and economy must be (further) assessed; it is necessary to consider questions of social ontology explicitly. This is a task I take up in Part III. Even at this stage, however, there are sufficient pointers in what has been explicitly covered to anticipate that although the possibility of social science is sustained, investigations of social objects will often require very different methods, techniques and criteria from those employed in the natural sciences. If, for example, knowledgeability presupposes that the social structures acted upon are enduring to a degree, the fact that such structures are dependent upon inherently transformative human agency suggests that they will be only relatively so, being faster, or more temporal-spatially restricted, than (most of) the objects of natural science. Economics, then, in a more obvious way than natural science perhaps, is necessarily historical-geographical. In addition, if by virtue of their dependency upon human agency, social structures must be continually reproduced, it follows that economics is concerned with inherently dynamic matters. The terminology of process is thus fundamental to social science. By this I do not mean (as some do, see e.g. Langlois, 1986) merely a sequence of events. Rather process denotes here the genesis, reproduction and decline of some structure mechanism or thing, the formation, reformation and decay of some entity in time. Clearly, if society is intrinsically dynamic, social science must be alert, and its methods tailored, to this condition. Furthermore, the dependence of social structure upon © 1997 Tony Lawson
intentional human agency also entails an inescapable hermeneutic moment in social science, one that (for reasons to be discussed in Chapter 15) may be of greater consequence than any similar or comparable moment in natural science. In addition, the impossibility of engineering, and the absence of spontaneously occurring, closed social systems, necessitates a reliance upon non-predictive, purely explanatory, criteria of theory development and assessment in the social sciences. Such considerations as these already indicate that many social scientific methods and ways of proceeding will be irreducible to those of the sciences of nature. These are matters to be developed below in due course. The primary conclusion here, though, is that despite peculiarities of social phenomena, the (knowable) structured and open conditions of the social realm appear to allow a project of investigating social forms that can be scientific in the sense of natural science. However, whether or not, or the degree to which, a successful social science is achieved, of course, is down to the local conditions and the activities of those, including economists, who remain concerned with investigating substantive social issues.
© 1997 Tony Lawson
4 THE LEGACY OF POSITIVISM
We are now in a position both to explain the recurrent problems of contemporary economics, including theory/practice inconsistencies of the sort discussed earlier, and also to assess how the various difficulties of the discipline can be resolved. In fact it is easy enough to see that the problems reviewed in Chapter 1 all turn upon an uncritical acceptance of certain results of positivism, and in particular the constant-conjunction conception of scientific laws upon which the deductivist account of explanation rests. Given a social world that is fundamentally open, this orientation was always going to cause problems.
THE FAILURE OF ECONOMICS AT THE LEVEL OF RESULTS Most immediately, because scientifically interesting event regularities seem rarely to occur in the social realm, and are certainly not ubiquitous, it is hardly surprising that economic forecasters (econometricians) do not forecast particularly accurately, and axiomatic deductive explainers (‘theorists’) are unable to illuminate. The import of this recognition, however, is not that social explanation is thereby impossible. Rather, it is that we must embrace a very different conception of explanation to the deductivist covering-law model. Specifically, social explanation, appropriately conceived, is not the attempted deduction of events from sets of individual conditions and constant-conjunction ‘laws’, but the identification and illumination of structures and/or mechanisms responsible for producing, or facilitating, social phenomena of interest. We will see in due course that such knowledge is not only achievable but quite sufficient for purposes of policy formulation and the like.
CONFUSION AT THE LEVEL OF METHOD How are the various observed theory/practice inconsistencies to be explained? Consider the first noted inconsistency at the level of method. The obvious © 1997 Tony Lawson
reason economists should so often act inconsistently with their official ‘methodologies’ is that they more or less have to if they are to hold any hope of illuminating social phenomena, of producing relevant economic research. For standard, officially subscribed to, methods and techniques are wielded on the implicit presumption of the availability, indeed ubiquity, of constant conjunctions, when the social world is open and hardly susceptible to local closure. Needless to say there can be no guarantee that unofficial practices will be successful. However, given that econometricians often appear to know prior to any formal analysis when certain sets of projections or estimates are at least plausible, it appears that some insights are, in some manner and at some stage, being gained, that on occasion economists must be allowing for the degree of openness and structured nature of the social system. But what explains the persistence of the official stance? There are doubtless numerous reasons for it. In some cases it will be sheer opportunism. Given the significant pressure within the academy to conform, to go with the mainstream, some economists may well feel obliged to submit to rhetoric, if not actual techniques and practices, rooted in the dominant post-Humean paradigm. Some of it will be simply self-misrepresentation. This may be significant even if, on occasion, some economists with a mainstream affiliation are being insightful in their research and policy formulations. For although it is reasonable to suppose that on such occasions these researchers are acting on something like the transcendental realist perspective defended here, it cannot be assumed that they are aware that they are. And they need not be doing so continuously or consistently. Whatever else is the case economics cannot be defined as what economists claim to do, or necessarily (especially when they are endeavouring to act upon positivist injunctions) what they in fact do. In the light of tendencies and counter-tendencies at play, indeed, it is hardly surprising to find that the methodological perspectives to which allegiances are claimed are not, but may be held to be, the real bases upon which practices are formulated, including any through which insights and breakthroughs are made. Mostly, though, the explanation is simply an unwillingness even to question certain fundamental methods regarded as proper, an orientation turning on the continuing neglect of ontological enquiry. Such is the prestige of deductivist methods for some, a prestige founded upon an erroneous perception of the generality of their successes in the natural realm, that the record of failure so far in the social realm does little to dent this endeavour to persevere with them, or to appear to do so, or at least to acknowledge the correctness of doing so. It is precisely this dominant attitude of course that I am attempting to challenge with this book. But, whatever the explanation, from the perspective opened up it is quite explicable (if nevertheless problematic) that the methods and activities which economists actually and reportedly follow should in fact diverge significantly. © 1997 Tony Lawson
PROBLEMS IN SOCIAL THEORY How are we to account for the further point of tension noted in Chapter 1, namely an apparent inability or reluctance of mainstream economists to accommodate within their theories and ‘models’ their explicitly acknowledged intuition that people exercise real choice? The explanation can be seen to depend once more upon the mainstream adherence to deductivism and in particular its associated conception of scientific laws. Of course a reliance on any given set of methods entails that limits are put upon the sorts of substantive claims that can be officially entertained and practically ‘applied’. If the methods sanctioned are determined a priori (and in particular ignoring insights from ontology) it is likely that the limits implied will be sufficiently misunderstood and/or unexamined as to lead to unnecessary confusions in substantive enquiry. And it is precisely the prior commitment to deductivist method, and in particular the associated conception of ‘laws’, that necessitates that real choice is treated, in the end, as illusory. For, on accepting the deductivist framework, and so the restriction that economic behaviour be explicable via the constant conjunction form, it cannot be allowed that if a person does y under conditions x the person in question could really not have done y under these conditions, i.e. that y is in fact chosen. The acceptance of certain positivistic results in the social realm leads necessarily to substantive conceptions in which human choice is foreclosed.1 It should be clear that on accepting the framework of transcendental realism problems of this sort do not arise; the status of human choice is preserved. By acknowledging a realm of structures and mechanisms which are irreducible to actual phenomena including human activities, but which govern, facilitate, produce and/or condition them, the determinism of positivism is avoided. Thus while structures of language facilitate speech acts they do not determine what is said. Space for real choice is retained. Similarly the highway code (as with natural mechanisms such as gravity) facilitates safe driving without determining the journey taken; the market mechanism facilitates buying and selling without necessitating any purchase. The positivist account of the structure of scientific laws and the deductivist mode of explanation which it grounds, then, unlike the transcendental realist conception, cannot meaningfully accommodate the reality of human choice. In fact, the positivistic account of science encourages a conception of human beings as merely passive. This follows straightforwardly from positivism’s claims as a specific theory of knowledge. For, just as any theory of knowledge presupposes an ontology (which, in the case of positivism, consists in atomistic events given in experience) so it also presupposes a social theory, i.e. some account of human agency and institutions. Specifically, these must © 1997 Tony Lawson
be of a form to enable knowledge of the specified type to be achieved. Positivism, then, supports a conception of human agents as passive sensors of atomistic events and recorders of their constant conjunctions. And just as the positivist conception of science is uncritically accepted in much of contemporary economics so the associated specification of the human agent as the passive receptor of atomistic events goes relatively unchallenged. In economics this association of the positivist conception of the structure of science with a portrayal of individuals as merely passive, is probably achieved by way of resonance as much as anything else. But however that may be, from the vantage point gained it is easy enough to make sense of the observation, recorded in Chapter 1, that knowledge acquisition, as represented in mainstream economics, is little more than the recording of atomistic events or states of affairs. As we have also already noted, the events or states of affairs deemed relevant to economic behaviour are interpreted as (price) ‘signals’ to which agents are usually assumed to respond in optimising fashion. Although some variations on these themes can be found, there appear to be none that undermine the basic conception of agents as automata with knowledge analysed in a purely individualistic way. The sort of social theory that can be associated with the transcendental realist conception is necessarily very different. This claim will be developed in due course. But already our reconsideration of the activity of experimental control has indicated that human capabilities and institutions must be recognised as of a kind to facilitate the manipulation of aspects of reality—so that the latter may be more readily assessed/ revealed. Certainly, from this transcendental realist perspective, and in contrast to the Humean reduction of knowledge to impression, it is possible to explain the existence and reproduction of scientific institutions, including laboratories and processes of training, as well as scientific activity and, in particular, scientific work.
A CONTRADICTORY ORIENTATION TO METHODOLOGY I turn, finally, to the noted incongruity between official and actual stances on methodology. The task here is to make sense of the apparently paradoxical situation wherein economists, and particularly those in the mainstream, feel the need to denigrate explicit methodological/philoso-phical reasoning while seemingly being unable to refrain from perpetual methodological reflection and commentary themselves. Consider first why it should be that, despite an often avowed opposition to methodological enquiry, orthodox economists are continually drawn back to it. An obvious reason (if among many) is the repeated failure of officially © 1997 Tony Lawson
sanctioned methods. And the explanation of this failure is precisely that all such accepted methods presuppose, for their validity and widespread usage, the ubiquity of spontaneous event regularities, while significant social regularities of the sought-after kind have yet to be discovered. Once more the uncritical adherence to deductivism forces orthodox economists to act against themselves. In turn, of course, a continuing refusal to question the basic belief that the identification of event regularities (or the construction of hypothetical ones) remains essential to science, ensures that any revised methods or approaches are also found always to fail. Such failures then serve to reinforce a priori dispositions, allowing economists to emphasise ex posteriori with Hahn (1992a, 1992b) that methodology has been found to be fruitless in practice. The real failure, however, is that such methodological enquiry as occurs within economics, certainly by those working within the mainstream, continually neglects explicit ontological investigation, and indeed merely takes the deductivist framework as given. Certain premises, it seems, are always, if perhaps unconsciously, treated as sacrosanct. It is thus opportune at this point to recall Whitehead’s (1926) assessment, being mindful, in particular, of the unexamined presupposition of contemporary economics that the results of science, including economics, necessarily accord with the constant conjunction form: When you are criticising the philosophy of an epoch, do not chiefly direct your attention to those intellectual positions which its exponents feel it necessary to defend. There will be some fundamental assumptions which adherents of all the variant systems within the epoch unconsciously presuppose. Such assumptions appear so obvious that people do not know what they are assuming because no other way of putting things has ever occurred to them. With these assumptions a certain limited number of types of philosophic systems are possible, and this group of systems constitutes the philosophy of the epoch. (Whitehead, 1926:69) Once we break from positivist influence, and in doing so acknowledge explicitly an ontological domain, and specifically a realm of intransitive objects which exist independently (at least in part) of the scientific enquiry of which they are the objects, the noted problem of methodology dissolves. Methodological analysis can then be seen to possess the potential for making a critical and fruitful contribution to the conduct of science, including economics. Of course, this is precisely what I am endeavouring to achieve here. It is also a theme I take up in Chapter 5 below, where the process of knowledge acquisition sponsored by transcendental realism is expanded upon. © 1997 Tony Lawson
What, though, explains the orthodox hostility to methodology in the first place? No doubt numerous factors are significant. In Chapter 1 I noted that the mainstream opposition to methodology appears to be almost a matter of common sense, a stance that follows naturally in the light of the project accepted. Given that I have identified mainstream economics as deductivist and in so doing attributed its enduring problems to positivist influence, an obvious question to pose is whether the noted opposition to methodology is, in part at least, a positivist hang-over as well. I think it may well be. In Chapter 21 noted that any theory of knowledge, including the positivist or Humean account of impressions, presupposes an ontology (in positivism it is of atomistic events and their constant conjunctions). In the current chapter I have further observed that any theory of knowledge also presupposes a social theory (in positivism it is of people as passive sensors of atomistic events). We can now recognise yet further that any theory of knowledge also entails a theory of methodology. For each theory of knowledge carries implications regarding the manner in which its characteristic results are achieved. In the positivist case the implicit theory of methodology is clear. Because reality is effectively defined as that which is given in experience, all experience is held to be certain. In consequence, the knowledge process is necessarily viewed as monistic, the accumulation of incorrigible facts. The upshot is that there is little scope for scientific, or philosophical, criticism of any kind. The status quo is always preserved. The result is a conservative ideology which serves to rationalise contemporary orthodox practice; a perspective notoriously expressed within positivism itself precisely as a generalised denial of the usefulness of methodology/philosophy. I do not want to suggest that reasoning of this sort is consciously followed by economists. I have already suggested that it is more likely that once certain central strategic results of positivism are accepted, such as its conception of ‘causal laws’, other results of that framework may be taken on board by way of resonance. But whatever the mechanism, we can now see that the noted untenable attitude to methodology ultimately stems (as in the end do all the identified problems) from a generalised neglect of ontology. The basic mistake, one which facilitates many others, is a failure to elaborate in an explicit fashion the nature of the objects of enquiry, the intransitive objects of social science. The majority of mainstream economists, rather, come to their researches with a set of theories, procedures and techniques at hand, and the question of whether the intended objects of analysis are of a sort, or nature, to lend themselves to investigation by way of such methods is rarely put. In consequence, without a clear conception of the intransitive objects of analysis, scientific criticism becomes unintelligible and philosophical criticism of scientific practice impossible, so that, once more, philosophical reflection can end up serving no purpose other than, through self-denial, sustaining the status quo. In this economists are committing their own scientistic fallacy. © 1997 Tony Lawson
In brief, the predominant result of the neglect of ontology in recent economic theorising has turned out to be an implicit and uncritical reliance upon various results of positivism. And all the noted failings and tensions of the contemporary orthodox economics project appear, one way or another, to stem from its being rooted in the results of this misguided perspective, and especially from an adherence to a conception of the structure of science, or its ‘laws’, which that perspective grounds. It is a conception which, in the social domain particularly, is of only limited relevance. The way forward lies in rejecting the positivist theory along with the deductivist mode of reasoning which it conditions, and accepting the transcendental realist perspective on science and explanation in its place. Whether such a strategy proves sufficiently fruitful in practice is, as I have already acknowledged, down to the labours of social investigators, including economists, amongst others. But, whatever else is at stake, this strategy does at least hold the advantage of avoiding the anomalies of the mainstream project, anomalies—or inconsistencies—which, from the perspective now set out, are found to be quite comprehensible.
© 1997 Tony Lawson
5 THE NATURE OF THE ARGUMENT
The persistent failures and incoherencies of modern economics have been explained and a potentially fruitful way forward identified. And this result has been achieved primarily by way of methodological/philosophical enquiry. This warrants emphasis. For it appears that even among methodologists and meta-methodologists there is a consensus developing that methodological enquiry can achieve no such thing. A recently dominant view is that normative methodology is necessarily foundationalist and thereby arbitrary and unwarranted. Any critical input from methodology/philosophy is supposed to be undesirably imposing,1 or even impossible.2 Instead, this line of reasoning goes, methodology should take the form of (or be replaced by) something like an endeavour merely to describe the practices, rhetoric, or contexts of contemporary economists. Certainly a characteristic of much recent writing of a methodological kind in economics is a relative absence of criticism of economists’ actual practices.3 Against this background it is important to recognise that, despite the obviously normative character of the argument which I have been making, no such foundationalist position is taken. Because there is likely to be some doubt about this, however, I want now to retrace my steps to some extent and consider the nature of the argument in some detail. I intend to explain further the reasons for taking the specific path I have followed and to indicate the status of the results obtained. Up to this point I have skipped over some of the philosophical detail and context in order to convey more straightforwardly the basic argument. However, I am aware that I have left some of the steps only briefly justified and certain aspects more or less unaddressed. This latter set of omissions most obviously includes any explanation of the label transcendental in distinguishing the perspective defended. This chapter, then, also constitutes an attempt to fill in the details by going through the argument once more. Such an undertaking should clarify how the various threads of the argument tie together.4 © 1997 Tony Lawson
THE CASE AGAINST NORMATIVE METHODOLOGY First of all, why is it supposed by some meta-methodologists that methodology could not have a critical input into the doing of economics? In stark terms the argument runs roughly as follows. If methodologists presume to take as premises arguments from ‘outside’ or ‘above’ the discipline in question, their input must necessarily be irrelevant—or, equivalently, if not irrelevant then not external after all. 5 If, however, methodologists take, as premises, arguments, aspects or claims which are internal to the discipline, then they are necessarily taking sides with, or at least (and probably arbitrarily) making judgements on, some substantive contribution and can have nothing extra to add.6 When stated in this way it is easy enough to see that both strands of the argument must be wrong. In regard to the first strand, methodological suggestion of consequence for economics can certainly derive from an investigation of determinate, historically specific, social forms including practices, and yet emanate from outside of economics.7 However, it is essential to recognise in this the modal status of external results, criteria and comparisons. This tends to be missed by those who deny any feasible role for methodology which draws on ‘external’ results. Certainly, there is no basis for supposing that results or orientations of one domain can be automatically and unthinkingly applied in another. But there can still be a role for such results in framing the questions which are put, in pointing to the sorts of practices, etc., whose possibility may be worth investigating. We reason by analogy from one sphere to others all the time. In science, ideas and insights are continually abducted from one field to another. Where, for example, a new symptom of physiological disorder is identified it is not unusual for medical researchers to hypothesise a virus to be the cause on the basis of previous experiences in different domains. But to do so marks the initiation of an investigation, not its conclusion. It indicates one possible direction of research, one that has proven successful in other contexts. The same can be as true of scientific practices as it is of theories. If certain practices judged as scientific—say experimental practices—have been found to be illuminating in one sphere, it makes sense to investigate the possibility of similar successes with such practices is some prima facie different sphere. Assuming, then, that there is a role for philosophy in any domain at all (a contested point to be considered below) there is no a priori reason for supposing that results established in one sphere cannot be useful in framing possibilities for another—no a priori reason, that is, to accept Weintraub’s assertion that ‘there is no position totally apart from the doing of economics which can inform the consideration of the doing of economics’ (Weintraub, 1989:486). In short, the legitimate rejection of normative © 1997 Tony Lawson
methodologism, i.e. of merely asserting the value to one realm of a methodology that is (assumed to be) of widespread relevance to a different one, does not entail the impossibility of any legitimate and fruitful philosophical influence across scientific domains. There remains the second strand of the argument against normative methodology. The question here is how methodology or philosophy can legitimately or fruitfully adopt premises internal to a discipline or subject area? Specifically, how can it avoid merely siding with one or other theoretical position, with nothing of its own to add? It is essential to recognise here that the alternative to an explicit philosophy of science is not an absence of philosophy. Rather, it is an implicit and often bad philosophy. And, whatever may be the theoretical or substantive orientations of contemporary economists (whether econometricians, axiomatic-deductive theorists, hermeneuticists, and so on) their practices are all underpinned or informed by (competing) science-oriented philosophies of some sort. Of course, much of this is often tacit or unacknowledged, and it may be in contradiction with other beliefs. But it is precisely because of this that philosophical analysis can go to work. There is always the possibility that explicit methodological investigation of scientific practice, or other social forms, can make a contribution by rendering explicit some knowledge that is already implicit but unrecognised, and perhaps, in the reporting of economists (or whoever), openly contradicted. As Kant argued it is a function of philosophy to analyse concepts which are already given but confused. We have already seen how confusion is rife in the contemporary discipline of economics where theory/ practice inconsistencies abound. A fundamental objective for philosophy, then, is immanent critique. Specifically, philosophy can expose internal incoherences in beliefs implicit in practices, or it can demonstrate how certain beliefs cannot in fact accommodate something which, through practice, is actually being done. In short, philosophy as a critical rational activity is always a real possibility.
A QUESTION OF SCIENCE If the potential for philosophy to contribute to scientific practice is something that cannot reasonably be denied, the question that I am interested in here, and have been addressing in the previous chapters, is whether it is possible to go further and use philosophy to investigate the possibility of economics as social science. Specifically, is it possible to elaborate an adequate conception of science, or scientific method, and use it to assess the scientific possibilities for, or scientificity of, actual economic practice? In pursuing this line of questioning I am persuaded not only that the current crisis in economics © 1997 Tony Lawson
urgently requires a reconsideration of the question of naturalism, i.e. of whether social phenomena are susceptible to explanation in essentially the same sense as are natural phenomena (that is, scientifically), but also that recent developments in the philosophy of science allow such a reconsideration. It appears to be precisely at this point, in pursuing this line of questioning, however, that I part company with many, and perhaps even the majority of, economic methodologists and meta-methodologists. For this group of people the attempt to distinguish science has apparently been largely abandoned. The reasons for this are varied. On occasion, a few, like McCloskey (1986) and Mirowski (1989), seem even to adopt an a priori stance, being dismissive of any attempt to delineate science as an (apparently necessarily positivist or ‘modernist’) endeavour inevitably bound up with an attitude of arrogance: In practice methodology serves chiefly to demarcate Us from Them, demarcating science from nonscience. Once the modernists have founded a Bantustan for nonsciences such as astrology, psychoanalysis, acupuncture, nutritional medicine, Marxist economics, spoonbending, or anything else they do not wish to discuss, they can get on with the business at hand with a clear head. Methodology and its corollary, the Demarcation Problem (What is Science? How is It to be distinguished from nonscience?), are ways of stopping conversation by limiting conversation to people on our side of the demarcation line. The replies to such scepticism about the uses of methodology and epistemology have been unpersuasive. Indeed, it has not usually been thought necessary to stoop so low as persuasion. (McCloskey, 1986:26) Without exaggeration, one may say that there has been a surfeit of breastbeating, wailing, crowing, and soul-searching over the question of whether economics is a science. In the modern era, this has mainly served as a prelude to a round of smug satisfaction and self-congratulations...; Let us begin by admitting that there exists no scientific method, no set of timeless criteria that the program of economic research could adopt and embrace in order to guarantee its own scientific status and legitimacy. This lesson is the legacy of the decline of positivist philosophies of science in the late twentieth century. Juxtapose this fact with the hypothesis that economic research has always met with the greatest difficulties in establishing the credibility of its results and in fending off charges of charlatanism and quackery. Consequently, the pressure to usurp the legitimacy of science has always weighed down economic research; further, there is no sign on the horizon that this weight will be lifted any time in the foreseeable future. (Mirowski, 1989:356, 357) © 1997 Tony Lawson
However, resistance to the sort of strategy that I am proposing also emanates even from many methodologists who are prepared in principle, and even keen, to advocate a more clearly evaluative stance towards economic practice. An example is provided by Caldwell’s (1989) methodological, or critical, pluralism, which explicitly accepts the task of evaluating the ‘methodological writings of economists’ (1989:42, 43). Caldwell is rather uncommitted about how the evaluations might be carried out: ‘There is a variety of criteria of assessment; there are many ways to evaluate methodologies’ (ibid.: 45). And a few lines on he adds: ‘The “pluralism” in critical pluralism shows up, then, in the plurality of methodologies that are evaluated and in the plurality of criteria that are used to assess them’ (ibid.: 45). But one strategy that he does appear to rule out firmly is precisely the one that I am interested in: of elaborating a conception of science and using it to assess the scientific possibilities for economics (ibid.: 44, 45).8 It seems that the main reason for the recent popularity of the antidemarcationist stance (at least amongst those who adopt a positive orientation to the idea of a critical evaluation of practices regarded as scientific) is an ex posteriori recognition that previous attempts to distinguish science, including some very influential endeavours, have failed. In truth, in some quarters at least, the question of demarcating science from non-science has underpinned much of the discussion and debate amongst economic methodologists for the past ten to twenty years (see for example the useful discussions in Backhouse, 1994; Blaug, 1980, 1994; Caldwell, 1982, 1988a, 1989; Hands, 1985, 1993a, 1993b; Hausman, 1989, 1992 and Mäki, 1993a). Central to much of this has been Popper’s falsificationist criterion for distinguishing science,9 and the epistemological issues which this raised.10 And it is because Popper’s attempt in particular to distinguish science from non-science is now recognised to have failed, being inconsistent with the actual practices of successful science,11 that the demarcation question appears recently to have been so widely abandoned. Caldwell, in fact, acknowledges explicitly that it is just because of this failure that his particular position is put forward: ‘Had [such attempts] been successful, there would be no need for critical pluralism’ (Caldwell, 1989:44). Yet, for Caldwell and others like him who talk regularly of science12 and/ or of the scientific community13 there arises an obvious dilemma. For either there are real differences between the practices of scientists on the one hand, and those of TV watchers, sun-bathers and those who chew gum, by virtue of which the label scientific is applicable to the former and not (typically) to the latter, or there are no grounds for consistently using the term scientist to designate one rather than another group of people. On this point Popperians like Blaug (1994:110) are correct. To persist with talk of science, but to reject © 1997 Tony Lawson
the slide to nominalism, is to concede the case for a real definition. The latter, I suggest, is ultimately what Caldwell (and most others) must do. In short, it is no more the case that providing useful methodological analysis can be ruled out a priori than that it can be inferred that pursuing an account of science is necessarily doomed to failure. If previous attempts at either endeavour have proved to be unsuccessful or even debilitating so far, the only compelling conclusion is that we should attempt to learn from such failures; not that we should abandon entirely the respective enterprises involved. Certainly, I know of no argument to the contrary that can be found in the writings of those in disagreement with this conclusion. Rather the supposition, sometimes stated but usually implicit, is that failure and/or misapplication so far, entails failure and/or misapplication in all cases. This, I am attempting to demonstrate, is not a conclusion that follows in practice, any more than it does in logic.
REALISM AND SCIENCE Once the case for a real definition of science is conceded and the task of pursuing it is accepted, there are conceivably many routes or approaches open. The account of science that I have found most compelling and have been putting forward in the preceding chapters is a realist one. A scientific realism, let me reiterate, is a theory that the ultimate objects of scientific enquiry exist and act (mostly) quite independently of scientists and their practices. And the particular conception that I am arguing for is a metaphysical realism, consisting in an account of what the world must be like before it is investigated by science, and for scientific activities to be possible. Such a realism neither turns on, nor endorses, a realistic interpretation of any substantive theory. But, once obtained, it does provide insight as to how science can be done. Clearly the possibility of such a metaphysical realism introduces explicitly the issue of a philosophical ontology—theories or questions about the nature of being and existence explicitly enter the picture. But, of course, questions or issues of (philosophical) ontology arise with every commentary upon science. Every philosophy presupposes a reality. Anyone who seeks to deny, or to banish, ontology only succeeds in smuggling it in illicitly. We have seen that those such as Hume who maintain that all we can have are experiences or impressions presuppose a (knowable) world of atomistic events given in experience. But this reliance on an implicit ontology is equally true of ‘post-modernists’ and others who wish to imply that knowledge is mere human construction. Thus, those who would reduce science to conversation presuppose objects of that conversation. Even if © 1997 Tony Lawson
they claim that conversation is only about conversation they thereby introduce a distinction between conversation as an (intransitive) object of knowledge, and conversation as a (transitive) form or medium of knowledge. Similarly, the attempted reduction of methodology to (or its attempted replacement by) history of thought or naive descriptivism, presupposes a reality of economic thought, practices, conversations, or whatever it is that is being re-told or simply described. Realism, then, is ineluctable; all theories of knowledge, science oriented philosophies, social practices, etc., presuppose an ontology, i.e. some conception of the nature of reality. Equally, though, contending theories of ontology can sponsor quite different conceptions of science. The important question is how best to proceed in elaborating a sustainable ontology for, or metaphysics of, science. In pursuing this question in the previous chapters I have been significantly influenced by Bhaskar’s (1978) elaboration of his ‘transcendental realist’ position. Starting from premises concerning experimental practices in science Bhaskar derives an account of a metaphysics of science through investigating the intelligibility of such practices. That is, Bhaskar proceeds by enquiring what the world must be like for experimental procedures and practices to be possible.14 Such a mode of enquiry is referred to as transcendental. This explains the use of the term for the conception of reality so derived. Because various misconceptions of the nature of this sort of enquiry can arise, however, I want to trace out explicitly the various steps involved. Recently the perspective systematised as transcendental realism and defended here has been gaining some attention in economics. But it is equally clear (especially in its negative reception in some quarters) that the nature of ‘transcendental enquiry’ is frequently misunderstood, as is the status of any results thereby established. Let me then, if briefly, indicate essential aspects of the mode of reasoning in question, as well as the standing of results obtained.
TRANSCENDENTAL ANALYSIS The argument used to establish a transcendental realist metaphysics of science constitutes a particular example of a general mode or method of reasoning that takes the form of an enquiry into the conditions of the possibility of some central, especially significant, or pervasive, feature of our experience. Now if, like Bhaskar, we follow Kant in referring to the latter general mode of enquiry as transcendental, it warrants emphasis that there is nothing inherent in the method which forces us to accept the individualist and/or idealist mode in which Kant framed his own enquiries. Several points require elaboration here. First, the features of experience to be accounted for and the way they are © 1997 Tony Lawson
philosophically conceptualised (i.e. the premises for the transcendental inference) may each be corrigible, contested, historically transient and spatially highly restricted. Second, any human activity which is to be so accounted for may depend on an array of powers that people possess as physical, biological and social agents, and not merely those most directly implicated in thinking and perceiving. Third, the analysis itself, which in the manner of its premises may be normatively corrigible and contested, may give rise to results that are realist, not idealist, and so epistemically relativist, not absolutist; and it may support conclusions which may only be domainspecific. On this account of transcendental analysis, then, both premises and conclusions of philosophical argument necessarily constitute contingent claims. If in the manner suggested it turns out that philosophy can so establish synthetic a priori knowledge (of the sort we are hoping to achieve) it can only be in the relative or conditional sense noted. That is, the application of philosophy (in the form of transcendental analysis) necessitates prior conceptualisations of determinate social forms, while any results must be interpreted as conditional and expressed in hypothetical terms. On this conception, transcendental enquiry does not exist independently of the various sciences and other social practices, nor is it used to investigate a different world: philosophy treats the very same world as science. And on Bhaskar’s conception, and perhaps generally, philosophy investigates that world transcendentally; it aims to reveal what our conceptions of certain generalised practices presuppose about that world. It represents, to repeat, an investigation that necessarily takes contingent historical premises and specific social conditions, and aims to produce hypothetical and conditional conclusions; an investigation which can never be foreclosed, it is always open to elaboration and transformation. On this conception, there is no philosophy in general, only philosophy of particular, geo-historically determinate, social forms. A question of obvious significance here is how the premises for a transcendental enquiry are to be selected without implying an arbitrary, ‘external’, unjustified, or otherwise unpersuasive commitment to the epistemic (or other) worth of the practices described? Science, or research activity, never takes place in a philosophical vacuum. All science-related practices are informed by one or more prevailing philosophies of science. Transcendental enquiry, as with philosophy generally, takes the form of immanent critique. The premises chosen for transcendental analysis should be acceptable to the investigator (if conditional synthetic a priori knowledge of possibilities for science is the aim). But at least as fundamentally the premises should be acceptable to, and if possible of paramount importance to, proponents of
© 1997 Tony Lawson
doubtful conceptions of philosophy currently dominant, or at least in contention. In other words, the premises of transcendental arguments will be descriptions of practices given prominence by influential, or potentially influential, science-oriented philosophies. In this way successful transcendental arguments will constitute transcendental refutations (or confirmations) of pre-existing accounts of science. Thus conceived, transcendental analysis is ultimately a process of determinate negation. The target will be a (dubious but influential) science-oriented philosophy which focuses on some more or less inadequately analysed activity. The aim is first to render explicit the ontological, sociological and other presuppositions of the science-oriented philosophy, and hence set it in the form of a scienceoriented realism, sociology, and so on. The analysis is completed through demonstrating how, when adequately analysed, the activity in question is seen to presuppose a transcendental realist conception incompatible with that of the science-oriented philosophy thus critiqued.
EXPLAINING EXPERIMENTAL ACTIVITY The argument set out in Chapter 3 is just such a determinate negation. The basic question posed is: how are experiments possible? The philosophical context is one in which experimentation is heavily sponsored in Kantianism (or transcendental idealism) and empiricism. For such traditions, regularities of the form ‘whenever event (type) x then event (type) y’ are at least necessary (transcendent idealism) and perhaps sufficient (empiricism) conditions for causal laws (and other features of knowledge). And these traditions have seized upon the experimental establishment, knowledge and practical application of such ‘laws’ as being of fundamental epistemic significance in science. Bhaskar’s substantial contribution is to demonstrate that a condition of the intelligibility of the experimental generation and practical application of scientific knowledge is that the objects of science are real structures and mechanisms which exist and act independently of the patterns of events they govern. And on the basis of this analysis a rational account of scientific development, as we have seen in Chapter 2, is easily stated. In this light, we can see how philosophy of science has the potential to provide a directional input into the practices of science. For although our analysis suggests that when scientists are practising science they are implicitly acting upon something like transcendental realism, it does not follow that transcendental realism, or any other philosophy, is always or consistently acted upon, or dominant, or even acknowledged. It is for this reason that Bhaskar (1979) is able to conclude that ‘one is...qua philosopher of science, at © 1997 Tony Lawson
perfect liberty to criticise the practice of any science for its lack of scientificity’ (1979:15). Nothing in the foregoing should be taken to imply that philosophy can do the actual work of science for it. If the elaboration of a transcendental realist perspective provides grounds for supposing that science can successfully uncover structures and mechanisms that govern some identified phenomenon of interest, philosophy cannot do the work of uncovering. This is the task of science alone. Philosophy, however, is able to make a difference to science in the manner noted: by, amongst other things, affecting the questions put to reality, and the manner in which this is done. Given, moreover, the disarray, confusion, and uncertainty that abounds in modern economic research, any critical rational philosophical ground-work or prompting of this sort seems particularly likely to bear fruit.
THE POSSIBILITY OF ECONOMICS AS SCIENCE How is the foregoing realist analysis of natural science used to assess the possibilities of, and/or for, social science including economics? The central result utilized in the preceding chapters is that the essence of science lies in the move, at any one level, from some manifest phenomenon to the structures which generate it. It follows that a questioning of the possibility of naturalism amounts to investigating the feasibility of making a comparable move in illuminating the objects of the social realm. From this perspective we can see immediately that previous (positivist) injunctions from the philosophy of natural science have been unsound. For, not only has the possibility of naturalism been begged at the outset, but also (natural) scientific methods have been inadequately, or misleadingly, expressed. Specifically, the conception of science which underpins naturalistically inclined contemporary orthodox economics, which is reaffirmed for economics in continuing positivist injunctions, and which is reacted against in the explicitly anti-naturalistic alternative provided by subjectivists, hermeneuticists and others (see Chapter 13), depends upon the popular constant-event-conjunction formulation. However, it has been found that the intelligibility of the experimental establishment and application of scientific knowledge presupposes that the objects of science are real structures and mechanisms which exist and act independently of the events they co-determine, so that causal laws must be analysed as tendencies, which are manifest as event regularities (including probabilistic ones) only under relatively unusual, typically intentionally engineered, conditions. Thus, contrary to specific claims of Popper, Hempel, Blaug and Hutchison and the tacit presuppositions of anti-naturalists like Hayek and Shackle,15 deducibility from event regularities is neither sufficient nor even necessary © 1997 Tony Lawson
for natural scientific explanation. There is an ontological gap between a causal law and its empirical grounds. In short, positivist methodological injunctions into economics are not only unwarranted but false in their conceptions of natural science, while the most familiar subjectivist or hermeneutical contrasts stand in need of reorientation and perhaps rejection (Sofianou, 1995; Runde 1997). Even so, it does not (yet) follow that seeking results involving the popular constant conjunctions of events is not a viable generalised goal for those investigating the objects of the social realm. For if it is a mistake to impose a priori a conception of natural science as necessarily relevant to investigating social objects, it must also be a mistake to suppose that a conception that is only of limited relevance to natural science is, by that reason alone, of little relevance to an investigation of the social realm. Whether or not it is of relevance can be determined only through exploring the possibilities for social investigation directly. We are though in a position to recognise that the premise for positivist incursions into the social realm, namely a belief that positivism provides an adequate conception of proper (natural) scientific method, is erroneous. And as this premise appears to be a dominant factor in explaining the desires of many economists to conform to the injunctions of Popperians and Lakatosians, etc., our analysis is not without significance. In any case, the ‘whenever event x then event y’ conception of natural science stands discredited. Certainly there are grounds for supposing that the transcendental realist conception of natural science facilitates a fairer comparison or standard when exploring links between the putative sciences of society and of nature. But, in the given context, that is essentially all that it achieves. To determine whether a social science, including economics, is possible it is necessary to look to the social realm directly. Why, though, have I been taking the route of questioning the possibility of naturalism at all? Why, in seeking to elaborate the possibilities for social investigation, is a conception of natural science even required? To be sure, such a conception helps debunk the sort of misconceptions of natural science which underpin positivistic injunctions. It is also true that taking this route is legitimate and carries the potential for bearing fruit. But still, in questioning what economists can usefully aim at, why not start out from premises concerning social investigative practices and elaborate a realist metaphysics of social science directly? Why take the detour of elaborating a metaphysics of natural science and then questioning the appropriateness of the conception of science thereby supported to understanding social life, even if addressing these questions is possible? Why not just employ the transcendental method exemplified by the analysis of the experimental situation, but leave aside the natural scientific results thereby obtained?
© 1997 Tony Lawson
One answer is that a priori there is no good reason to proceed other than on both fronts, or even to see them as but two strands of the same project. Certainly, the real possibilities for social analysis must be determined by considering aspects of the social domain directly. But purposefully to ignore a potentially relevant and useful set of results could be just as limiting as imposing a philosophy of natural science at the outset. There are, however, other, ex posteriori, reasons for taking the route in question which stem from the nature and situation of the contemporary economics discipline. Specifically, when we look to the practices of economists as a source of premises for transcendental enquiry, we find that there is no obvious set of widely accepted and acceptable practices that are suitable for our purposes. Note that the premises employed in our analysis of the natural science situation turn on aspects of experimental practices which are accepted by all the contesting parties. Within natural science, in other words, such practices are, where feasible, a part of the day-to-day procedures. Disputants of substantive debates all look to experimental work to check out some mechanisms empirically, or to confirm, or falsify, their own, or others’, positions. Equally, experimentation is sponsored, usually as an activity of paramount importance, by the dominant, post-Humean, (nontranscendental realist) philosophies of natural science: empiricism and Kantianism alike. In economics, by contrast, the various practices which abound (econometrics, axiomatic-deductive reasoning, single-exit modelling, the compositive method of reasoning, pattern modelling) not only are the focus of substantive theoretical controversy, but also presuppose different, and generally quite incompatible, conceptions of the economy and/ or society. Do we not, then, hit upon an insurmountable problem here? Even granted that Bhaskar’s analysis provides insights into the nature of scientific practice, what use to us are such results if we cannot examine the possibilities of proceeding in similar ways in the social domain? What advantage is there to the framing of interesting questions if we do not have the means of adequately investigating them? At this point should we not just follow Caldwell (1988a) and include all practices sponsored by people calling themselves economists in the economics’ (‘scientific’) melting-pot: How are we even to know what counts as ‘economics’ if we do not try to solve some sort of demarcation problem? Methodological pluralism looks to the practitioners of economics to see what economics is. Now, of course, the practitioners disagree among themselves about this. Even a casual observer of the discipline will discern a huge and amorphous mainstream that is surrounded on all sides by the heterodoxy, a group so diverse that it sometimes seems that the only bond among them is a © 1997 Tony Lawson
dissatisfaction with mainstream analysis. Which groups should be included? The pluralist answers: all of them. (Caldwell, 1988:235–6) With perhaps as much justification Caldwell might equally have answered: ‘none of them’. For not only do the most prominent practices of different traditions of economists conflict, and often in ways that presuppose incompatible accounts of social reality, but equally all the most prominent approaches are rather unsuccessful even on their own terms or criteria. Forecasters, as we have noted, do not forecast particularly successfully; axiomatic-deductive explainers do not explain anything of interest or value. Economic science, in short, cannot be reduced to what economists do.16 Against this background of failure, indeed, it seems not at all unreasonable to be questioning whether a social science of economics is even possible, and if it is possible, whether, in fact, the child is yet to be born. Anyway, the variety, and significantly contested nature, of contemporary practices renders it impossible to determine the scientific possibilities for economics, either by accepting all of them, or by focusing on any one of them, without immediately begging or abandoning the question at issue. How, then, is it possible to proceed? If we cannot start out from some agreement on (aspects of) social scientific activities or practices we must seek to identify, instead, some more or less universally recognised features of substantive social life itself, features which do not load the dice in favour of one or other conception of social scientific practice (or ‘economics’) at the outset. And it is because of this situation that a questioning of the possibility of naturalism, employing the transcendental realist conception of science sustained, provides not only a legitimate route but one that emerges as the most promising of the options available. However it seems now that we must commit ourselves to the content of an area of investigation, specifically the social realm, and particularly to the nature of (or, at least, aspects of) its object(s), i.e. society and/or economy, before deciding whether the objects of the relevant domain can in fact be studied scientifically? But this was always inevitable. The question ‘can there be a science, or indeed any form of investigation, of XYZ?’ cannot even be addressed without some prior conception of XYZ. Methodology is always going to be irrelevant without some prior account of a putative object of study. Methodology, to repeat, is necessarily practically conditioned. It exists in tandem with, not apart from, science or, more generally, human life. Its object is the same reality; its premises are always substantively rooted. In short, the situation focused upon in determining an account of (natural) science is quite different to the one that must be faced up to in the social realm. In the analysis of the experimental situation, the starting point has © 1997 Tony Lawson
been the insight that science exists and that it includes experimental activities amongst its practices. We questioned what the world must be like for such scientific practices to be possible. Here, in contrast, the actuality of a social science cannot be taken as given or uncontentious. Rather, we are questioning its possibility, and the best we can do is take the (currently) most adequate account of (natural) science as conditioning. This account, I have argued, is provided by Bhaskar’s transcendental realist metaphysics of (natural) science. The situation facing us, then, is that we have a conception of (natural) science at hand (obtained through reflection on natural scientific practices) and we wish to determine whether there exist objects of a particular domain of interest (here specifically the social domain) with features or properties which allow their being studied scientifically.
HUMAN INTENTIONAL AGENCY If ‘dependency upon human intentional agency’ can be accepted as criterial for phenomena to be regarded as social, then human intentional agency, along with the capacity of choice which is analytic to it, provides a convenient, widely accepted17 premise for initiating our enquiry into the possibility of social science. And as we have seen it is easy enough to show that the reality of choice presupposes that the world, both natural and social, is open. Immediately, then, we can confirm that the content of the familiar positivist injunctions into the social realm (including, in particular, the insistence that economists accept the universal validity of methods or modes of explanation that take the event regularity conception as at least a necessary component of science) are invalid as prescriptions appropriate to social research. If there are any persistent event regularities at all in the social realm—and complementing the above derivation from the phenomenon of choice is the ex posteriori experience of a complete absence of any scientifically significant ones (see e.g. Lucas, 1976)—they will most likely be intentional products like the regular re-occurrence of annual holidays, or of shopping hours. Economic orthodoxy, then, which is a barely disguised manifestation of positivist philosophy at play in economics, ultimately cannot sustain the reality of people exercising their power of choice even though the latter is an aspect of human behaviour that it claims explicitly to sponsor.18 But if the social world is open (in the sense that events, including actions, could always have been otherwise), we have still to establish that a social science is possible; this remains the central concern. Interpreted from the perspective of transcendental realism, the appropriate question to pursue, and which I addressed in Chapter 3, is thus whether there are social structures of some kind, i.e., social phenomena which facilitate, but are irreducible to © 1997 Tony Lawson
human actions and other actualities, which can be identified. For the answer it is necessary to look to substantive social investigations themselves. We can acknowledge once more that should social structures exist and endure, they will be distinguishable from purely natural ones by virtue of a dependency upon human agency. Thus conceived, a specific social structure, if such exists, is something which, if the human race were to disappear tomorrow, would ‘vanish’ along with it. But it must be left to the work of social research, including economics, not philosophy, to uncover such social structures as there are. If we turn to social investigations the obvious candidates for social structure, as we saw in Chapter 3, are rules, relationships, positions and the like, as components, and perhaps constitutive, of societies. Existing studies of all identified putative societies, whether industrial or not, couch their explanations in such terms. Such ‘societies’ are rule-governed in language use, games, modes of exchange, etc. They are typically structured along hierarchical lines, with positions (kingships, tribal-leaderships, teaching posts, studentships) occupied by specific individuals. And such positions are always defined in relation to others—teachers to students, tribal leaders to other tribal members, landlords to tenants, and so on. Notice that even most self-styled ‘methodological individualists’ do not flinch from making reference to (or at least from using the language of) social structures such as these in their contributions (see Hodgson, 1988 for an assessment). Rather, members of the latter group seem to be motivated by a felt need to distance themselves from others (whoever they may be) who apparently attribute emergent human properties, such as intentions and purposes, to some or other ‘aggregate’ or ‘collectivity’. Conceptions of social structures such as rules and relationships, then, figure widely in explanatory contexts. This recognition, however, is not yet sufficient to establish the autonomy or reality of the putative items expressed. For it may yet be suggested that the categories in question are merely ‘theoretical constructs’ or some such, used purely to organise our interpretations and understandings—of, say, the (routine) patterns of everyday life. In consequence, if a case is to be made for the possibility of naturalism, it is necessary to establish that such putative objects of scientific investigation as those in question are both irreducible and real. An essential point here, let us once more recall, is that science employs not only a perceptual but also a causal criterion for the ascription of reality to a posited object. The latter turns on the ability of an entity whose existence is in doubt to bring about changes in material things. It is this causal criterion, and not perceivability, that is satisfied by magnetic and gravitational fields, and which is grounded in Bhaskar’s analysis of experimental situations outlined earlier. The task, then, is to demonstrate that, but for society, or at least for © 1997 Tony Lawson
various social structures, certain physical conditions, including actions, would not be performed. To the extent that this proves possible we are justified in accepting the structures as real. And, as we saw in Chapter 3, this is easily achieved. Such (intentional) human activities as speaking, writing, driving on public roadways, cashing cheques, playing games and giving lectures would be impossible without rules of language, highway codes, banking systems, rules of play and teacher-student relationships. The latter are all structures which pre-exist and make a difference to (constrain as well as enable) the sort of activities which take place. Their pre-existence establishes their autonomy just as their making a difference establishes their reality. In the manner just elaborated, then, the real possibility of social science has been established. It has been achieved by way of a transcendental argument from intentional agency as dependent upon social material causes (so establishing the latter’s existence), which are available to be drawn upon prior to any individual act (so establishing their relative autonomy and rendering them susceptible to social investigation). That is, a transcendental argument from intentional agency (a premise accepted by the super-naturalist and anti-naturalist traditions in economics alike) establishes the sui generis reality and the temporal pre-existence of forms of social structure, as its necessary condition and means.
LIMITS AND LIMITATIONS I want to round off this chapter by standing back a little and briefly drawing out, or further underlining, the limits and limitations of what is being maintained. In the light of the emphasis taken in this chapter especially, the project advanced in this book may strike some as being insufficiently pluralist, or at least overly committed to too definite a direction of research. In case this should be so I want to conclude this part of the discussion by stressing once more the open-ended nature of both the philosophical project supported as well as the sort of economics it sponsors. First, although the emphasis throughout has been realist, and in particular I have allowed that things exist and act independently of their descriptions, this stance is complemented not by an absolutist or foundarionalist position in knowledge but by an epistemological relativism, by the thesis that we can only know these things under particular (historically and socially relative), and potentially transformable, descriptions. Clearly, if the (intransitive) objects of knowledge exist (largely) independently of our knowledge of them, such knowledge as we actually possess cannot be identified, or be said to be in correspondence, with such objects; it is not reducible or equivalent to them.
© 1997 Tony Lawson
Knowledge, rather, exists in a historically specific, symbolically mediated and expressed, practice-dependent, form. Thus many insights of the ‘sociology of scientific knowledge’ tradition (that scientific practice is a social process, set in a social context, with results that constitute a social product) can be accepted (see Mäki, 1993b, for a useful discussion). We can also acknowledge that pragmatic goals, including narrowly economic incentives, are often influential, an insight recently systematised by Hands (1994) and others under the heading of ‘the economics of scientific knowledge’. Even so, the recognition of a social and pragmatic dimension to all social being, including scientific practice, does not necessitate an ‘anything goes’ position, or anything like it. While a realist position such as sustained here acknowledges an inevitable role for purely pragmatic aspects and implicit criteria (and often explicit ones too, such as standards of rhetoric, simplicity, elegance, coherence), it does not necessitate a judgemental relativism in the process. Rather, the attachment of epistemological relativism to ontological realism facilitates a judgemental rationality. Second, the transcendental realist project itself is self-consciously a nonfoundationalist, fallible, transient, practically conditioned, social product. It is one that is partial in focus and open-ended in its development. It does not suppose disinterested states of, or for, any party to, or subject of, analysis. Indeed, its practical, and practically conditioned, character is reflected in an explicitly acknowledged dependence upon extra-philosophical interests and standpoints as premises. The latter are commitments which can be justified, if at all, only through the wider reach of historical-social experience. Nor does it suppose self-evidence (or simplicity, etc.) of premises, arguments or conclusions. Experimental practices have provided the entry point in deriving a conception of (natural) science because (among other things) these practices tend to be regarded as fundamental to the sciences of physics and chemistry. These are the usual contrasts in assessments of the scientificity of other putative sciences. But there is no pretence that the analysis thereby conducted is comprehensive or indeed anything more than a ‘first’ step. Clearly, the insights sustained need to be re-examined and extended through an enquiry that draws upon the range of sciences, in a process that can be expected to be reciprocally enriching. Equally, such analyses and comparisons might be extended to the realm of non-scientific activities. Certainly, there can be no presupposition that the sciences, let alone a sub-set of them, are exclusively capable of providing fruitful forms of contrast for social scientific activities such as economics. Rather, the sciences, or any sub-set, can be expected to provide one particular focus, perspective, angle, or slant on reality, albeit, in some cases, a significant one. To repeat, it is not being suggested that the transcendental realist perspective is infallible, complete, comprehensive or necessarily enduring. © 1997 Tony Lawson
However, in that it is uniquely consistent with the practical presuppositions and substantive contents of the currently most highly regarded and successful sciences, I think it can be accepted as the best account of science that is available. But that is all. Let me also note, parenthetically, that the use of the term transcendental in the heading used to describe the theory of ontology defended, is, in the light of the discussion of transcendental argument, both comprehensible and yet, I think, questionable. It is comprehensible precisely because the theory of ontology, the realist perspective in question, derives support from a particular transcendental argument from experience. It is questionable because of the particularity of the argument. For it is at least conceivable that alternative conceptions may eventually also derive support by way of transcendental argument from (the same or alternative) accepted features of experience, thus facilitating different, perhaps competing, accounts equally qualifying for the ascription of ‘transcendental realism’. A better label then would be one that captured the structured, intransitive, nature of the ontology elaborated above. Perhaps ‘structural realism’ or ‘transfactual realism’. However, I persist with the transcendental realism label here because it is now ingrained in the (fast growing) literature and because, for the time being at least, there is no comparably viable alternative theory of ontology on the scene. But this usage may have to be temporary. Of course, if competing ‘transcendental realist’ accounts are so derived they will have to be selected amongst on the basis of their relative explanatory powers, according to their relative capacities to illuminate a range of generalised features of experience, their abilities to withstand attempts at determinate negation, and so on. Third, if the results set out above support the possibility of naturalism, i.e. the thesis that the study of social objects can be scientific in the sense of natural science, the options opened up necessarily take a highly qualified, specifically non-reductionist and non-scientistic, form. That is, the conception defended neither asserts the identity of the subject matters of the natural and social realms (reductionism) nor denies (as does scientism) that there are, or may be, significant differences in methods appropriate to studying social and natural objects, differences which turn on the peculiarities of their respective objects. Even so, an account of science has been set out under which the legitimate, if more or less specific, methods of both natural and social sciences can fit. Finally, while philosophy of science or methodology has been shown to carry the potential to intervene, actively and productively, in the practices of science, it cannot do the job of science. If philosophical analysis such as presented above reveals the need for scientists to uncover structures that govern phenomena of interest, it cannot do that job of uncovering. It can ‘underlabour’ © 1997 Tony Lawson
for science19 in clearing up inconsistencies and confusions, and it can act as a ‘midwife’ in helping new sciences to emerge. But that is as far as it goes. Of course, if we must accept that philosophy is restricted to a servicing role, it warrants emphasis that within the contemporary discipline of economics there remains rather a lot of servicing to be done.20
© 1997 Tony Lawson
6 TOWARDS A RICHER ONTOLOGY
My objective in this book is not only to explain the disarray of contemporary economics but also to work towards an outline of a more viable alternative approach or project. In pursuing this latter goal I have found it necessary to turn to ontological enquiry. Indeed, it is evident that a fundamental reason for the widespread disarray in contemporary economics is precisely a generalised neglect of ontological enquiry. This neglect is itself underpinned by the epistemic fallacy, by the supposition that statements about being can always be rephrased as statements about knowledge (of being), that ontology can be reduced to epistemology. With ontology unavoidable this error functions only to cover the generation of an implicit ontology, in which the real is collapsed onto the actual which is then anthropocentrically identified with, or in terms of, human experience, measurement or some other human attribute.1 Such an ‘actualism’ thereby serves to deny the differentiation of the world, its depth, and the openness of the future.
A STRATIFIED REALITY In the more constructive part of the analysis the central emphasis so far has been on establishing that reality cannot be so reduced to the actual course of events and states of affairs. That is, reality is not only open but also stratified, the real (including the necessary and the possible) is irreducible to the domain of the actual (instances of the possible, the actual course of events) which in turn is irreducible to the empirical or conceptual. This ontological stratification has been found to be a very precondition of ex posteriori natural scientific successes, and to provide the basis for a science of society in the sense of natural science. It might also be added that it is clearly a condition not merely of science but also of the arts, sport, and such day to day activities as cooking, shopping and having a conversation. © 1997 Tony Lawson
But if the primary focus so far has been upon the stratification of reality, I want to round off this part of the book by briefly indicating that further related insights are also implied by (and sometimes explicit in) the arguments already laid out. In particular, the world, including the social domain, can also be seen to be characterised by negativity (real absences), transformative human intentional causality, emergence and internal-relationality or totalities. At this point, though, I only touch on each part of this claim; I wish merely to indicate that the noted features are implicated by the discussion so far, and thereby to sign-post something of the direction taken in later parts of the book.
NEGATIVITY, INTENTIONAL CAUSALITY, EMERGENCE AND INTERNAL-RELATIONALITY If we necessarily accept the causal (as well as perceptual) criterion for ascribing reality (to repeat, it is only on this criterion that we recognise the existence of magnetic fields, i.e. through their effects) it follows that absences can be real. Droughts in crop-growing regions, government non-intervention in a specific sphere of the economy, and public transport strikes are just as causally efficacious and a fortiori real, as, respectively, rain, active governmental regulation and trains and buses which run on time. Human agency has already been elaborated as, in effect, intentional causality. It contributes, purposefully, to (intrinsically dynamic) processes and states of affairs that, in the absence of such agency, would not normally have come about. As we shall see in Chapters 12 and 13, real reasons must be causes in intentional human doings. And these may be routinised, tacit, unconscious, multiple and contradictory. But human agency is not only intentional causality: it is embedded intentional causality. Philosophically, it is sustained by an emergent powers materialist orientation (see Chapter 13). Emergence may be defined as a relationship between two features or aspects such that one arises out of the other and yet, while perhaps being capable of reacting back on it, remains causally and taxonomically irreducible to it. Real emergence clearly turns on the openness of being, a rejection of actualism (the doctrine that the real is reducible to the actual), and necessitates a conception of levels or strata (the physical, the biological, the social, etc). Our pursuit of a separate science of the social sphere, centred upon the intentionality of human agency and involving a recognition of the reality and relative autonomy of actionconditioning social structure, amounts to an acknowledgement of the irreducibility of society to nature. Indeed, it is clear that the form of materialism to which we are committed2 ultimately entails the unilateral © 1997 Tony Lawson
ontological dependence of social upon biological upon physical forms coupled with the taxonomic and causal irreducibility of each to any other. Thus, although, for example, we can acknowledge the geo-historical emergence of organic from inorganic matter and of human beings from hominids, when we come to explain those physical and biological states which are due, in part, to intentional human agency it is necessary to reference properties, including powers, not designated by physical or biological science. In short, the human world must be recognised as an irreducible and causally efficacious dependent mode of being. The reality of emergence leads in turn to the possibility of a multiplicity of modes of determination. That is, any event, state of affairs, or action can be facilitated and/or constrained by different types of laws simultaneously. Thus physical laws and my physiological state both enable me to write this book, whilst setting limits to the manner in which I do so. But they do not determine the content. The latter is also dependent upon social, including cultural and scientific, as well as psychological modes of determination. If machines are governed by the laws of physics or mechanics, economic factors determine where and when such laws apply. My driving is conditioned by mechanical principles, gravitational tendencies and the highway code alike. This discussion of emergence and multiple modes of determination leads to the remaining category of totality. Totalities are systems of (or which include) internally related elements or aspects, i.e. systems wherein some aspects constitute conditions for the existence or essence (characteristic features and ways of acting) of others (see Chapter 12). These are systems wherein the form or structure of the inter- or intra-dependencies or combinations causally affect the elements, and the form and structure of the elements causally influence each other and so also the whole. Paradigm examples include melodies and paintings; each comprise totalities of internally related parts. Each must be conceived as a totality or whole and cannot be understood merely through a separate analysis of each component. As Keynes observes ‘Any judgement as to the beauty of an entity must be derived from contemplation of it as a whole and not deduced from judgements concerning its various parts’ (1904–12). It is easy to see that internal-relationality pervades the social realm. Just as a landlord is an essential condition for a tenant so an employer presupposes employees, and a teacher presupposes ‘students’, and so on. And, equally clearly, powers which such types of individuals can draw upon depend in some way (to be elaborated in Chapter 12) on the relations they stand to one another. Similarly, forms of actions which individuals can and do undertake in strikes, wars, riots and celebrations are significantly determined by the event as a whole and the relations in which they find themselves. Employment practices and labour relations fostered by firms are often highly internallyrelated to the local environment, and in particular to the opportunities, skills © 1997 Tony Lawson
and perceptions of the local labour supply. In ‘industrial districts’ firms are often able to operate as they do only because of the local concentration of others producing complementary products. An industrial locale may be regarded as an internally related spatial network. Even this book is composed of sentences whose elements are semantically bound in a network of relations, and is existentially constituted by the academic project of economics, amongst other things. In short, the world (as we shall explore further in due course) is densely (if not exclusively) populated by totalities. All such features require elaboration, and this is provided in due course under the heading of critical realism. At this point, if to repeat, I merely want to observe that they are already implicated in the picture so far outlined and sustained, that it is already clear that the conception which we must embrace is of a world that is complexly structured, open, intrinsically dynamic, characterised by emergence and so novelty, and inclusive of totalities and causally efficacious absences, amongst other things. And we can immediately see just how shallow in comparison is the ontology (of atomism and closure— see Chapter 7) presupposed by the scientific ideal which informs the project of economic orthodoxy. Thus, in contemporary economics the (usually implicit) social landscape is marked by determinism instead of transformative intentional agency, stasis rather than change, extensionalism (formulated by Hume as the doctrine that events are everywhere ‘loose and separate, conjoined but never connected’) rather than internal-relationality, actualism rather than openness, depth and emergence, and monovalence to the exclusion of negativity. It involves a flattening of a rich complex panorama onto a single plane, a set of lines. And this is all because of the insistence, which can with reason be described as a dogma, that scientific results must be forced into the ‘whenever this then that’ form. I expand upon the social ontology just sketched, along with its implications for economic science, in Part III to Part V. First, though, I return to the more critical discussion; I elaborate in further detail the assessment so far sustained that various strands of the orthodox project are undermined by an acceptance of the noted conception of science. For it will seem to many that I have dealt with some of the most cherished aspects of contemporary economics far too lightly, that rather more argumentation is required. To this latter end I now turn, in Part II, to demonstrate at greater length how not only the more traditional features of mainstream thought but also recent developments, including proposed ‘remedies’ for persistent problems, are explicable by way of the analysis and perspective here sustained. In this I shall also indicate the manner in which the mainstream project is found so often to set off in the wrong direction, and detail the sorts of regresses which inevitably ensue.
© 1997 Tony Lawson
Part II
PROBLEMS OF CONTEMPORARY ECONOMICS
© 1997 Tony Lawson
7 ECONOMETRICS
The intention of this and the following four chapters is to indicate further how the already highlighted features of the deductivist account of explanation, and especially its associated, ultimately positivistic, conception of significant results or scientific laws, are manifest in the most prominent strands of contemporary economics and are responsible for so many of the subject’s problems. I start with that set of activities and methods usually collected under the heading of econometrics. For these can be seen to rest most directly upon the post-Humean conception of scientific laws in question. What, then, is econometrics? Broadly speaking econometricians concern themselves with attempting to determine constant event conjunctions such as I have been discussing, albeit of a probabilistic sort. Not infrequently these sought-after relations are interpreted as ‘causal’. However, this usage of the term is basically Humean. Specifically, it indicates that some ‘dependent variable’ y is held to be functionally related to a set of ‘independent variables’ x, so that movements or variations in the latter give rise to predictable variations in the former. Such ‘variables’ are taken to represent economic events or states of affairs.1 Given this set-up, a primary task of econometrics is to provide quantitative content to any such hypothesised relationships. This involves estimating values of parameters of the sought-after relationships (and of associated probability distributions) and is usually achieved through the application of mathematical and statistical methods to the analysis of economic data (see e.g. Pesaran 1987:8). It is true that quantitative work in economics is not exhausted by this. For example, the determination of summary statistics of a body of data such as means or growth rates of some phenomenon can be included under the rubric of econometrics. Such activities are not being questioned here; there is no suggestion that where the latter and related activities are feasible their pursuit is illegitimate. The specific realist emphasis that I am pursuing is not an antiempirical one. It is also the case that econometricians occasionally concern themselves with ‘time-varying’ parameter models whose understanding requires a modification of the discussion which follows. 2 But the overwhelming bulk of econometric work appears to be characterised by the © 1997 Tony Lawson
search for event regularities (albeit probabilistic ones), and it is this orientation that will be taken as given below.
AN ABSENCE OF ECONOMIC ‘LAWS’ The most telling point against this econometrics project is the ex posteriori result that significant invariant event regularities, whether of a probabilistic kind or otherwise, have yet to be uncovered in economics, despite the resources continually allocated to their pursuit. Fifty years ago Haavelmo (1944) justified his efforts in developing the ‘probability approach in econometrics’ with the observation that ‘economics, so far, has not led to very accurate and universal laws like those obtaining in the natural sciences’ (1944:15). With the passage of time this situation does not seem to have changed significantly.3 Econometricians continually puzzle over why it is that ‘estimated relationships’ repeatedly ‘break down’, usually as soon as new observations4 become available.5 In Part I of this volume, this continued failure is rendered intelligible. Social reality is shown to be dependent upon transformative human agency and thereby intrinsically open and hardly susceptible to scientifically interesting local closures. Given this understanding it is not surprising that any approach whose methods presuppose a ubiquity of local closures should fail. However, if the argument that I am making is correct it seems that I ought to be able to explain more than this. Specifically, I ought to be able to throw light upon the nature or direction of systematic attempts by econometricians to resolve their problems. Previous failures have certainly induced thoughtful and influential responses, including proposed ‘solutions’. Given that all such ‘solutions’ remain premised, albeit implicitly, upon the positivist conception of scientific laws, it may be feasible to anticipate the sort, or direction, of any tendencies set in train. I ought to be able to show not only that ‘solutions’ of the sort proposed are not solutions at all, but also that they are, in fact, actually intelligible manifestations of a still inadequately analysed problem. This is my immediate objective here: to explain the direction of systematic responses to econometric failure. In this, however, I avoid the invidious strategy of picking out individual econometric applications (as well as the accusation that my observations apply only to some carefully selected ‘strawperson’) by focusing on rather prominent and influential commentaries which recommend ways in which to proceed. There is no shortage of suitable candidates. However, over the last twenty years or so one response to the failings of econometrics has dominated others in terms of its notoriety. I refer to the extra-ordinarily influential ‘Lucas critique’ and the set of responses that © 1997 Tony Lawson
it inspired. For my purposes it is clearly convenient to examine the nature of this critique and the reactions to it. To the extent that these can be shown to be manifestations of the sorts of tendencies which are to be anticipated from the perspective I am defending, the explanation I am putting forward can be seen to be that much more firmly grounded.
THE ‘LUCAS CRITIQUE’ Before identifying the tendencies in question let me briefly review the essentials of Lucas’ argument. Lucas’ entry point is the observation, noted in Part I, that theory and practice are out of step with each other. Specifically, Lucas points to the ‘many signs that practicing econometricians pay little more than lip-service to the theory’ which underpins that practice: The most striking [sign] is the indifference of econometric forecasters to data series prior to 1947. Within the theory of economic policy, more observations always sharpen parameter estimates and forecasts, and observations on ‘extreme’ x values particularly so; yet even the readily available annual series from 1929–46 are rarely used as a check on the post-war fits. A second sign is the frequent and frequently important refitting of econometric relationships. The revisions of the wage-price sector now in progress are a good example. The continuously improving precision of the estimates of [parameters] within the fixed structure..., predicted by theory, does not seem to be occurring in practice. Finally, and most suggestively, is the practice of using patterns in recent residuals to revise intercept estimates for forecasting purposes. For example, if a ‘run’ of positive residuals (predicted less actual) arises in an equation in recent periods, one revises the estimated intercept downward by their average amount. This practice accounts, for example, for the superiority of the actual Wharton forecasts as compared to forecasts based on the published version of the model. (Lucas, 1976:22–3) The import of these signs is that they indicate that the econometric models actually constructed, the probabilistic relationships identified as connecting sets of measurable events or states of affairs, are unstable. The theory upon which most econometric analyses are based presupposes the existence of stable parameter relations. But the observed practices of econometricians indicate that the models actually derived are not sufficiently stable to allow the
© 1997 Tony Lawson
successful forecasting of events occurring outside the period for which the models were initially constructed. Lucas locates the source of the observed theory/practice inconsistencies in the continued failure to obtain successful results by way of sticking to accepted theory. This of course is quite consistent with the conclusions of Part I. However, instead of interpreting this failure as an outcome of working with an untenable conception of science, Lucas suggests a rather more substantive cause of the problem. Specifically, in formulating an explanation of the noted instability in econometric formulations Lucas proceeds as follows. First, he emphasises that the ‘structure’ of any sought-after econometric relationship (connecting ‘variables’ denoting measurable events or states of affairs) depends upon ‘decision rules (demand and supply functions) of agents in the economy, and these decisions are, theoretically, optimal given the situation in which each agent is placed’ (Lucas, 1976:25). Second, he suggests that the appropriate or optimal decision rules will vary systematically according to the pattern or structure of the series of events that the decision makers are optimising over. Third, he infers that if there are changes to the patterns of the series that the agents are optimising over, changes which Lucas attributes to adjustments in ‘policy rules’, then the inevitable result is instability in the structures of relationships connecting economic events. Herein lies the basic problem with econometrics which Lucas identifies. The usual methods of obtaining ‘structurally stable relationships’ for simulating the effects of alternative policies require that the sought-after event relationships be invariant with respect to changes in policy rules, while considerations about the behaviour of optimising agents give reason to doubt that such invariant relations can arise:6 To assume stability of [the sought-after econometric relations] under alternative policy rules is thus to assume that agents’ views about the behaviour of shocks to the system are invariant under changes in the true behaviour of these shocks. Without this extreme assumption, the kinds of policy simulations called for by the theory of economic policy are meaningless. (Lucas, 1976:25) Lucas adds that ‘Everything we know about dynamic economic theory indicates that this presumption is unjustified’ (ibid.: 25—emphasis in the original). Interestingly, the solution to the noted difficulties originally advanced by Lucas does not really focus on econometric practice at all. Rather it turns on the way in which policy rules are formulated and introduced. In other words, Lucas addresses the policy-maker rather than the econometrician. The basic idea is that the simpler the rules, and the more they are communicated in © 1997 Tony Lawson
advance, ensuring that they will be correctly anticipated by the relevant population, the easier it will be for the econometrician to model optimising behaviour:7 If the policy change occurs by a sequence of decisions following no discussed or pre-announced pattern, it will become known to agents only gradually, and then perhaps largely as higher variance of ‘noise’. In this case, the movement to a new [structure of the economy], if it occurs in a stable way at all, will be unsystematic, and econometrically unpredictable. If, on the other hand, policy changes occur as fully discussed and understood changes in rules, there is some hope that the resulting structural changes can be forecast on the basis of estimation from past [experience] (ibid.: 40–1) Lucas concludes his paper: In short, it appears that policy makers, if they wish to forecast the response of citizens, must take the latter into their confidence. This conclusion, if ill-suited to current econometric practice, seems to accord well with a preference for democratic decision making. (ibid.: 42)
REACTIONS TO THE ‘LUCAS CRITIQUE’ This, however, if the end of the original paper, is not quite the end of the story—at least not for the majority of econometricians who have taken this ‘Lucas critique’ seriously. Inevitably perhaps, essential aspects of what is thought by Lucas to constitute a solution to the difficulties detected have been reinterpreted by others as, or transformed into, a hypothesis about the conditions actually prevailing. Specifically, in the ensuing econometric practice the hypothesis of rational expectations has moved centre stage as its proponents argue that changes in government policies are, most of the time at least, really very well understood after all and facilitating of predictable responses. The emergent project is thus one of building econometric models that account for systematic behaviourial responses to changes in policies, culminating ultimately in the search for econometric methods capable of isolating parameters that may, after all, be invariant to changes in policy rules. Several strands of this latter project are in evidence. At one extreme are the contributions of Sims (1982) who appears to believe that economic theory can © 1997 Tony Lawson
provide few restrictions for specifying the econometric relationships. The conclusion thus inferred is that economists should concentrate upon largely uninterpreted reduced forms. At the other extreme are the contributions of those who believe that economic theory can indeed provide explicit a priori equation restrictions. The latter contributions—assessed in Hoover’s very useful study as comprising the ‘most popular’ (Hoover, 1988:192) response to the ‘Lucas critique’—are described by Sargent (1986), an occasional collaborator with Lucas, as follows: Recent research has been directed at building econometric models that take into account that people’s behaviour patterns will vary systematically with changes in government policies—the rules of the game. Most of this research has been conducted by adherents of the socalled hypothesis of rational expectations. They model people as making decisions in dynamic settings in the face of well-defined constraints. Included among these constraints are laws of motion over time that describe such things as the taxes people must pay and prices of the goods they buy and sell. The hypothesis of rational expectations is that people understand these laws of motion. The aim of the research is to build models that can predict how people’s behaviour will change when they are confronted with well-understood changes in ways of administering taxes, government purchases, aspects of monetary policy, and the like. (Sargent, 1986:3) Further on in the same text Sargent broaches the topic of new econometric methods. He observes: A major research effort is currently under way by economists to develop theoretical and econometric methods capable of isolating parameters that are structural in the above sense, that is, parameters that are invariant under government interventions in the form of changes in the rules of the game. This is a very ambitious undertaking, one that is in many ways more difficult and ambitious than was the impressive effort of the Cowles Commis-sion in the late 1940s that created the econometric methods that made Keynesian econometric models possible. For what is required is a theoretical and statistical framework that permits the economist to estimate how private agents’ decision rules or strategies depend on the decision rules or strategies used by the government. Any successful version of this research effort will embody the principle that the parameters of private agents’ decision rules are not among the free parameters of the model, but are themselves © 1997 Tony Lawson
functions of (among other things) parameters describing the rules used by the government. Achieving success in this endeavour requires that many new methods and results be achieved in technical aspects of econometrics and dynamic economic theory. Lucas and Sargent have formalized the ideas behind this research effort in the following way... Impressive technical progress is being made in this research endeavour, and there are grounds for being reasonably optimistic that economists will eventually be able to deliver econometric methods that are useful for predicting the effects of changes in government policy rules. However, it is wise to keep in mind what an ambitious task this is. It requires several leaps of faith in economic theory to hope that this line of research will realize all of its aims, although the success of this research line will eventually have to be judged not in absolute terms but vis-à-vis alternative lines of research designed to achieve the same objectives. (ibid.: 11–13) To repeat, the main responses to the Lucas critique turn either upon fashioning the economy after the needs of the econometrician (Lucas), or upon imputing such powers of foresight to lay agents and of technique to econometricians, that changes in policy rules are not after all debilitating of econometric analysis (Sargent and Sims). Such responses are certainly remarkable. In the open social system in which we live, characterised as it is by fundamental uncertainty, the hypothesis of rational expectations is hardly realistic. Although human beings knowledgeably and skilfully negotiate and accomplish their every day tasks, the knowledge actually possessed is obviously a far cry from the requirements of this particular hypothesis about expectations. Indeed, the suspicion must always have been that the resort to such a hypothesis is little more than an act of desperation. I now want to explain this response from the perspective that I am defending; to show that it really is nothing more than a transformation in the form of the symptoms of disorder rather than anything approaching a solution of the problems to which Lucas draws attention.
SYSTEMATIC RESPONSES TO GENERALISED PREDICTIVE FAILURE My objective here is to identify at a general level the sorts of responses that might be expected if, as I am arguing, the problems of the discipline turn on a generalised presumption of the ubiquity of spontaneously occurring closed systems (while social phenomena are by and large generated in a system that © 1997 Tony Lawson
is intrinsically open). Once this is achieved, I will indicate how the ‘Lucas critique’ and the responses it encouraged in the practices of econometricians can be recognised as a particular manifestation.8 First, though, let me add a point of clarification. My aim here is to anticipate and/or explain responses to econometric failure. Thus I am not concerned with the plethora of strategies which are, or could be, adopted prior to failure, i.e. within a definite programme of statistical inference in any given area of substantive research. That is, my focus in this chapter is not with this ‘planning stage’ of the analysis, as Leamer labels it, but rather with the stages of ‘criticism’ and ‘revision’ which follow once failure is admitted.9 My concern, in fact, is one that is not usually addressed within the perspective of standard theories of statistical inference; nor, I think, can it be in any meaningful way.10 Furthermore, we shall see that the tendencies induced by continuous econometric failure are quite the opposite of the seeking-out of parsimonious representations which planning-stage strategies often encourage.11 Let me also emphasise, finally, that I am concerned with the question of strategic responses. For in the face of repeated failure there may be a multitude of more or less arbitrary reactions. And indeed it is a telling feature of the econometrics literature generally that many practitioners, whether or not researching an area where hypothesised econometric relationships have recently broken down, appear to take no notice whatsoever of previous studies and formulations.12 The result is a plethora of seemingly quite unrelated ‘models’ or specifications with little sense of cohesion or progression. Sometimes the adopted stance is little more than the ‘fitting’ of data to hand, with minimal reliance on theory. Other, possibly less ad hoc responses, include (repeatedly) transforming data or modifying estimation techniques. Such responses are not my focus here. Rather, I am concerned with the sorts of systematic options available to those who seek a programme of progressive improvement, a programme that at a minimum recognises a need to situate, explain and attempt to learn from previous failures or apparent successes.
REGULARITY STOCHASTICISM My starting point is a recognition that econometricians, albeit implicitly, are in effect committed to a particular formulation of the post-Humean scientific ideal. I refer to a metaphysical thesis that can be labelled regularity stochasticism. According to this thesis, for every (measurable) economic event or state of affairs y there exists a set of conditions or events, etc., x1, x2... xn say, such that y and x1, x2... xn are regularly conjoined under some (set of) ‘well-behaved’ probabilistic formulation(s). In other words, stochastic closures © 1997 Tony Lawson
are everywhere assumed to hold; for any (measurable) economic event y a stable and recoverable relationship between a set of conditions x1, x2... xn and the average or expected value of y (conditional upon x1, x2... xn), or some such, is postulated. My objective here is to determine how an adherent of regularity stochasticism might sensibly proceed if stable event regularities of the soughtafter kind are not successfully identified, if any apparently established regularity is found always soon to break down. This is the question that must at some level be addressed by any practising econometrician. Of course, even if regularity stochasticism is not a valid doctrine some constant conjunction of events of the sought-after kind may come about, at least over some limited region of time-space. Nothing in the preceding discussion rules out such a possibility. And, indeed, there can be no a priori argument to exclude such a possibility. Econometricians seriously concerned with their discipline, however, need more than this. Given their tacit support for the substantial resources devoted to the econometrics project they need ways of proceeding which enhance the possibility of identifying structurally stable relationships and with some reasonable degree of frequency. In consequence, an obvious course of action to take is to attempt to identify certain basic conditions which, if satisfied, would ensure that a stable econometric relationship holds. The question, then, is what, if any, set of basic, collectively sufficient, (types of) conditions can be identified? Now at some level all econometric analyses rest upon, or presuppose, the agency of individuals of some kind. Consumers’ expenditures (and so consumer expenditure functions) presuppose consumers, production presupposes producers, and so forth. In practice such economic individuals as are explicitly theorised by econometricians include not only human agents but also firms, households, and trades unions. What set of conditions on a system of such individuals, then, would ensure that an event regularity of the sought-after kind obtains?
EXTRINSIC CLOSURE First, it is clear that, for any chosen outcome or ‘dependent variable’ y of interest, the identification of a particular set of individuals and conditions that are thought to be ‘explanatory’ of it may not be sufficient for an event regularity also to be identified. For, in a complex system such as the social world there are likely to be individuals and conditions, or ‘variables’, over and above those explicitly identified that do, or at least could, influence y. If, then, an event regularity of the sought-after kind, between y and some specified set of conditions, is to be guaranteed as well as recoverable, an obvious strategy is to seek to identify a particular set-up, a transformation of © 1997 Tony Lawson
the conditions, which effectively entails restrictions on the unelaborated conditions or potential influences on y prohibiting their interference with the event regularity being sought. Somehow the system of interest must be ‘closedoff’ from their effects. The requirement of satisfying restrictions of this sort can be referred to under the heading of the extrinsic closure condition. In fact, if a definite set of conditions or factors has been identified, and let us assume for the moment that these influences on y are (taken to be) stable, then it is clear that in order to guarantee a stable and recoverable event regularity, either any influences not explicitly accounted for must be constant in their operation or, more generally, the set-up must be of a sort that any such influences are orthogonal to (uncorrelated with) those conditions explicitly focused upon. The obvious special case of this restriction is that any individual or system of analysis be physically isolated from any such potential, but nonexplicitly identified, influences. From this perspective it is not surprising to find that contemporary orthodox economics looks so much to assumptions of isolated systems (‘closed economies’) or individuals. And we have now identified the first obvious response to instability in estimated econometric relationships: to suggest that the system has so far been incompletely described, that extrinsic closure has not been achieved, that some significant influence or condition has been erroneously omitted from explicit formulation within the system.13
INTRINSIC CLOSURE Even if a system were effectively isolated from non-orthogonal external influences an event regularity is not yet guaranteed, at least where the individuals of analysis are characterised by intrinsic structure and complexity. For each such individual of analysis may, in repeated or identical extrinsic conditions, behave differently by virtue of different internal states. Prima facie, of course, most individuals are so structured. Thus the manner in which a given person may answer a specific question, conduct himself or herself in the market-place, or perform some other specified act, is likely to depend upon whether he or she is alert or sleepy, optimistic or pessimistic, starving or well fed, in pain or not in pain, etc. Likewise a firm’s (‘entrepreneur’s’) response (in terms say of employment offers) to a fixed increase in demand for the firm’s product may vary according to whether the firm is currently oriented towards manufacturing or service provision, whether or not it has recently installed labour saving technology, and indeed on the technical and social organisation of the firm quite generally. It follows that if an event regularity is to be guaranteed a second obvious strategy is to seek some set of restrictions on the intrinsic structure, organisation or nature of any (and every) individual © 1997 Tony Lawson
of analysis which guarantees that a given fixed set of conditioning factors, x1... xn, exert a constant effect on y, the outcome variable of interest. For obvious reasons the requirement of satisfying such a set of restrictions can be gathered under the heading of the intrinsic condition for a closure. One possible restriction to attempt to shore up in this regard is that of intrinsic constancy: that an explicit specification be given for each of the various intrinsic possible states of any individual, where it is supposed that each satisfies the requirement of an internally or intrinsically constant structure. Notice that if the basic individual or unit of analysis is an entity like the firm (rather than, say, an entrepreneur) then the obvious condition to seek to satisfy is the intrinsic constancy not only of the workplace, etc., but also of each human agent involved in producing the outcome of interest. Now if the intrinsic structures of each and every relevant (state of each) individual of analysis were constant, and also a fixed set of conditions or influences was in play, this would not yet guarantee an event regularity, i.e. that given identical conditions of action or operation, the same outcome must ensue. In order for the intrinsic closure condition to be guaranteed, a second requirement is reducibility: that each individual of analysis is so structured or organised that, for any given intrinsic state and set of conditions of action, only one outcome or ‘exit’ is possible, and one that is predictable from a knowledge of the prevailing set of conditions.14 In other words, the outcome variable of interest must be determinate, the structure or organisation of each individual must be such that, for any given (and constant) intrinsic structure, the outcome of interest (or average outcome in repeated conditions) must be deducible from, or reducible to (in the sense of being expressible as a constant function of) the conditions or states of the system. Typically, additivity appears to be the assumed principle of organisation. But, however that may be, to guarantee that given ‘forcing’ conditions give rise to a determinate, in principle conditionally-predictable, outcome, the obvious strategy is to seek to identify a set-up wherein the conditions of intrinsic constancy as well as reducibility are satisfied. The latter couple, then, constitute the most obvious restrictions to seek out which ensure the satisfaction of the intrinsic condition for a closure. Both of these restrictions are automatically satisfied if any and every relevant individual can be characterised atomistically—in effect as lacking intrinsic structure. For then each reaction is only and always but a passive response to external impinging forces or stimuli. We can thus take note of a second response encouraged by a lack of stability in any estimated econometric relationship: to suggest that the individuals of the analysis have yet to be given a sufficiently atomistic description. In other words, while an attention to satisfying the extrinsic condition encourages the response of broadening or extending the system, a concern to satisfy the intrinsic condition for a closure encourages the response of ‘digging deeper’. © 1997 Tony Lawson
AGGREGATION Where the noted intrinsic and extrinsic conditions are satisfied it seems that a closure must come about. In principle, then, such conditions (and the identified modes in which their satisfaction are likely to be sought) are all that we are looking to isolate here. In actual studies, however, econometricians primarily concern themselves with macro or aggregate or system-wide outcomes or effects, i.e. with effects at a level of aggregation which is different to that at which theory can sensibly be applied. Frequently, as in orthodox ‘theory’, the question of interest is whether all individual actions or responses are, or rather could be, in some sense mutually consistent—encouraging conceptions of equilibrium (or other ‘solution concepts’) in economics (see Chapter 8). In econometric ‘modelling’, by contrast, the question of most apparent interest is whether the sort of event regularities expected at the level of some (sub-set of) individual(s) can also be anticipated at some more aggregate level. That is, although econometricians typically theorise at the level of some individual, say at the level of a consumer or producer, observations or data employed in econometric analyses usually relate to outcomes for a substantially larger sector or collectivity. For example, they often relate to the total of consumers’ expenditure, or to the production of an industry or even of the economy as a whole. In principle, the aggregate outcome or dependent variable can always be determined derivatively by first seeking to ‘explain’ the corresponding outcome at the level of each (type of) individual (or anyway at some more disaggregate level) and then summing. Certainly, by characterising the various states of all individuals as crypto-atomistic there is a presumption that their effects can be added together or at least combined in some mechanistic fashion. For it follows from their atomistic characterisation that any responses to given conditions are independent of the particular combination of states or individuals or of their inter-relationships. In other words, by assumption no principle of organic unity, or totalities, can be allowed to operate, whereby a system has properties irreducible to its component parts. There is nothing in such a ‘summing’ procedure that is in contradiction with an adherence to regularity stochasticism. In practice, however, such a move is usually infeasible given the (aggregate) nature of the ‘data’ available. In any case it is not how econometricians tend to proceed. Typically aggregate level relationships are pursued; the attempt is made, for example, to relate aggregate consumers’ expenditure directly to total personal sector disposable income.15 The guarantee of such an aggregate or macro-relationship, one taking the sought-after event regularity form, clearly requires something more than the satisfaction of the intrinsic and extrinsic closure conditions identified above. © 1997 Tony Lawson
It also requires some principle of composition or combination, or aggregational condition, for an (aggregate) closure. Populations are heterogeneous.16 Reactions to any given economic change of interest will vary according to whether the individuals in question are rich or poor, members of large households or small ones, operating big firms or small ones, capital intensive firms or labour intensive ones, and so on—and the composition of any group can change. Some set of conditions or restrictions is required of any system or collectivity of analysis, then, which ensures that stability in responses nevertheless emerges at the aggregate level of focus.17 Formally, the most obvious sufficient strategy to pursue is to demonstrate either that the individual equations satisfy some set of ‘integrability conditions’ such that aggregate stability is seen to follow without necessitating much attention to distributional information explicitly, or that the structure of the distribution of individual attributes relevant to the analysis at hand is such that it can be, and is usefully, included explicitly as a tractable component of the model specification. An alternative—more restrictive but obviously more compelling—sufficient strategy, of course, one that effectively constitutes a limit, or degenerate, case of the two situations just noted, is to attempt to select the group or collectivity of analysis in such a way that its members are, after all, in a relevant sense, homogeneous. Here we have one final obvious response to equation instability, and specifically to instability in any aggregate or macro-formulation: to argue that the system has yet to be given a sufficiently disaggregated or micro description. Now it is important to be clear what I am suggesting. To repeat an observation made above, even if regularity stochasticism is an untenable metaphysical thesis a restricted event regularity may by chance arise. Such a possibility is not theoretically refutable. The import of the preceding outline, rather, is that conditional upon accepting regularity stochasticism as true, if the intrinsic and extrinsic—and, at the level of aggregate or macrorelationships, the aggregational—conditions are satisfied, an event regularity of the sought-after kind must ensue. Equivalently, conditional upon regularity stochasticism being true, the failure or breakdown of a postulated event regularity of the requisite sort entails that at least one of these closure conditions does not hold.
REINTERPRETING THE ‘LUCAS CRITIQUE’ How does this discussion throw light upon the Lucas critique and the responses that it has elicited? I think it explains them. From the perspective opened up the ‘Lucas critique’ is straightforwardly interpretable as an argument that at © 1997 Tony Lawson
least one of the identified closure conditions is not satisfied in econometric models. And the responses of econometricians to this critique are just as readily interpretable as attempts to satisfy the closure condition(s) in question, in order that the econometrics project might finally be shorn up. Of course, neither Lucas nor those econometricians influenced by his argument elaborate their projects in the manner now set out, and so their strategies must be reconstructed. But certainly the Lucas critique and the more prominent responses at the level of econometric practice can be straightforwardly rendered comprehensible from this perspective. What, first of all, does Lucas have to say about, or what does his analysis reveal concerning, the satisfaction of the intrinsic condition for a closure? For Lucas the unit of theoretical analysis seems always to be the human individual. Intrinsic constancy is then rendered satisfied (in theory) by supposing only one intrinsic state for all individuals: the assumption adopted is that individuals are only and always economic optimisers. Reducibility is achieved by so specifying the set-up that a (unique) optimum always exists relative to the prevailing conditions: ‘decisions’ are to be always ‘optimal given the situation in which each agent is placed’ (Lucas, 1976:25).18 If Lucas’ theorising is typically at the level of the individual human agent, his primary concern appears to be with macro-econometric relationships, that is with event regularities at an aggregate level. What then does Lucas have to say that bears upon the aggregational condition for a macroeconometric closure? It turns out that the shoring-up of this condition is essentially fudged. At first the signs are that this will not be the case. Indeed, Lucas starts out lamenting the theoretical gap he perceives between the ‘micro-economic’ formulation of individual behaviour and econometric attempts to determine the aggregative behaviour of the system as a whole. However, as the discussion progresses, and illustrative examples are provided, it emerges that the micro-specification for, say, individual consumer behaviour, is interpreted, as suits his purposes, as an individual or aggregate (average) formulation inter-changeably. The problem of aggregation, in other words, is implicitly assumed anyway as the behaviour of any sector of the economy, or the economy as a whole, is equated with that of an (optimising) individual. Of course, the main focus of Lucas’ criticism, if implicitly, is the nonsatisfaction of the extrinsic condition for a closure. Basically, his argument is that human behaviour is treated by econometricians as though it were unaffected by (that it is isolated from the influences of) changes in government policy rules, when in practice this is highly unlikely. And the response of practising econometricians to this critique is to endeavour to shore up this
© 1997 Tony Lawson
extrinsic closure condition by attempting to internalise explicitly within the econometric formulation this newly identified influence. But from the perspective now opened up it can be seen that Lucas has really only touched the tip of an iceberg. Why restrict the focus of extrinsic, unaccounted for, influences to government policy rules? Even if we wish to posit perpetually optimising agents, there are all sorts of reasons to doubt the existence of the type of conditions necessary for optimising to proceed—so casting doubt on whether the intrinsic condition can be satisfied. Such reasons also suggest that the system can rarely if ever be viewed as one of isolation— so suggesting problems in satisfying the extrinsic condition for a closure. We live in an open economy, subject to the influences of erratic movements in world trade, multiple regionalised wars, frequent institutional collapses, the regular re-arranging of trading policies, agricultural gluts and failures, novel political developments including arbitrary, unannounced foreign policy changes bearing on matters relevant to domestic consumers and traders, etc. Equally, of course, all such factors are relevant to the requirement of meeting the aggregational condition. With investors or traders, say, responding to developments in different institutional, political and cultural contexts, how reasonable is it to suppose that stability can result at the aggregate level—that is, for formulations for imports, exports, investment, etc., specified for an economy as a whole? Once, furthermore, it is allowed that human agents are internally structured and complex and possess the capacity always to have acted otherwise the whole framework appears to go by the board. The import of this discussion then is that it enables us to render intelligible a set of regresses that are widely observed. They explain the most familiar set of reactions to econometric failure. For, from the perspective opened up it can be recognised that the ‘break-down’ of estimated econometric relationships encourages either or both of two broad types of response. First, failure can be rationalised as the erroneous omission from the analysis of some significant factor—thereby suggesting a need to shore up the extrinsic closure condition. The second response is to argue that the system in question requires a more micro and/or atomistic formulation—thereby recognising the requirement, and attempting to satisfy the aggregational and/or intrinsic closure conditions. It is these implicit attempts to shore up the identified closure conditions that account for the familiar pattern of responses to econometric failure: of expanding ever further or ‘digging’ ever ‘deeper’. The ‘Lucas critique’ and the reactions it encouraged merely constitute a recent and influential example.
© 1997 Tony Lawson
THE FALLACIES OF ATOMISM AND ISOLATIONISM Ultimately, then, regularity stochasticism in economics is seen to be underpinned by two dogmas, which can fairly be labelled the atomistic and isolationist fallacies respectively. These are the suppositions that all behaviour or aspects can in some form be given, or reduced to, a (crypto-) atomistic representation19 and treated analytically as though effectively isolated from all other factors not explicitly identified.20 It is under the influence of these fallacies that econometricians attempt repeatedly to shore up the closure conditions. And in the limit the tendencies so set in train amount to the search for systems so large that they exclude nothing, couched in terms of single individuals so small that they include nothing. This is the direction in which the familiar responses to econometric failure ultimately lead. The alternative, efficacious solution, of course, one that is possible only when the positivistic conception of science is emphatically rejected, is to posit an autonomous, structured, intentional agent somewhere between the universe and the atom, that acts in a complexly structured, fundamentally open, internally-related and changing world. An adherence to regularity stochasticism, and more generally the postHumean ideal for science, then, sets us off searching in the wrong direction. The Lucas critique and the responses it stimulated are just recent manifestations of this problem. But it is a problem that pervades econometrics (as it is usually practised), as well as the many other activities of economists that are rooted in positivistic theory. The alternative recourse is to recognise the greater explanatory adequacy of the perspective provided under the heading of transcendental realism and accordingly accept its re-formulation of the possibilities and tasks of economics as science. We might note, finally, that the arguments of this chapter can be reconstructed as a series of embedded generalisations. A set of conditions which together are sufficient to ensure a closed system, the obtaining of a constant conjunction of events, has been identified. The Lucas critique suggests one reason why one of these conditions may not be met. From the perspective opened up, however, it is easy enough to suggest a variety of reasons why each of the closure conditions identified are unlikely to be met very often. In its turn this generalisation of the Lucas critique of econometric practice is itself generalised by the realist critique of the post-Humean conception of science upon which econometric practice, amongst much else, rests. It follows that so long as we accept the post-Humean or positivistic ideal and focus upon the conduct of, and real possibilities for, econometric practice, then the sort of problems identified by Lucas, and here generalised, must be interpreted, along with Lucas, as largely debilitating for economics as science © 1997 Tony Lawson
and policy evaluation. However, once the whole positivistic conception is rejected as untenable and the viability of the transcendental realist alternative perspective and project is recognised, it can be seen that there are, after all, no inevitable problems in any of this for economics as science, or indeed (as I will demonstrate in due course) as input into policy evaluation and design.
© 1997 Tony Lawson
8 ECONOMIC ‘THEORY’
The single most prestigious branch of modern economics is a project referred to by its mainstream proponents as pure theory, economic theory or just plain theory. Its mathematically (or rather formalistically) inclined adherents likewise refer to themselves (and usually only to themselves) as theorists. In recent years, various problems have come to light with the ‘theory’ project1 that have led even its proponents to question its capacity for accounting for real world events or for assisting in policy formulation. In consequence, as with the problems of econometrics identified in the preceding chapter, numerous diagnoses have been made and various solutions have been suggested. In the event, however, the remedies proposed, if in the context sometimes remarkable, have failed to make any headway in terms of facilitating a project that is seemingly capable of illuminating social reality. And the reason for this, as in the case of econometrics once more, is simply that the problems identified do not fully represent the real nature of the relevant disorder. My aim in this chapter is to elaborate further how the almost unquestioned reliance upon a deductivist mode of explanation along with its associated conception of scientific laws, accounts for the noted shortcomings. I also show how it explains the particular forms of ongoing attempts to resolve recognised difficulties of the ‘theory’ project, and why such attempts inevitably fail.
NOMINAL FEATURES OF THE ORTHODOX ‘PURE THEORY’ PROJECT What, first of all, is the nature of this project? If it is commonplace to find references to something called ‘economic theory’ or ‘pure theory’ in modern literature, there is some disagreement about what it comprises. In the assessments of many critics (and some proponents) the project is characterised according to various recurring nominal substantive features. Specifically, economic ‘theory’ is widely interpreted as a body of substantive thought that, © 1997 Tony Lawson
amongst other things, focuses upon individuals rather than collectivities; upon exchange activities rather than production or distribution; upon optimising (maximising or minimising) behaviour rather than satisficing or habit following; upon conditions of perfect competition rather than oligopoly or monopoly; upon structures facilitating constant (or decreasing) returns to scale rather than increasing returns; upon presumptions of perfect knowledge and foresight or ‘rational expectations’ rather than uncertainty or ignorance; upon end-states, fixed points, or equilibria, rather than processes in time; upon functions (utility, cost, preference, profit) that are well behaved (where appropriate, convex, differentiable, fixed, well ordered over all the arguments, etc.) rather than otherwise. If such nominal characteristics are often associated with ‘economic theory’, and are undeniably pervasive, it is also noteworthy that many ‘theorists’ have regarded most of them as not being in any way essential. This, for example, is the position found in the various meta-theoretical contributions of Hahn, or at least those made prior to the last few years or so.2 In both his Jevons’ memorial lecture entitled In Praise of Economic Theory and the introduction to his collection of essays entitled Macroeconomics and Equilibrium, a somewhat different account is given. Specifically, only a restricted set of features of the ‘neoclassical’ economic theory project are explicitly identified as essential. Three which are enumerated are 1 an individualistic perspective, a requirement that explanations be couched solely in terms of individuals; 2 an acceptance of some rationality axiom; and 3 a commitment to the study of equilibrium states.3 In addition to these features Hahn also expresses a preference for highly general, simple theories over particular ones and a conviction of the need to draw an explicit distinction between axioms and assumptions. Both axioms and assumptions are to serve as premises of theoretical analysis. But whereas axioms (such as the claim that people always act rationally) are thought, on grounds of empirical adequacy, to be beyond question, assumptions are recognised as less secure. Quite often it is anticipated that acceptance of such assumptions is temporary; they are contemplated merely in order to facilitate particular analyses getting off the ground. From this perspective, much about the contemporary state of affairs is explained. For example, a realisation that many of the nominal features noted by critics are accepted by ‘theorists’ only under the category of assumptions (as opposed to axioms) is consistent with their being interpreted as inessential to the theory project and ultimately discard-able. The metaphor of scaffolding is often invoked in this regard indicating a presumption that most of these assumptions can be eventually replaced or dismantled leaving the essential © 1997 Tony Lawson
structure of the building intact. In turn, the particular focus upon (rational) individuals explains why so many of the prominent nominal features (hypotheses regarding beliefs, preferences, resources, etc.) relate to aspects of an individual’s situation. And this focus also accounts for the concentration on exchange activity rather than production or distribution. For it is only the former that can be interpreted at all plausibly as the momentary interaction of independent individuals. In short, although not every traditionally recurring aspect of the orthodox project is accounted for by the sort of core claims asserted by Hahn and others (see below) it is clear that a significant number of them are.4
CHANGING FEATURES In the light of this traditional orthodox view of things, certain recent ‘internal’ developments appear rather significant. For the project’s central figures appear now to be changing their assessments about the nature of the project’s character. Specifically those who, like Hahn, have hitherto emphasised the virtue of focusing more or less exclusively upon the decisions of rational individuals, on equilibrium states, and, perhaps most significantly, upon simple highly general theories, appear now to be supporting quite different positions. In place of rational individuals we find, on the one side, stochastic, fixed (context-independent) rule-following automata or ‘particles’ (Canning, 1990) and, on the other, groups or societies somehow coordinated so as to act as, or on average as, a single rational individual (Kirman, 1989). In place of equilibrium states, any of a variety of system solution concepts seems in order. And, most notable of all, in place of simple, general premises and the simple ‘elegant’ theorems which they facilitate, the advice that is being given to potential ‘theorists’ is to focus more upon complex and specific premises and to embrace the methods of computer simulation. In this, indeed, it is being predicted that ‘economic theory’ will even need to address itself to matters of biology, sociology and history. As Hahn himself has recently written: Not only will our successors have to be far less concerned with the general (leave alone the ‘generic’) than we have been, they will have to bring to the particular problems they will study particular histories and methods capable of dealing with the complexity of the particular, such as computer simulation. Not for them the grand unifying theory of particle physics which seems to beckon physicists. Not for them, or at least less frequently for them, the pleasures of theorems and proof. Instead the uncertain embrace of history and sociology and biology. (Hahn, 1991:50) © 1997 Tony Lawson
EXPLANATORY FAILURE What, then, is going on? What explains this apparently dramatic change in the stance or perspective countenanced? The cause of it all, it would appear, is a final, if reluctant acknowledgement that the highly generalised models of ‘economic theory’ are too general to entail anything of interest. The problem with the ‘theory’ project is simply that it does not throw light on any actual economic phenomena. The result that appears to have tipped the balance derives from orthodox theory’s central focus on both general equilibrium theory and game theory in particular, wherein apparently insuperable problems are acknowledged in the continuous search for possible situations of equilibrium, that is for situations in which all agents’ plans are in some sense mutually compatible. Specifically, it has been widely observed in the highly generalised frameworks such as advocated previously by Hahn that if an acceptable specification of an ‘economy’ allows of one equilibrium it typically allows of many; and even if any such equilibrium solution is found to be unique there is no obvious mechanism in the standard orthodox specification of an economy to ensure that the economy could get there.5 Such results are clearly not encouraging if orthodox ‘theorists’ ever wish to explain real phenomena, or to influence policy formulation.6 From this perspective the change of orthodoxy’s direction is largely comprehensible. In essence, the sorts of moves noted are designed to render the premises of theorising less general or ‘arbitrary’ so that specific conclusions might be entailed. Several strands of this activity are discernible. One move is to replace traditional very general ‘assumptions’ by more specific, albeit often increasingly implausible ones. In particular, the specifications of preferences and beliefs are made ‘stronger’. Thus, we find assumptions of interpersonal utility comparisons (ideal utility functions) in welfare economics and cost benefit analysis; restrictions on the possible distributions of income in general equilibrium theory (Hildenbrand, 1983a); assumptions of ‘similar’ preference functions in ‘social choice theory’ (Caplin and Nalebuff, 1989); conceptions of representative consumers in macro economics, and so on. A second response is to endogenise more of the basic conditions or ‘assumptions’ of the system. A recent example in general equilibrium and game theory is that, implemented under the heading of ‘rational expectations’, of endogenising beliefs and/or expectations, allowing the latter to be ‘determined’ as an equilibrium condition of the model.7 A further line of response is the noted replacement of the axiom of rationality by the claim that agents follow fixed, context-independent, rules. This move appears to have been accepted for an environment of random shocks and motivated by the observation that, in physical systems, the © 1997 Tony Lawson
presence of randomness in the movements of atomic particles often prevents any such system from settling down in unstable equilibria or limit cycles. Indeed, a finding of statistical mechanics is that in a system with a very large numbers of particles, the average position and velocity of these particles (when averaged over the entire population), are ‘well behaved’. The point then is to seek analogous results for economics by considering human agents as particles, as simple rule followers or some such, where beliefs and preferences can be described by points in some high dimensional space. When two or more agents interact, their beliefs and perhaps preferences, are endogenously amended according to the rules that they follow, so that the agents move to a new point. If the number of agents is large enough, the behaviour of any one, or so it is assumed, will essentially be random. In consequence it is supposed that, with beliefs and actions averaged over the entire population, average outcomes will be ‘well behaved’ and hence predictable (see Canning, 1990). One final move that we can consider here is that taken by those ‘theorists’ who appear most concerned that the project should be rendered as realistic as possible. This is the recourse of specifying the ‘assumptions’ of analysis, not arbitrarily, nor by endogenising them as part of the system solution concept, but by incorporating, where possible, the findings of empirically based studies. Solow (1986:24–5) for example, argues that economic theory must be modified to accommodate the fact that in reality individual decisions are conditioned by context-specific values and habits. This strand of thought certainly supports the conclusion that economists must pay more attention to matters of history, biology and sociology. And it is the complexity of any such models as would result that encourages the prediction of Hahn and others that theorists must eventually be concerned with such procedures as computer simulation. So what is the outcome or prospect? Is this turn to historical, biological and/or sociological matters a basis for expecting greater empirical success, for anticipating models of greater relevance to policy analysis? Are there now grounds for supposing that the often-noted fictitious nature of ‘economic theory’ is about to be transcended? Do these proposed transformations or adjustments even entail that the ‘economic theory’ project is, or will be, no more; that, in effect, it will be or has been transcended, as many of its critics appear to hope and applaud? I think that all of these questions must be answered in the negative. Such changes as have been occurring represent merely the latest set of manoeuvres aimed at prolonging the life of an essentially misguided project. The nature of the proposed changes, it will be seen, are such that they necessarily leave the project still incapable of dealing with its manifest problems. They do not facilitate a theory capable of explaining real-world events or of assisting policy formulation. In elaborating © 1997 Tony Lawson
now why this must be so various other issues can also be addressed. Why is it the case, after all, that in transforming the ‘theory’ project, certain sorts of paths have been followed and not others? What, in any case, explains the previous persistent emphasis upon individualist explanation and rationality? Hahn does not elaborate. And why now condone a conception along the lines of agents as naive, fixed, context-independent, rule followers? What, moreover, explains the continuing focus upon equilibrium states or similar system solution concepts? These sorts of issues can be made intelligible and the futility of the orthodox project as a whole rendered apparent. But in order to do so it is necessary, first, to acknowledge a more adequate account of the real or essential nature of the orthodox ‘pure theory’ project than has been elaborated so far in this chapter.
ECONOMIC ‘THEORY’ AND DEDUCTIVISM There is one essential feature of the ‘economic theory’ project, of course, that I have not so far emphasised in this overview of the project’s more prominent aspects and recent developments, but which is at least as fundamental as any other. It is a feature, moreover, that does not turn on the choice of substantive premises. I refer to the mode of theorising or, specifically, the form of explanation that is fundamental to it, one that I referred to in Chapter 2 as deductivist. Recall that to explain something is to provide an account (the explanan(s)) whereby that something (the explanandum) is rendered intelligible. According to deductivism, as I am using the term, to be able to explain an actual event or state of affairs is to deduce a statement of it from a set of initial or boundary conditions plus universal ‘laws’ (constant conjunctions of the form ‘whenever event x then event y’). For adherents of this conception, furthermore, the explanation of such laws, as well as theories and sciences, likewise proceeds by deductive subsumption. Now although deductivist explanation does not usually figure explicitly in characterisations of the project provided by its proponents, its centrality cannot be doubted. While the deducibility requirement, that the explanadum be deducible from the explanans, is more or less always transparent, the covering-law aspect, namely that at least one universal law (of the form ‘whenever event x then event y’) be specified, is usually met by the axioms.8 Certainly, axioms take the ‘whenever this then that’ structure and typically serve to relate variables expressing types of events or states of affairs. Moreover, some ‘theorists’ even take axioms to be justified claims about actualities of the social world, about events, etc. that actually occur:9 ‘It is not that [axioms]...are divorced from experience or observation but rather that
© 1997 Tony Lawson
they mark the stage beyond which one does not seek to explain’ (Hahn, 1984:6). Elsewhere, and more generally, Hahn writes: For I do not wish to deny that there are empirical regularities of human economic behaviour awaiting discovery. But I claim that these will be, as it were, much deeper down, more elementary and closer to the form in which axioms are postulated than are the complex, institution and history dependent ‘facts’ of the econometrician. (ibid.: 332)
Axioms are not plucked out of thin air and far from distancing the theorists from what somewhat mysteriously is called the ‘real’ world they constitute claims about this world so widely agreed as to make further argument unnecessary. (Hahn, 1985:5) In short, if a reliance upon the deductivist mode of explanation is not always explicit in orthodox accounts, it is not denied.10 Rather a presumption of its centrality and indeed universality in science is essentially taken for granted; so much so that any attempted defence or justification of it is considered unnecessary.11 Those critics who venture to suggest otherwise tend merely to be summarily dismissed. To restate Hahn’s (1985:9) conclusion, already noted in Chapter 2: ‘If one is kind to such critics [of deductivism] one interprets them as signalling that they do not care for these axioms and these assumptions. In any case all theory in all subjects proceeds in this manner.’ That is, it ‘consists of logical deductions from axioms and assumptions’. In a word, the relevance of (typically formalistic) deductivist reasoning in economics, and thus of any philosophical perspective upon which it ultimately rests, are regarded as being beyond question. But are such matters beyond question? Is the status of deductivism as a universally applicable mode of reasoning secure? Are its philosophical underpinnings sound? Despite the confidence with which ‘economic theorists’ hold to the deductivist mode of explanation in general, and to formalistic modelling which it facilitates in particular, it is precisely this line of reasoning, or the presupposition of its universal legitimacy, that, in the first part of the book, has been rejected as quite untenable. We have seen in fact that it is exactly the presumption in question that accounts for the project’s numerous difficulties. Specifically, because the project rests upon an implicit commitment to identifying or formulating regularities of the form ‘whenever event x then event y (albeit perhaps in an a priori or introspective fashion), its legitimate application is restricted to those very special situations in which scientifically significant event regularities are (or might be expected to be) © 1997 Tony Lawson
forthcoming—which, in the economics sphere, may be hardly any situation at all.12
GENERALIST CLAIMS AND THE QUESTION OF DETERMINACY Note that the foregoing deductivist characterisation of orthodox theory, and in particular the observation that the enterprise depends upon the principle of empirical invariance (i.e. upon the premise that laws or significant claims in economics are, or depend upon, event regularities) is in no way undermined by the recognition that the axioms of ‘economic theory’ have tended to be postulated at a very general level. It is true that this tendency has encouraged the assertion that any of a range of outcomes or events is often possible. But it is equally the case that any such range of possibilities is not considered to be a feature of the reality being expressed; the assertion made is not ontological but epistemological. Such ‘possibilities’, in other words, arise only as a result of the current limitations of contemporary economic theory, i.e. because models which command attention have yet to be fully specified. In consequence, the basic deductivist characterisation of orthodox theory as set out above is not in this way compromised. Let me expand upon this last claim. Although the axioms of orthodox theory are often formulated at a relatively high level of generality, these general claims (as already noted) are nevertheless usually interpreted as, or as depending upon, event regularities of our real social world, and ones that can be regarded as quite secure (I am not at this point concerned with the validity of this latter claim that axioms are empirically justified, which I dispute; I merely note its existence). Thus, for example, an axiom of the form ‘firms maximise some function of profits’ is, despite its apparent generality, seen as a putative real-world event regularity. Hahn, for example, writes: The axiom that firms maximise some function of profits is stated as such because the theorist is not proposing to answer the question why firms should do so. But it is not plucked out of the air or from dreams. It encapsulates an empirical phenomenon which many practical people and economists believe to be the nature of the capitalist. It does so at a more general level than, say, the pay-back theory but it is every bit as empirically motivated. (Hahn, 1984:6) Now not only are highly abstract claims of the sort in question considered to be empirically adequate by orthodox economists, but so far it is only © 1997 Tony Lawson
hypotheses elaborated at this level of abstraction that are held to be empirically secure. Of course, it is not the case that orthodox analysis never proceeds at a lower level of generality. But because such lower level specifications, or ‘special theories’, as abound—for example specific profit, production, or preference functions—are considered to be more questionable in the light of empirical evidence, deductions derived when employing such formulations are regarded as less secure. For this reason, indeed, and as noted above, the more specific or lower level event regularities postulated in orthodox theorising are often referred to not as axioms at all, but as assumptions. Thus, Hahn (1985) asserts: ‘That managers have preferences is an axiom; that they take a particular form, for instance that they are linear in expected profit, is an assumption...’ (1985:11). ‘That people have preferences and try to satisfy them we treat as an axiom while universal perfect competition, for instance, must count as an assumption’ (ibid.: 10). ‘If you want more out you have to put more in, that is you must supplement the axioms with assumptions, e.g. that production functions are Cobb-Douglas’ (ibid.: 13). The point then is that such indeterminacy as can be found in orthodox theorising and which is often even applauded as a virtue, does not arise because the event regularity conception of science is in any way rejected. Hahn is explicit that he does ‘not want to deny that there are empirical regularities of human economic behaviour awaiting discovery’. Rather, such indeterminacy as arises in orthodox models does so only because secure or empirically adequate knowledge is not yet available concerning event regularities holding at the lower levels of abstraction, that is at the level that Hahn refers to as ‘special theory’. It arises only when the model has yet to be fully specified. In consequence, when ‘theorists’ have praised their theories for not placing ‘single valued restrictions’ on the world, that is for not predicting specific economic outcomes or events, this claimed virtue must be seen for what it really is: a rhetorical device to mask what, on orthodoxy’s own terms, is merely a state of ignorance.13 Hahn, indeed, more or less explicitly states as much: Lastly, I want to argue that it is one of the virtues of theories derived from axioms more ‘fundamental’ than those used in special theory, that they usually do not yield single valued restrictions on the world.... This is a virtue because the economist is thereby restricted from claiming more than he has reasons for claiming. The axioms have summed up what one regards as pretty secure empirical knowledge. The set of outcomes which are possible is simply the reflection of our lack of knowledge. A special theory can usefully narrow them down. But our confidence in the special hypothesis is smaller than in the axioms. A © 1997 Tony Lawson
claim of only one outcome should always include the proviso that given our state of knowledge there are also other possibilities. (Hahn, 1984:7) In sum, the high level of generality which traditionally has characterised much of ‘economic theorising’ is quite consistent with the deductivist conception of explanation as elaborated above, including the event regularity view of laws (here referred to as ‘axioms’ and ‘assumptions’), on which this explanatory form depends. In consequence, despite some of its rhetoric, including the sometimes found explicit dissociation with positivist thinking, this project must be recognised as flawed in precisely the manner of all those projects which take the post-Humean conception (including structure) of science, or its laws, to be necessary (if not sufficient) for understanding and enlightenment.
A DEDUCTIVIST TRADITION I also note, if somewhat parenthetically, that the advocacy of this conception in economics is hardly new. I have focused on Hahn’s views of the situation because he is one of very few contemporary mainstream economists prepared to reflect at all explicitly on what is done. But this coyness on the part of leading economists is a relatively recent phenomenon. If we look back forty or fifty years or more, we find that explicit reflections by prominent economists are rather more numerous. Consider, for example, Robbins’ (1932) Essay on the nature and significance of economic science, which for many still provides the most acceptable definition of what economics is about (vis ‘Economics is the science which studies human behaviour as a relationship between ends and scarce means which have alternative uses’ (1932:16). Here Robbins concludes: In the light of all that has been said the nature of economic analysis should now be plain. It consists of deductions from a series of postulates, the chief of which are almost universal facts of experience present whenever human activity has an economic aspect, the rest being assumptions of a more limited nature based upon the general features of particular situations or types of situations which the theory is to be used to explain. (Robbins, 1932:99–100) He even supposes that: The efforts of economists during the last hundred and fifty years have resulted in the establishment of a body of generalisations whose © 1997 Tony Lawson
substantial accuracy and importance are open to question only by the ignorant or the perverse. (ibid.: 1) Now Robbins does assert that whereas natural science scientific generalisations are known ‘only inferentially’ economic ones are known to us by ‘immediate acquaintance’. But all that is concluded from this is that there is less reason to doubt the existence of the entities referred to in economic theories than those of natural science (ibid.: 105). Certainly, the presumption is that economic science is about reality: It is a characteristic of scientific generalisations that they refer to reality. Whether they are cast in hypothetical or categorical form, they are distinguished from the propositions of pure logic and mathematics by the fact that in some sense their reference is to that which exists, or that which may exist, rather than to purely formal relations. In this respect, it is clear, the propositions of Economics are on all fours with the propositions of all other sciences. As we have seen, these propositions are deductions from simple assumptions reflecting very elementary facts of general experience. If the premises relate to reality the deductions from them must have a similar point of reference. (ibid.: 104) However, and like Hahn once more, Robbins is of the view that the supposed empirically grounded propositions of economics are all formulated only at a very general level.14 And he too doubts whether concrete quantitative laws of economics can be practically derived. Part of the reason for his holding this view is a recognition that there are usually a large number of determinants of economic phenomena. But partly the reason is a realisation that certain factors, especially individual valuations and technical conditions, will be changing over time. This does not lead Robbins to suppose that by collecting large bodies of statistics and standardising for numerous factors it is impractical to obtain regularities and so perhaps ‘elasticities’ which hold for a period. He recognises, in fact, that ‘Rough ideas relating to the elasticity of demand in particular markets are indeed essential if we are to make full use of the more refined tools of economic analysis’, and may even be useful for predicting over a period. Robbins is clear though that ‘they have no claim to be regarded as immutable laws’ (ibid.: 109). At different points in time, it seems, different sets of data must be used to estimate relevant functions: ‘If we wanted to be helpful about herrings we should never dream of relying on the researches of [someone]...working in 1907–8. We should work the whole thing out afresh on the basis of more recent data’ (ibid.: 109). But, however that © 1997 Tony Lawson
may be, economic ‘laws’ of a general kind do exist, facilitating a deductivist, predictive, economic science: But to recognise that Economic laws are general in nature is not to deny the reality of the necessities they describe or to derogate from their value as a means of interpretation and prediction. On the contrary, having carefully delimited the nature and the scope of such generalisations, we may proceed with all the greater confidence to claim for them a complete necessity within this field. Economic laws describe inevitable implications. If the data they postulate are given, then the consequences they predict necessarily follow. In this sense they are on the same footing as other scientific laws, and as little capable of ‘suspension’. (ibid.: 121) With the passage of time Robbins’ assessment of the successes of the discipline (in terms of establishing a body of the sought-after generalisations) is seen to be somewhat exaggerated (to put it mildly). But the nature of the enterprise remains. Indeed, the primary difference between the assessments which are explicitly laid out by the likes of Hahn and Robbins and actual explanatory endeavours, at least as they are formally reported in mainstream journals, is that the latter are rather less circumspect or cautious in the way in which deductivist procedures are adopted. They are wielded, indeed, as if ultimately applicable to the ‘explanation’ of seemingly any situation or phenomenon, at apparently any level of generality/specificity, and for any space-time region. In short, I am suggesting that they are treated by many, and perhaps most, as if they are (uniquely) applicable to the understanding of all social, including economic, phenomena. If my thesis is correct that the problems for the ‘theory’ project stem from its uncritical adherence to deductivism, we might expect to be able to explain not only the project’s continuing lack of insight, but also something of those aspects that have so far been left unaccounted for. Specifically, we might expect to be in a position to render intelligible not only the more familiar or traditional range of premises and orientations adopted by orthodox ‘theorists’, but also, and perhaps especially, the sorts of novel moves that have recently been countenanced in an attempt to overcome the project’s numerous failings. My objective here is to demonstrate that the thesis in question can indeed shed insight on the so far unaccounted for features of the orthodox ‘theory’ project. And the obvious way to proceed in this is to repeat the strategy of the previous chapter and attempt to identify the sorts of options which might appear to be (or have been) available for those orthodox theorists who fail to appreciate (or refuse to accept) the inherent limitations of the deductivist project (i.e. that it presupposes a ubiquity of social closures, while, in the © 1997 Tony Lawson
event, social closures of an interesting kind appear to be extremely rare). Specifically, I want to examine the most likely responses of those who continue to act on the supposition that deductivism constitutes a more or less universally applicable mode of economic explanation. For this should at least indicate what it is that such ‘theorists’ are hoping to achieve. And, as we shall now see, it is indeed by addressing this question that those features of ‘economic theory’ which have so far been left unaccounted for can finally and straightforwardly be explained.
REGULARITY DETERMINISM A presumption of the universal applicability of the deductivist mode of explanation must ultimately rest upon an adherence to a metaphysical thesis that is referred to here as regularity determinism. According to this thesis (which is clearly a special, or degenerate, case of regularity stochasticism identified in the previous chapter), for every economic event or state of affairs y there exists a set of events or conditions x1, x2,... xn, such that y and x1, x2... xn are regularly conjoined under some (set of) formulation(s). The relevant question to pursue, then, is what sort of constraints does such an orientation place upon, or at least encourage in, theory construction? It was observed in the previous chapter that all economic ‘theorising’ is couched in terms of economic individuals of some sort, whether consumers, employers, households, firms, banks, trades unions or whatever. Once more, then, an obvious condition to seek to satisfy is that any individual (or set of individuals) is so intrinsically constituted or organised that under any repeated, completely specified, or isolated set of states or conditions of action x1, x2... xn the same outcome for y is guaranteed to follow.
INTRINSIC CLOSURE This is the restriction that we have previously referred to as the intrinsic condition for a closure. It entails two essential aspects. A first constraint to impose, or to seek to guarantee in the course of economic reasoning, is that of intrinsic constancy: that the internal, or intrinsic, structure of any (delineated state of any) individual of analysis be constant. The second obvious requirement (of any theoretical construction that is to satisfy the intrinsic condition for a closure) is reducibility: that the overall outcome event, for any state description, be reducible to the system conditions obtaining. Clearly the conditions of intrinsic constancy and reducibility are both automatically satisfied if any and every relevant individual is characterised © 1997 Tony Lawson
atomistically, as in effect lacking intrinsic structure. For then each reaction is only and always but a passive response to external, impinging, forces or stimuli. And we shall see below that such advantages or attractions of atomistic conceptions do indeed render intelligible fundamental aspects of substantive developments within the mainstream project.
EXTRINSIC CLOSURE Before considering such matters, however, it must be registered once more that a satisfaction of the intrinsic condition does not by itself guarantee a closure. A satisfaction of the intrinsic condition ensures that if the elaborated set of conditions or states x1... xn entails a complete list of potential ‘influences’, then the outcome for y can always be determined. A potential problem for deductivist attempts to predict or ‘explain’ any actual outcome arises, however, with the possibility of other conditions, including possibly novel factors, that are not included in the specified set of conditions x1, x2... xn, but which, nevertheless, are capable of influencing the actual outcome y. For deductivist explanations of actual events to be guaranteed, then, such influences must be either internalised within the system or shown to be constant in their action. Once more, then, we now have a second restriction, an extrinsic closure condition, for mainstream economists to build into their models: that only the explicitly elaborated conditions x1, x2.. xn have a systematic, non-constant, influence on the outcome event y in question. We saw in the previous chapter that the extrinsic condition is most straightforwardly satisfied where the actions of the relevant individuals are physically isolated from all the (non-constant) conditions not explicitly set out. In other words, while the desire to satisfy the intrinsic closure condition encourages conceptions of individuals (or their states) as crypto-atomistic, a similar need to satisfy the extrinsic condition encourages formulations in which each individual acts in relative isolation. In such a scenario, of course, the overall outcome of a system composed of numerous such individuals can then be determined merely by adding together (or otherwise mechanistically combining) their separate responses. Let me once more attempt to clarify the argument. As with the doctrine of regularity stochasticism already examined, even if regularity determinism is an untenable metaphysical thesis, a restricted event regularity may by chance come about. That is, even accepting that all individuals are internally structured and complex and act in an open, structured and changing world, it is still theoretically possible that an apparently stable relationship between conditions and some outcome event is observed, at least over some restricted region of space-time. Such a possibility, if on the face of it highly unlikely in © 1997 Tony Lawson
any form that is scientifically interesting, is not theoretically refutable. Those who emphasise the theoretical nature of the orthodox enterprise, however, require much more than this. The desire15 of ‘theorists’ to ‘explain’ events in some theoretical manner, combined (if implicitly) with a belief in the universal relevance of the deductivist mode of explanation, sets them deriving theoretical constructs that have conditions to guarantee the ‘whenever this event then that event’ formulation built into them. Of course, there is unlikely to be a unique set of sufficient conditions of this sort. But the intrinsic and extrinsic conditions just outlined, by focusing directly upon both the nature of individuals and also their conditions of action, do seem to constitute obvious choices for most occasions. And the most straightforward conceptions to pursue in the endeavour to satisfy these two conditions, I have suggested, are those of atomistic individuals acting in conditions of relative isolation.
REINTERPRETING THE ‘THEORY’ PROJECT The various substantive moves of orthodox ‘theorists’ that have so far been left unaccounted for are now easily explained. For they are nothing other than specific attempts to provide accounts of atomistic individuals acting in relatively isolated conditions in order to ensure that the identified closure conditions are met. Of course, economic ‘theorists’ do not explicitly elaborate their project in these terms (any more than econometricians do with their own project) and so underlying strategies must once more be reconstructed. But certainly the traditional premises and recently revised outlooks of ‘theorists’ are easily explained from the perspective now set out. Consider first the intrinsic closure condition. How is this condition addressed in contemporary economics? What sorts of moves are made, if any, in orthodox theory to ensure that any individual of analysis is of a kind to act always in the same way in identical conditions? It must by now be obvious that this has usually been achieved in contemporary economics by (1) making human individuals more or less the only unit of analysis, (2) characterising human nature in such a way that such individuals are only and always ‘economically rational’, i.e. economic optimisers (thus achieving intrinsic constancy), and (3) specifying some objective (utility, preference, profit) function as fixed and permitting of a (locally at least) unique optimum and so single course of action (ensuring outcome reducibility). It is clearly due to the implicitly or tacitly recognised need to provide some basis for supposing the satisfaction of the identified intrinsic closure condition that epistemological individualism and the rationality ‘axiom’ (and associated types of assumptions) have been so stubbornly maintained. However, we can now clearly see that if the latter axiom is an obvious one © 1997 Tony Lawson
to hold on to, it is not after all necessary. Assumptions such as that made by Canning (1990) that agents always follow simple, fixed, context-independent, rules will do equally well. All that is required is that human agents rebound in atom-like fashion to external influences. Indeed, in this light it is not surprising that Canning should attempt to model human agents by drawing directly upon the physics of atomic particles. Conditional upon the methods of deductivism (as opposed to the substantive or material forms of the deductively structured orthodox formulations) constituting all that is essential to the orthodox project, any such move as Canning’s is quite intelligible and, indeed, legitimate. For, at the end of the day, all that is required on this front is that, in mainstream theoretical specifications, any agent of analysis is constrained in determinate, deducible ways given the situation. Ultimately the need, given the specified situation, is for a single, predictable outcome or ‘exit’.16 The extrinsic condition for a closure is clearly addressed by treating each agent as acting within an isolated and typically highly limited fixed set of conditions. Thus the focus is usually upon buyers and sellers interacting in a quite impersonal way in their attempts to satisfy their preferences in situations more or less completely specified by given endowments and prices. Production, too, if implausibly, is often reduced to the activities of relatively isolated ‘entrepreneurs’. It thus takes the form of entrepreneurs exchanging inputs for outputs on the basis of some fixed technique of production or some such (see e.g. Hahn, 1984:80). And when a collection of agents and their interactions is the focus, the set-up is usually conceived of as an economy or game in its entirety, as some closed off, completely described and physically isolated, system. As Kapp (1976:210) observes in his summary assessment of ‘neoclassical’ economics: ‘Concrete economic systems or processes are...believed to be adequately represented as isolated, selfcontained and self-sustaining, closed mechanical processes with definable boundaries.’ If it is the case that individual agents have traditionally been specified as acting in a highly restricted set of supposedly isolated conditions, one obvious response to the failure of the ‘theory’ project is to explore alternative, more complex formulations. In other words, in the place of overly simple formulations aimed at satisfying the extrinsic condition, we now see the latter being addressed through certain history- and context-specific features of each agent’s (nevertheless still isolated) situation being introduced into the formal specification. Perhaps the most significant point here is that until recently the possibility of deductivist analysis has rested upon the sought-after event regularities consisting in or resting upon a not too large set of conditioning factors in order to reduce the complexity of the analysis. In recent years, however, the accelerated rate of availability of computer resources has © 1997 Tony Lawson
entailed that more complex formulations can be addressed. This explains why at this point in time more complicated, historical, biological and sociological, specifications or conditions are at least being openly and positively contemplated.
SOLUTION CONCEPTS What can be said, though, about that one feature of orthodox theorising not yet adequately addressed: the insistence upon some sort of equilibrium notion as a fundamental component of analysis? Notice that this is not something that can be deduced from specifications at the level of individuals: it is a system-level (global, macro, or aggregate) property.17 How, then, is the equilibrium orientation to be explained according to the perspective I am defending? The answer, of course, turns on the dependence of the deductivist project upon a set of postulated behavioural regularities or formal equations. For a question which arises naturally in such a framework is whether there exists a set of parameter values or conditions which renders any such set of equations or formulations mutually consistent. The notion of equilibrium serves precisely to capture any such condition. Thus, in its most common and weakest form, an equilibrium is interpreted just as a state in which all actions and/or plans and/or expectations (depending upon the set-up) are mutually compatible, wherein the system equations are mutually consistent. From this perspective it is not surprising that, with the repeated failure of the project to enlighten, and with various novel specifications (consistent with a satisfaction of the intrinsic and extrinsic closure conditions) being explored, a range of novel ‘solution concepts’ or types of equilibrium, are being continually elaborated. For clearly different specifications are required to fit the specific details of the models in question. And given that all methods or formulations that presuppose or entail a closure are, in the end, bound to be unenlightening of a social world largely insusceptible to closure, the one question that will continually attract attention, and when answered provide some semblance of an achievement, is whether it is possible to demonstrate that the formal system at hand is at least internally coherent. From this perspective, the orthodox attention to equilibrium notions, or more generally ‘solution concepts’, if largely besides the point, is nevertheless not surprising.
© 1997 Tony Lawson
ECONOMIC ‘THEORY’ AS DEDUCTIVISM It is clear that the various features of the ‘economic theory’ project traditionally held as essential, along with the more recently proposed revisions of perspective, can easily be explained once the unquestioning, uncritical, orthodox adherence to the deductivist mode of reasoning is acknowledged. Given the revealed flexibility of that project at the level of substantive premises, including axioms as well as assumptions, and its apparent inflexibility at the level of its mode of explanation, I suggest that an adherence to deductivism in the context of attempting to understand social phenomena be recognised not merely as fundamental to, but actually constitutive of, the ‘economic theory’ project. In other words, in the absence of any alternative hypothesis that is capable of accounting for the nominal or manifest characteristics of orthodox contributions examined throughout, I suggest that the mainstream ‘theorising’ project be recognised just as the uncritical adherence to, or persistence with, the deductivist mode of explanation in the context of economics.
COMPETING INTERPRETATIONS But is it really the case that there is no sustainable alternative interpretation? At this point I should probably address Hausman’s (1992) recent account of The inexact and separate science of economics (to use the title of his book). I take it that Hausman uses the term ‘economics’ here to mean that project which mainstream economists refer to as ‘economic theory’, though he himself expresses it as ‘neoclassical economics’ 18 or ‘orthodox theoretical microeconomics and general equilibrium theory’.19 If indeed he and I are attempting to characterise the same project then Hausman offers an alternative assessment, one that provides the most systematic and comprehensive treatment of the interpretation that most critics of orthodox ‘theory’ seem to hold, albeit one, nevertheless, that I think is ultimately inadequate. What then does Hausman mean when he suggests that this reputed ‘science of economics’ is ‘inexact and separate’? It is a consequence of the discussion in Part I of this book that predicates such as ‘natural’ ‘social’ ‘biological’ ‘and economic’ are likely to be better suited to differentiating distinct kinds of mechanisms rather than events. For in the production of an event in an open system, many of these predicates may be simultaneously applicable. If I understand Hausman correctly it is essentially this that he is getting at with his ‘inexact and separate’ characterisation. Thus, the term inexact signals that certain social, or anyway ‘market’, phenomena will be determined by © 1997 Tony Lawson
causes other than those posited in ‘economic theory’,20 an observation on which we can readily agree. And the term separate expresses a belief that orthodox economics is defined in terms of its own given set of causal factors (rather than outcomes), albeit a set, according to Hausman, which (or the set of ‘laws’ of which) is already completely known. The domain of economics is ‘then the realm of social phenomena in which those causal factors predominate’ (Hausman, 1994:208). What is this set? In Hausman’s assessment: ‘Economic phenomena are the consequences of rational choices that are governed predominantly by pursuit of one’s own consumption and profit. In effect, economics studies the consequences of rational greed’ (1994:208). Now we can readily admit that rational greed has been the most dominant or recurrent of the ‘theory’ project’s supposedly explanatory factors. However, we have also seen that the orthodox ‘theory’ project is significantly more flexible than this and indeed is repeatedly changing its substantive claims and orientations, often involving some explicit alternatives to the presumption of rational greed. This characterises much of recent game theory, for example, where assumptions of ‘irrationality’ are often in play, and where, in some situations, conceptions of rationality seem peculiarly ill-defined (on all this see e.g. Abreu and Pearce, 1984; Basu, 1990; Binmore, 1990). Hausman does recognise this situation. However he writes: Not all of what is called economic, even orthodox neoclassical economics, is concerned with a peculiarly economic realm. Inquiries in game theory, for example, which shade into work in standard economics and are carried on by many of the same theorists, often relax common motivational assumptions (such as self-interest, commodity bundles as the sole objects of preference and independence of utility functions). The strategic interactions with which game theorists are concerned consequently need not lie within the specifically economics realm or domain. But to recognize that some of economics does not concern this domain is consistent with accepting the existence of an economic realm and demanding that economic theory span it. (Hausman, 1994:207) The assessment that mainstream economists may theorise about something other than an ‘economic realm’ may seem both legitimate and reasonable. But, given the extent to which, at least until very recently, game theory has come to dominate the concerns of ‘economic theorists’ and to absorb space in journals regarded as mainstream, it surely strains credibility to see it as lying outside the formal bounds of orthodox economic theory? This is true even if (as with most other recent substantive concerns of mainstream economists) it © 1997 Tony Lawson
turns out only to be a passing phase. Moreover, the form of enquiry at issue here is one that is concerned with a realm that is not so much distinct from that which Hausman characterises as economics as, at times, actually inconsistent with it. Economic theorists developing game theory, for example, may be concerned not only with a causal mechanism other than rational greed but also its denial. Notice, here, that while an objective of Part I was to determine a sustainable conception of science, and in pursuing this objective I was content to allow that economic science may be something other than what mainstream economists mainly do, the objective currently under consideration is rather different: to characterise mainstream (or ‘neoclassical’ or ‘orthodox theoretical micro...’) economics. The accepted aim is to identify what mainstream ‘theorists’ actually do, not, for example, what they might or ought to do. Hausman is explicit that ‘what follows is a description, not an evaluation’ (1994:205—see also 1992:90). In consequence, it seems that any conception sustained ought to bear some resemblance to, or at least be able to accommodate without strain, the most prominent recent activities and concerns of those regarded by themselves, and by everyone else, as proponents of orthodox economic theory. As the subject moves on a danger with Hausman’s strategy is of defining ‘economic theory’ mainly in terms of what ‘economic theorists’ used to do. Parenthetically, albeit as a partial redress and/or foil to Hausman’s emphasis, we can also take up Mirowski’s (1994) puzzle here. In questioning the nature of ‘neoclassical economics’ and adopting a rather broader orientation to the subject, Mirowski suggests that although its opponents often treat it as a monolithic doctrine, the historian is forced to concede that, in fact, it is best described as a sequence of distinct orthodoxies, surrounded by a penumbra of quasi-rivals; and that it is this, more than any deductive or inductive ‘successes’, which accounts for its longevity. (Mirowski, 1994:68) After tracing out a succession of ‘formalisms’ in the history of the project Mirowski feels that we are inevitably led to the question: ‘does neoclassicism consist of anything more than bald assertion of continuity in the face of repeated ruptures every two or three generations? (ibid.: 69). My response, of course, is to observe that such problems of internal coherence as face Hausman’s account or as pointed out by Mirowski only arise if we focus on substantive or ‘theoretical’ content. Once we recognise that orthodoxy, in all its various forms, turns on its (deductivist) methodology there are no problems of reconciliation at all.21 © 1997 Tony Lawson
POTENTIALS AND ACTUALITIES It may clarify matters if I provide an alternative formulation of the thesis I am defending. My argument is (clearly) not that people never act in an economically rational way, nor that firms never seek to maximise profits, nor for that matter, of course, that they never follow rules. The point, rather, as I will discuss in more detail in Chapter 13, is that these or related notions must be conceived in terms of potentials; as potentials that may or may not be expressed, and if expressed that may or may not be actualised because of countervailing tendencies. Although human beings seem to possess a capacity to act rationally, we may not always seek to exercise this capacity, and even when we do our actions may be dominated by some rather irrational and other counter-tendencies, ones we may even regret as soon as they are expressed. The fundamental error of orthodox theory here, at least where realisticness is a goal, is not its focus upon such conceptions as rationality or profit seeking per se;22 the problem is its presumption that such matters, at some level at least, are always expressed in terms of actualities rather than capacities. And with the need to pre-specify one ‘actuality’ when many are really possible, the usual choice has been the situation in which some narrowly conceived ‘economic goal’ is achieved. In the context of specific economic models, agents have usually ended up optimising something; a potential is reduced to its exercise and its exercise successfully actualised. Of course, this latter sort of strategy is necessitated, and so explained, precisely by the prior commitment of ‘theorists’ to regularity determinism. This emphasis upon actualities (whether real or fictitious) is essential to the deductivist ‘whenever this then that’ explanatory form. However, when seen in this light it is apparent that in place of premises asserting optimising behaviour, assumptions of sticking to given rules will ultimately do, or anyway appear worth examining (as recent developments bear out); as indeed are any specifications which facilitate deducible outcomes. It is such an actualism, a flattening of capacities into actualities, whatever the precise nature of the ‘theoretical’ content or specification, that is the generalised material form or counterpart of the deductivist structure of science and explanation.
PROSPECTS Let me finally note that if this assessment of ‘economic theory’ (as deductivism applied in an attempt to understand social phenomena) is correct, it is possible to anticipate once more the direction of future tendencies encouraged by a persistence with the theory project. Or, more accurately, it is possible to © 1997 Tony Lawson
anticipate the sorts of tendencies likely to be induced by those economic ‘theorists’ who are most concerned to render their project realistic, or at least endeavour openly to ‘explain’ actual events and situations. For we can see that concern about the continued explanatory failure of the project is likely to induce one or both of the two types of general response already seen to characterise the project of econometrics. That is, the strategy encouraged is to suppose either that the extrinsic condition for closure has been inadequately addressed, with the obvious conclusion that more (possibly context-specific) influences must be explicitly identified; or that the intrinsic closure condition has yet to be adequately addressed, with the obvious inference that the individuals of the analysis have yet to be given a sufficiently atomistic formulation. In other words, the maintenance of the deductivist project sets in train two regresses, with the common limit point, once more, of a universe of empty individuals. Under the guidance of the fallacies of atomism and isolationism we yet again inevitably set off in the wrong direction. The recent recommended turn to computer simulations is just an early manifestation of this. The alternative and efficacious response is to posit instead an autonomous, complexly structured agent somewhere between the universe and the atom that acts in an equally structured and fundamentally open, internally-related and changing world. In other words, the path to transcending the problems of contemporary economics entails recognising the superior adequacy of the perspective provided by transcendental realism and accordingly accepting its reformulation of the possibilities and tasks of economics as science and explanation. If economics is a discipline in some disarray, it is particularly noticeable that its most prestigious branch, ‘economic theory’, is quite unequal to the tasks of explaining real world events or informing policy analysis. The argument of the current chapter is that this lack of relevance is in some significant part due to an uncritical acceptance, within contemporary orthodox economics, of a particular mode of explanation, namely deductivism, including an associated, indeed conditioning, conception of science, necessitating the formulation of laws or law-like statements which at some level express constant event conjunctions. Until it is recognised that such a conception can be of only limited use in the fundamentally open social world in which we live, one that seems hardly susceptible to interesting local or regional closure, pure theorising as traditionally understood, i.e. as the universal application of the methods of deductivism, seems destined to undergo a variety of analytical contortions, manoeuvres and regresses of the sort identified, without actually resolving any of the discipline’s problems.
© 1997 Tony Lawson
9 ALTERNATIVES AND/OR PRELIMINARIES TO SUCCESSFUL EXPLANATION
The ‘theory’ project discussed in the previous chapter is unable to explain actual phenomena of the real (open) social world. But do orthodox ‘theorists’ usually or consistently seek to explain or even address actual phenomena? A belief that the project ought to be explanatory, that it should seek to overcome its explanatory failures, underpins the recent turn to methods of computer simulation and the like. For the goal here, by most accounts, is the generation of models sufficiently complex to accommodate the apparent intricacies of actual social phenomena and their relations. Even so the aim of formulating ‘models’ that should be confronted with data on actual economic phenomena has never been accepted by all ‘theorists’, or at least not consistently so. Hahn, for example, despite currently advising others to turn to computer simulations, has previously castigated attempts to confront ‘theoretical claims’ against empirical data.1 This stance reflects a recognition that the ‘theory’ postulates in question do not at all express the real mechanisms responsible for generating the outcomes recorded in economic data.2 Yet Hahn insists on persevering with the deductivist ‘theory’ project, including theory postulates of the sort in question, despite a realisation that the latter are hardly realistic. Thus, on occasions, or in some contexts at least, it would appear that the constant conjunction form and the deductivist mode of reasoning it underpins are held to be justified despite, or irrespective of, the fact that few if any interesting real world event regularities have been recorded. But if there is a presumption that such ‘theory’ models as abound need not be empirically adequate, how are they justified? How are they thought, or expected, to prove insightful? The category of fictitious idealisations is often held to be significant here. By idealisations I mean limit-types. These are entities, aspects or situations which are characterised by some feature that is perfect, complete, pure, or absolute in some (limiting) sense. Examples abound © 1997 Tony Lawson
in economics. The list includes conceptions of perfect competition, perfect foresight and/or knowledge (omniscience), rationality in the sense of optimising (maximising or minimising) behaviour, complete preference orderings, perfect capital markets, market clearing, and so on. Now the sorts of fictions which figure in mainstream economics are not all of this limiting nature (e.g. assumptions of three commodity worlds or two time periods). The point though is that whether or not they are conceived as idealisations, acknowledged fictions abound in economics. And the question I want to pursue is how any premises that are patently unrealistic can be expected to throw light upon actual phenomena of the real world. Indeed, I believe this to be the single most important question facing the advocate of mainstream economics. Of course, given the discussion of previous chapters, the version of this question of most interest here is how constant conjunction formulations (and the recourse to deductivist ‘modelling’ which a reliance upon such fictions facilitates) can be of any use in understanding phenomena generated within a social system that appears to be insusceptible to significant local closure.3 Two broad types of ‘justification’ for persisting with such formulations are in evidence. The first is to suggest that it does not matter (or at least it is not a total negation of the mainstream project) that few regularities of events have been uncovered, for ‘theorising’ of the sort in question has uses over and above explaining actual phenomena. The second is to interpret the sorts of models which economists formulate as necessary (or anyway useful) temporary devices in a step-wise process of elaboration which should eventually result in models sufficiently complex or realistic to provide successful explanations/predictions of actual phenomena. I now consider each of these types of justification. It should be clear that for each suggested defence a variety of specific positions is conceivable. In consequence I cannot pretend to provide a comprehensive coverage here, merely an illustrative selection of those positions which appear to be the better grounded. Nor do I wish to suggest that only one such strategy is always followed in any contribution focused upon below. I merely consider aspects of particular contributions to indicate the sorts of arguments to be found. In each case, though, it will be seen that there are few reasons for supposing that arguments of this type can succeed.
THEORISING AS AN ALTERNATIVE TO EXPLANATION In what way, if any, can mainstream theory be defended as constituting a project that seeks an alternative goal to the explanation of actual phenomena? Of course, it is trivially true that modellers can accept for themselves criteria of model or theory selection which need have nothing to do with any project © 1997 Tony Lawson
of increasing our understanding of social reality. Examples of criteria that might be invoked are elegance, simplicity, linearity and parsimony. These per se are not my concern. Rather the question of interest here is whether it is possible that mainstream (deductivist) theorising is somehow illuminating of social reality despite a lack of success according to the usual (deductivist) criteria of deducing or predicting actual phenomena of interest on the basis of empirical ‘laws’ and initial conditions. A presumption that I want first very briefly to consider and dispose of is that an understanding of irredeemably fictitious models which are acknowledged as representing only ‘imagined’ (i.e. non-feasible) economics can also provide insight into the workings of, and possibilities for, real economics. My criticism of this position, in brief, is that at the very least the mechanisms theorised have also to be real possibilities. An understanding of structures, mechanisms and conditions that could never be realised amounts to nothing but the comprehension of the formal properties of the models in which they are theorised. Let me briefly illustrate the position I am addressing, for it is common enough. First let us be clear that, typically, ‘theory’ models are indeed irredeemably fictitious. By and large such models constitute formulations of purely hypothetical mechanisms and states of the economy, bearing no relation to real phenomena; they are produced merely to enable certain outcomes or results to be generated. Consider Hahn once more: When a mathematical economist assumes that there is a three good economy lasting two periods, or that agents are infinitely lived (perhaps because they value the utility of their descendants which they know!), everyone can see that we are not dealing with any actual economy. The assumptions are there to enable certain results to emerge and not because they are to be taken descriptively. (Hahn, 1994:246) In short, the understanding which ‘theorists’ most obviously achieve is of the properties of formal models in which assumptions acknowledged to be fictions are embedded. Or, equivalently, an understanding is achieved of an imaginary world in which the assumptions formulated are satisfied. Now, the rather optimistic and frequently observed further presumption that I am primarily concerned with here is that an understanding of such ‘imagined economics’ can achieve more than this, and specifically that such an understanding can facilitate insights into the workings of real economics, or at least their real possibilities. This is a supposition, indeed, that on occasion appears to be accepted by Hahn. Consider, for example, his assessment of the benefits of analysing an economy wherein everyone is assumed to have ‘rational expectations’: © 1997 Tony Lawson
What a rational expectations theory provides is an understanding of an imagined economy which satisfies the assumption. As such it may be of great use. For instance, it allows us to study pathologies which cannot be traced to expectational mistakes. Or again it leads to an understanding of how market variables by revealing information to the uninformed reduce (or negate) the benefits of special information. Or yet again it allows us to grasp the informational disturbances introduced by an unknown monetary policy. It is a vulgar misunderstanding of theory and its aims to dismiss Rational Expectation theorising because of its assumptions. I repeat again that I do not hold this view on ‘as if’ grounds. (Hahn, 1985:11, 12) Now the relevant fact of the matter is that even if this assumption of rational expectations, or any other highlighted assumption, were considered to be of interest (in the sense of being plausible under certain conditions, or whatever), it could not allow us to study the suggested pathologies as real possibilities. This follows because the whole framework of assumptions in which any such a highlighted one is embedded is equally fictitious. Formally, if X implies Y, and we know that X is (mainly) false as a set of claims about actual or possible states of social reality, we can infer nothing about the real possibility of Y. Despite pretensions to the contrary, orthodox ‘theory’ cannot even shed light on the real possibility of situations occurring in society which might be characterised as types of economic equilibrium. Of course, if, in the open social system that is reality, we had grounds for supposing the set of claims denoted X were approximately true or a real possibility, say over some relevant region of space-time, then the reasoning involved could be expected to prove illuminating. In other words, if, under these conditions, we examined the consequences of assuming a counterfactual state of affairs that was nevertheless a real possibility, then some insight into social reality might be expected. But not only is the noted assumption of rational expectations not interesting in this way, in the deterministic set-up of deductivism it is difficult to sustain a notion of real possibilities anyway. In essence, what will be will be. If different patterns or outcomes are viewed as ‘possible’ in the context of some formal model this can only be due to incompleteness of specification and relative to our current knowledge, as we observed in the previous chapter. In such a situation it may be meaningful and revealing to assess the consequences of different sets of assumptions, that is, to explore the range of outcomes or solutions which relative to our current understandings remain feasible. But there is little to be gained by employing assumptions which specify situations which are already known to be nonactual and non-achievable. © 1997 Tony Lawson
At this point the red-herring of clarity and/or precision is often brought into the frame. It is intimated that formalistic modelling is more defensible than contending projects in economics by virtue of its clarity and precision alone. Now, clarity and precision with regards to assumptions, or more generally premises, is not the prerogative of deductivist reasoning. Certainly, any application of the retroductive, including transcendental, form of reasoning requires an explicit prior statement of the premises which are used to initiate the analysis. Nor, of course, is deduction per se ruled out in the latter, or in any other general approach to reasoning. The employment of deductive logic, where it is appropriate, is not the same thing as accepting the deductivist form of analysis (whereby the object always is to deduce specific claims about actualities from accepted ‘laws’ and initial conditions, possibly including its axioms and assumptions). But the essential point here is that if clarity and precision are not the preserve of formal modellers nor are they anything like sufficient for an understanding of real social phenomena. A farmer may, with as much clarity as any ‘theorist’, assume that his or her pigs can fly in a definite direction with a determinate velocity. But this exercise is unlikely to help the farmer in understanding the nature, speed or cost, etc., of any process whereby pigs can actually be brought to market, unless flying pigs are a real possibility. In the end mainstream theorists who adopt a similar position to Hahn are faced with an unpromising set of options. Either the goal is understanding, including explaining, real phenomena, which, given the deductivist framework, necessarily means seeking out event regularities that can be regarded as secure, or it entails abandoning the goal of understanding social reality in favour of the objective of determining (nothing but) the properties of irrelevant models. But if the latter goal of determining the entailments of fictitious irrelevancies seems pointless, that of achieving successful (deductivist) explanations of social phenomena appears hopeless. For those who hold out for successes on the latter criteria the most likely result is, in the manner discussed in Chapters 7 and 8, an everlasting journey of digging ever deeper, or constructing equations that are ever more complex, in the hope, sooner or later, of uncovering ‘whenever this then that’ formulations which successfully cover the phenomena. For those who take the former route of exploring the properties of formalistic ‘models’ these two regresses identified in the last two chapters can be avoided. But this is achieved only at the price of abandoning even the goal of understanding how real world events came about.4
© 1997 Tony Lawson
THEORY AS ONE-SIDED UNDERSTANDING But perhaps this is too quick? Is there not an alternative interpretation of orthodox theorising which avoids these sorts of conclusions? The version which I have examined here accepts that when assumptions recognised as fictitious are employed then, as Hahn (1994:246) puts it, ‘everyone can see that we are not dealing with any actual economy’. But what of the claim often heard and perhaps more frequently accepted, that assumptions recognised as fictitious do nevertheless provide an understanding of actual phenomena, albeit from one side, or under one set of aspects? On this interpretation, when economists assume people to be omniscient, infallible or whatever, they are really getting at the real nature of human beings, but only from one (the economic) side or angle. Economic theory does after all provide an understanding of social reality by getting at the essential nature of one side of it. This is not an argument I find in Hahn’s writings. And, as I have noted, there are unfortunately few modern economists who reflect as openly on what they are doing. Nevertheless, for completeness sake at least this position ought to be considered here. If contemporary economists rarely provide explicit commentary on the nature of their project, it bears repeating that this marks them off from earlier generations. Indeed the writings of earlier economists are a rich source of insight on the nature of, and possibilities for, social theorising. And one in particular stands out for its ingenuity and insight in putting forward the sort of position now in question. I refer to Menger’s (1883 [1985]) Investigations Into The Method Of The Social Sciences With Special Reference To Economics. For its influence alone, both historically, as a contribution to the famous Methodenstreit,5 and more recently, especially amongst strands of Austrianism and ‘new institu-tionalism’ (see e.g. Langlois, 1986:2), Menger’s arguments provide a rather appropriate choice of focus to illustrate the defence in question. Menger’s account is also interesting here in that, like transcendental realism, it is rooted in the Aristotelian tradition. Furthermore, Menger’s analysis takes as a starting point the observation that strict regularities of actual events not only do not occur in the social realm but hardly do so in the natural realm either. Indeed, it will be apparent that in many ways Menger is striving towards the sort of account defended here, making it, from our point of view, perhaps the most challenging or least straw-person-like contribution of its sort.
MENGER’S DEFENCE OF ‘THEORY’ Despite the noted orientation of his work, Menger ends up suggesting that even in the social realm strict laws understood as regularities of event-types, © 1997 Tony Lawson
regularities which in practice are associated with such acknowledged fictions as omniscient, infallible, only economically motivated ‘human beings’, can be formulated which are useful to social understanding. It is the tenability of this claim that I am here wanting to assess. In order to see how it is reached, and justified, it is necessary to consider Menger’s position in some detail. Menger’s Aristotelian starting point is that each concrete phenomenon of reality is typically both an individual of some sort and a member of a general category. Just as each apple, say, is an individual in virtue of its unique nominal characteristics, so each is also an example of a type or empirical form; in the present example of one called apples. Similarly, just as phenomena stand in individual relationships in time and space, say this apple in relation to this tree, so there are general or typical relationships that hold between the various types or empirical forms: all apples grow on (apple) trees. According to Menger these two aspects of phenomena, the individual and the general, facilitate two legitimate orientations of research. The first, the study of concrete phenomena considered under their individual aspect, is the task of the historical orientation of research or science, and the aim is the elaboration of individual knowledge. The second, the study of concrete phenomena considered under their general aspect, is the task of the theoretical orientation, and the aim is the production of general knowledge.6 This distinction applies in the social domain including economics as much as it does in the natural: Also in the field of economy we encounter individual and general knowledge, and correspondingly sciences of the individual aspect of phenomena and sciences of the general aspect. To the former belong history and the statistics of economy, to the latter theoretical economics; for the first two have the task of investigating the individual economic phenomena, even if from different points of view. The latter have the task of investigating the empirical forms and laws (the general nature and general connection) of economic phenomena. (Menger, 1883 [1985]:37) Although Menger holds that all types of knowledge are involved in the comprehension of social life, it is in connection with theoretical knowledge that the question of idealisations and exact event regularities arises. For this reason it is theoretical knowledge that I focus upon here.
THEORETICAL SCIENCE AND EXACT LAWS Now the types and typical relationships (or laws) that are the object of the theoretical sciences are not, according to Menger, equally strict in all cases: © 1997 Tony Lawson
A glance at the theoretical sciences teaches us rather that the regularities in the coexistence and in the succession of phenomena are in part without exception; indeed they are such that the possibility of an exception seems quite out of the question. However, some are such that they do indeed exhibit exceptions, or that in their case exceptions seem possible. The first are called laws of nature, the latter empirical laws. (ibid.: 50) Note that Menger is not making reference to transfactual tendencies (although this presumably is what he is striving for and underpins his intuitions) but to regularities in the succession of phenomena, where some regularities are, and others are not, without exception. Notice, also, that Menger’s claim is not that empirical laws hold in some realms only while laws of nature or exact laws hold in others. Rather, both types of laws are said to hold in all realms marking the sites of two investigative orientations within theoretical research, the realistic-empirical orientation and the exact orientation respectively. The difference is that the former deals with real types, with basic forms and such regularities as are observed in the succession and coexistence of actual phenomena,7 while the latter deals with ‘regularities in the succession of phenomena which do not present themselves to us as absolute, but which in respect to the approaches to cognition by which we attain to them simply bear within themselves the guarantee of absoluteness’ (ibid.: 59). How are the latter exact laws determined? The first step in the process, and it is a fundamental one in Menger’s reasoning, is to establish the simplest elements of reality. Such elements, according to Menger, may not exist in pure form in any concrete example. But they are a necessary basis and presupposition for exact laws. To illustrate what he has in mind here Menger refers to ‘absolutely pure oxygen, pure alcohol, pure gold’ as simple elements of the natural world, and to a ‘person pursuing only economic aims’ as an example for the social realm: [Theoretical research] seeks to ascertain the simplest elements of everything real, elements which must be thought of as strictly typical just because they are the simplest. It strives for the establishment of these elements by way of an only partially empirical-realistic analysis, i.e. without considering whether these in reality are present as independent phenomena; indeed, even without considering whether they can at all be presented independently in their full purity. In this manner theoretical research arrives at empirical forms which qualitatively are strictly typical. It arrives at results of theoretical research which, to be sure, must not be tested by full empirical reality (for the empirical forms © 1997 Tony Lawson
here under discussion, e.g., absolutely pure oxygen, pure alcohol, pure gold, a person pursuing only economic aims, etc., exist in part only in our ideas). However, these results correspond to the specific task of the exact orientation of theoretical research and are the necessary basis and presupposition for obtaining exact laws. (ibid.: 60–1) The second step of the process of determining exact laws of phenomena, consists just in identifying regularities that are held to be associated with these ‘simplest elements’ when the latter are isolated from other influences: In a similar way exact research solves the second problem of the theoretical sciences: the determination of the typical relationships, the laws of phenomena. The specific goal of this orientation of theoretical research is the determination of regularities in the relationships of phenomena which are guaranteed to be absolute and as such to be complete. We have already demonstrated that laws of this kind are not attainable in respect to the full empirical reality of phenomena, and, indeed, because of the not strictly typical nature of real phenomena. Exact science, accordingly, does not examine the regularities in the succession, etc., of real phenomena either. It examines, rather, how more complicated phenomena develop from the simplest, in part even unempirical elements of the real world in their (likewise unempirical) isolation from all other influences, with constant consideration of exact (likewise ideal!) measure. It does this without taking into account whether those simplest elements, or complications thereof, are actually to be observed in reality uninfluenced by human art; indeed, without considering whether these elements could be found at all in their complete purity. In this case it is also aware that a completely exact measure is not possible in reality. Science starts out, however, with these assumptions, since it would never be able otherwise to reach the goal of exact research, the determination of strict laws. On the other hand, with the assumption of strictly typical elements, of their exact measure, and of their complete isolation from all other causative factors, it does to be sure, and indeed on the basis of the rules of cognition characterized by us above, arrive at laws of phenomena which are not only absolute, but according to our laws of thinking simply cannot be thought of in any other way but as absolute. That is, it arrives at exact laws, the so-called ‘laws of nature’ of phenomena. (ibid.: 61)
© 1997 Tony Lawson
EXACT LAWS OF THEORETICAL ECONOMICS Let me illustrate Menger’s approach to theorising in the context of economics. If the objects of theoretical research in the natural realm are thought by Menger to be adequately characterised by such examples as absolutely pure oxygen, pure alcohol, or pure gold, the apparently analogous simple elements of the economic realm (the ‘most original factors of human economy’) are further elaborated as ‘the needs, the goods offered directly to humans by nature (both the consumption goods and the means of production concerned), and the desire for the most complete satisfaction of needs possible (for the most complete covering of material needs possible)’ (ibid.: 63). This last item is already something of a caricature. In fact it soon becomes clear that in focusing upon such supposedly simple elements of ‘human economy’, Menger is contriving a deterministic set-up in just the manner already described in the context of examining the consequences of the doctrine of regularity determinism in Chapter 8. Specifically, he formulates a framework in which human choice is effectively denied or anyway emasculated,8 and the supposed simplest elements of ‘human economy’ are considered as acting in a situation that is closed off from the influence of other factors. Given these presuppositions, Menger draws the implication that the task of the exact orientation of research is to elaborate the strict event regularities such as these considerations apparently facilitate: The function of the exact orientation of theoretical research is to apprise us of the laws by which not real life in its totality but the more complicated phenomena of human economy are developed on the basis of the thus given situation from these most elementary factors in human economy, in their isolation from other factors exerting influence on the real human phenomena. It is to teach us this not only in respect to the nature, but also to the measure of the above phenomena, and thus to open up for us an understanding of these which has a significance analogous to that which the exact natural sciences offer us in respect to natural phenomena. (ibid.: 63–4) By way of an example Menger considers exact research into price determination which, he supposes, teaches us...that the increase in need for a certain item of goods appearing in a definite trade area can under certain circumstance lead to a price increase which can be determined exactly according to measure (whether the increase in demand is the result of population © 1997 Tony Lawson
growth or of the greater intensity with which the need for the item concerned appears among the individual economic subjects. (ibid.: 70–1) What though are these ‘certain circumstances’, or ‘presuppositions’, under which the noted exact law would be observable? Menger lists four of them as follows: 1 that all the economic subjects considered here strive to protect their economic interest fully; 2 that in the price struggle they are not in error about the economic goal to be pursued nor about the pertinent measures for reaching it; 3 that the economic situation, as far as it is of influence on price formation, is not unknown to them; 4 that no external force impairing their economic freedom (the pursuit of their economic interests) is exerted on them. (ibid.: 71) Clearly the initially identified ‘simple’ or ‘elementary’ or ‘original’ factor of ‘human economy’ elaborated as the ‘desire for the most complete satisfaction of needs possible’ has here been idealised somewhat further. In the process ‘economic subjects’ are now also taken to be infallible, omniscient and unconstrained in their actions, with the result that Menger is well on his way to shoring-up the closure conditions identified in the previous two chapters. Menger readily acknowledges that his conceptions are, in effect, no more than fictitious idealisations. Indeed, immediately following his pricedetermination example Menger adds that: in practice ‘people in fact endeavour only rarely to protect their economic interests completely’ (ibid.: 71); they are generally ‘vague and in error concerning the economic means to attain their economic goals’; while their ‘economic freedom is not infrequently impaired’ in one way or another (ibid.). In consequence, Menger acknowledges: the real increase in need for an item of goods will not necessarily have as a result a real increase of prices corresponding exactly to the thus changed economic situation; indeed, there are circumstances in which it will not result in an increase at all. Accordingly, the law that the increased need for an item results in an increase of prices, and indeed that a definite measure of the increase of need also results in an increase in prices determined according to its measure, is not true—is unempirical, when tested by reality in its full complexity. But what else does this prove than that the results of exact research do not happen to find their criterion in experience in the above sense? The above law is © 1997 Tony Lawson
true in spite of all this, absolutely true, and of the greatest significance for the theoretical under-standing of price phenomena, as soon as it is merely considered from the point of view which is adequate for exact research. If one considers it from the point of view of realistic research, then, to be sure, one gets into contradictions. The error in this case, however, lies not in the law but in the wrong way of considering it. (ibid.: 71–2)
THE NATURE OF MENGER’S ARGUMENT What justification does Menger provide for his chosen way of proceeding? Why follow the exact orientation at all? In particular, why suppose that exact regularities associated with a quite fictitious account of the human being can be of help in the understanding of real social phenomena? There appear to be three basic components to Menger’s reasoning on this. The first is a recognition that ‘theoretical’ science seeks general knowledge, knowledge that stretches beyond immediate experience and thereby allows the possibility of influencing the course of events. However, Menger merely asserts that such general knowledge takes the form of ‘laws of the coexistence and of the succession of phenomena’. Such knowledge is held to be required for the derivation of unknown facts from known ones and the determination of various outcomes through setting conditions within our control.9 The second component is an association of hypothetical exact laws of the natural realm with such ‘simple’ or basic elements as illustrated by the examples of absolutely pure gold, oxygen, or iron. There are two strands to Menger’s reasoning here. The first starts from a recognition that at the level of actual things or phenomena there are no strict event regularities in evidence. From this Menger infers that if we generalise at the level of actual phenomena we obtain ‘only a deficient understanding of the phenomena, only an uncertain prediction of them, and by no means an assured control of them’ (ibid.: 59). From this he supposes that exact laws can only be associated with more basic non-empirical items, or types, such as absolutely pure gold. The second strand to Menger’s reasoning is to emphasise that exact laws have in practice been found to be associated with such simple elements, or more accurately that the ascription of exact laws to simple elements in natural science has been found ex posteriori to be successful. The third component is an inference that such possibilities as exist for the natural sciences must also exist for the social sciences. However, an analogy between the omniscient, infallible, unconstrained, only economically motivated agent as the element of ‘human economy’ and the identified ‘simplest’ elements of natural science is asserted but not elaborated upon. © 1997 Tony Lawson
Mainly Menger focuses upon the historical school’s opposition to his specific idealisations (especially the premise that people act only in their economic self-interest) and spends a considerable amount of time attempting to undermine this opposition. The argument here is essentially a negative one. That is, instead of arguing for his particular ‘simplifying’ claims, Menger points out that if, as the historical school argue, fictitious idealisations are to be ruled out a priori just because they are unrealistic, then by that same criterion all idealised claims utilised in the natural realm must similarly be jettisoned.10 However, because, as Menger supposes, natural scientific usage of idealised claims are found ex posteriori to be successful in the context of exact research, the general case against any idealisations, including the economists’ conception of the omniscient, infallible, unrestrained and only economically motivated human being, can thereby be recognised as invalid: If now we return to the so-called ‘dogma’ of human self-interest, which, in the view of the historical school of German economists, is supposed to form such a disturbing contrast to ‘full empirical reality’, there is scarcely need of any further discussion for us to recognize this view as a misunderstanding of justified methodological outlooks which guided the great founders of the ethical sciences in their scholarly activity. No more than pure mechanics denies the existence of air-filled spaces, of friction, etc; no more than pure mathematics denies the existence of real bodies, surfaces, and lines which deviate from the mathematical; no more than pure chemistry denies the influence of physical factors in the formation of real phenomena, or pure physics the influence of chemical factors, although each of these sciences considers only one side of the real world and abstracts from all the rest—no more than these does the economist assert that humans are actually guided only by self-interest or else are infallible and omniscient. He does not, because he makes the formations of social life the object of his research from the point of view of the free play of human self-interest uninfluenced by secondary considerations, by error, or ignorance. (ibid.: 87–8) It warrants a mention that in all of this Menger does not acknowledge any actual successes of the exact orientation in economics. Rather, he notes ‘the really lamentable state of economic theory’ (pp. 78, 79), and emphasises that the immediate goal is to ‘free exact economics from its errors and to fill in its gaps’ (emphasis in the original).
© 1997 Tony Lawson
UNANSWERED QUESTIONS It is clear that Menger’s argument raises as many questions as it answers. Why, for example, suppose that general knowledge which can facilitate capable human intervention must take the form of regularities in the coexistence and succession of phenomena? What, moreover, explains the association of strict regularities or exact laws with the ‘simplest elements of everything real’ (as illustrated by such examples as absolutely pure gold or oxygen)? And if Menger’s strict laws are associated only with such pure or simple elements why suppose that they are able to illuminate a world of more complex phenomena? How is the latter possible? In the event Menger hardly addresses any of these questions. The first is treated as though the answer were self-evident, the second is not addressed at all, and the third is essentially answered by assertion. In the latter case, Menger mainly repeats several times that knowledge of exact laws enables us to understand a particular side of phenomena, and that only a knowledge of exact laws produced by all of the sciences facilitates a full understanding of concrete reality. Thus, while Menger concludes that a knowledge of exact laws is useful ‘for all stages of the complexity of phenomena’ (ibid.: 69), little explanation is provided. At the same time he acknowledges that ‘the results of exact research...are true only with certain presuppositions, with presuppositions which in reality do not always apply’ (ibid.). Specifically they are true only when the identified elements exist in pure form isolated from other influences. But, the manner in which results of exact research are supposed to facilitate an understanding of a reality in which these ‘preconditions’ do not hold is left as something of a mystery. From the perspective of transcendental realism established in Part 1, however, it is easy enough to show how each of these rather vital questions is satisfactorily answered. Moreover, from this perspective we can see that Menger’s failure to address these issues at all adequately is a result of his holding to an overly partial, ultimately untenable, account of natural science. This partiality, in turn, is found to encourage a quite misguided assessment of what carries over from natural to social science. I now briefly run though these claims reconsidering Menger’s analysis of natural science and economics in their turn.
MENGER’S ACCOUNT OF NATURAL SCIENCE It is Menger’s view that natural science is primarily concerned with determining regularities in the coexistence and succession of phenomena. Thus, although Menger ought not to be construed as a positivist, he © 1997 Tony Lawson
nevertheless accepts the structure of science, or scientific laws, which that position encourages. Coupled with this is a failure to elaborate a structured ontology in the manner described earlier, or to do so in any sustained fashion. Yet even Menger’s absolutely pure gold and copper are structured complex things which possess powers by virtue of their intrinsic structures. And such structures themselves must be explained in terms of other causal factors. Whether this failure to elaborate a structured ontology is a cause or consequence of Menger’s holding to the event regularity conception of scientific laws, it is easy enough now to see that it is central to Menger’s inability to explain successfully either the location of his ‘exact laws’ or his perception that ‘theoretical’ science has wider applicability than the simple elements with which his strict laws are associated. Consider the question of why laws are conceptualised as event regularities, and why such ‘exact laws’ might be associated with absolutely pure states of metal and so forth, considered in conditions of their isolation. In Part 1 it was seen that (approximate) event regularities are sometimes achieved, albeit in situations where stable mechanisms are isolated from interfering ones and thereby empirically identified. Typically, outside astronomy at least, such situations correspond to conditions of experimental control. And, of course, Menger’s ‘simplest elements’, as illustrated by absolutely pure gold or alcohol, are precisely examples of the sorts of conditions or results that are (approximately) achieved usually only in the experimental situation. Thus, Menger’s ‘exact laws’ are not really laws at all but event regularities corresponding to the empirical identification of natural laws, namely of the transfactual tendencies that act in open and closed systems alike. Alternatively put, Menger’s ‘exact laws’ must now be recognised as (being formulated in terms of) an epistemologically significant result of human contrivance, not an aspect of reality that does its characteristic thing independent of human manipulation. It is this situation that explains both the significance of event regularities in science and the conditions of their realisation. On these matters any puzzles are now resolved. Nor, from the perspective in question, is the successful application of scientific results in a world of ‘complex phenomena’ a mystery any longer. For once it is recognised that science is concerned to identify and understand structures and mechanisms or their tendencies which, if triggered, go on acting both inside and outside of the conditions of their insulation and empirical identification, any further puzzle on this score is also resolved. If triggered, tendencies are operative, something is really going on, whether or not they are fulfilled. Even as I drive home, I make use of gravitational attraction between me (or my vehicle) and the earth’s surface (and tacitly also of my knowledge of it—I am aware that I will not, at any moment, he propelled into outer space) though at no point do I (necessarily) accelerate uniformly towards © 1997 Tony Lawson
the earth’s surface according to the event regularity encapsulated by the familiar ‘law of fall’.11 It must be admitted that Menger’s account (which shares various intuitions with the transcendental realist account of tendencies defended here) is far from being without insight on these matters. His exact laws can correspond to something real, wherein a stable mechanism or tendency is acting in relative isolation. The problem is that his account is overly partial. It treats the result of a particular or rare occurrence as though its occurrence were the general situation and/or primary objective of science. When a stable mechanism or tendency is triggered in an open system, a transcendental realist understanding of the situation may license a subjunctive conditional about what would have happened if the tendency had been acting in (relative) isolation. But even here the entirety of what we can know and understand and indeed of what is really, i.e. transfactually, going on, is not, and cannot be, captured in this way. Our understanding of a mechanism acting in an open system is captured not by a hypothetical statement about some event regularity that would have occurred if the situation had been different, but by an indicative and categorical statement describing a mechanism’s way of acting. To labour the point yet further, science does not seek to find out how pure copper behaves in a pure state only, and specifically to identify event regularities associated with any such pure state. It’s aim, rather, is to identify copper’s intrinsic condition, its electronic or ionic structure (or whatever), which is essential to all copper and by virtue of which copper’s powers are possessed and tendencies may be expressed. Of course, if a material substance such as copper can be isolated in its purity, any power possessed, say the full potential of copper to conduct electricity, may be that much more easily identified. But if triggered or exercised, any power so gauged will be in play even within materials existing in impure form. Copper’s tendency to conduct electricity, for example, is in play in the pieces of copper that serve as components in the old hi-fi system that I am currently listening to, although countervailing tendencies resulting from impurities modify, or in part determine (along with many other things), the outcome that eventually ensues, the quality of the music that I hear. The final question left without a satisfactory answer by Menger is how the results of ‘theoretical’ science can illuminate a world of complex phenomena. And we can now see that this omission must be attributed to Menger’s error in treating as fundamental that which appears most simple as opposed to that which is most essential (to what is going on). For this has the further undesirable consequence of discouraging, or masking, a recognition that complex things can be agents in their own right, with their own powers, constituting potential objects of scientific study at their own level.12
© 1997 Tony Lawson
Of course, Menger is correct to emphasise that the actual things or phenomena of the world are individuals. Specifically, ordinary things can be conceptualised, metaphysically, as compounds enabling us to make sense of the individuality of historical particulars, just as ordinary events can be conceptualised as conjunctures enabling us to make sense of the uniqueness of historical outcomes. Complex things are rooted in, and cannot escape from, the principles that govern, the components of their make-up. Even so, it is essential to recognise that the ‘lower level’ principles or laws (understood now of course as tendencies) usually only set limits upon and constrain the more complex, including ordinary, things of the world. That is, they do not necessarily determine all that is going on. Space is left for powers that lie at the higher level. My hi-fi system, for example, has a power of sound reproduction irreducible to the capacities of its individual components, a power that, albeit indirectly, makes a difference as to when the lower level laws in fact apply. The hi-fi’s potential, and so its tendency, when triggered, to rouse the sleeping baby or disturb the disgruntled neighbour, determine, in conjunction with my desire to placate the parties involved (amongst many other things), the conditions of its operation, i.e. when the lower level laws are actually exercised. And if complex entities such as a music system influence the conditions under which the physical and other laws apply they cannot be already manifest in the latter but must constitute objects of potential study in their own right. More generally, anything that is capable of producing a physical effect is real and potentially a proper object of scientific study. Of course a music system is humanly engineered and not, as such, an object whose ways of acting are necessarily something to be discovered (although they may be— e.g. a musical instrument). But complex objects need not be of this intended sort anyway. The emergence of organic life, of consciousness, and of society, for example, all constitute putative complex structures (that appear not to be products of any intended project) with powers at their own level capable of acting back upon, i.e. making a difference to, the materials out of which they were formed. All, then, are potential objects of scientific study, and all, of course, like their complexity, are real. In short the directive to seek for ‘exact’ laws at the level of simple states of things in Menger’s account is doubly misleading. For while it misses the point that low-level laws refer to powers and tendencies of novel kinds rather than in principle perceivable things isolated in some pure state along with their associated event regularities; it also discourages any recognition that the legitimate objects of science can be directly located at the level of complex objects and everyday things (although never of course just given in their appearance). © 1997 Tony Lawson
MENGER’S INFERENCES FOR SOCIAL SCIENCE If Menger’s failure to elaborate a structured ontology results in an overly partial account of natural science, the more obvious and salient manifestation of this failure lies in its encouragement of a quite erroneous account of social science including economics. And the basic mistake is apparent: because Menger confuses the (typically experimental) conditions of the empirical identification of a tendency with the ‘preconditions’ for the operation of a law (interpreted as a strict event regularity), he sets off seeking to identify such conditions, or conditions which are considered to be analogous to these, within the social realm. A practical problem which Menger faces with such a strategy is that scientifically significant event regularities of the sort in question, which in the natural sphere are mainly restricted to situations of experimental control, never seem to put in an appearance in the social realm. Of course, from the realist perspective here achieved, there is reason to doubt that they could; there are no obvious ‘conditions’ or ‘presuppositions’ to which they might be attributed. In other words, Menger’s inferences for social science are conditioned upon precisely that aspect of natural science which has least, and possibly not any, bearing for the social realm. The social world, as we have already taken note in Part I, and will explore in some depth in Part III, is highly interdependent: social structure depends upon (always potentially transformative) human intentional agency and cannot be separately isolated under experimental control, or any other, conditions. At the same time, human intentional agency, which includes the power of choice, can itself be exercised only via the means, media and resources of existing structures and conditions. Meaningful experimentation in economics, to repeat, appears to be infeasible. Without any ex posteriori successful location for attributing exact laws in the social realm, then, and lacking an adequate conception of a structured ontology, Menger reasons that economic laws (understood as strict regularities) must, in line with his understanding of natural science, be associated with the simplest or purest states of ‘human economy’. Believing, furthermore, that examples of gold which are absolutely pure may never be achieved, Menger is encouraged in the view that conceptions involving social items of similar unrealistic ‘purity’ can be fruitfully entertained. It is in this light that he assesses that fictions such as omniscience and infallibility can provide an acceptable basis on which social science can proceed. We can now see that such reasoning is fallacious. For science does not seek to isolate actual material things in pure form per se, i.e. merely to understand those pure states of actual things and to elaborate event regularities associated with these states. Rather it aims primarily to understand their characteristic © 1997 Tony Lawson
structures by virtue of which powers are possessed and tendencies may be expressed. The question of naturalism, properly interpreted, is whether specifically social structures, mechanisms, powers and tendencies govern social phenomena in the sense that natural ones do natural phenomena. And once these objects of science are properly identified, it is clear that statements expressing them need be no more idealised than the things or structures themselves. A model of the intrinsic structure of a real atom, for example, is not necessarily more perfect than the intrinsic structure of real atoms. Put differently, if it is correct to recognise, with Menger, that the objects of scientific investigation (such as the atomic structure of copper) are ideal with respect to the actual things of the world (such as individual pieces of copper), it is a mistake to suppose that such objects cannot also be real. For Menger the failure to elaborate a structured ontology necessitates a restriction of focus to, and the ascription of laws at the level of, events and states of affairs or things, and relations holding between them—where, in practice, all such postulated relationships may never actually occur. The position supported here, however, indicates that laws apply instead to really existing structures, and associated mechanisms, powers and tendencies. On the latter conception, but not the former, it is the world itself (including chemical elements, electrons, quarks, along with actions, reactions and fields) not our thought of it that must be recognised (when it is) as ideal. It follows that Menger’s preferred starring points for economics cannot be sustained. It is generic conceptions of agency or structure that ex posteriori have been found to carry over from the natural to the social realm. And while copper’s ideal electronic structure is real and accounts for copper’s power to conduct electricity, conceptions of omniscience or infallibility do not express structures or powers possessed by any human. They are not powers which are possessed but may or may not be exercised, and if exercised as tendencies may be in play unrealised because of countervailing metaphorical blinkers or whatever. They are not real human powers at all. It is an essential aspect of human nature and the human condition that people are fallible, uncertain about their situation and the future, restricted in options and motivated by a variety of concerns depending upon the societal positions and relations in which they stand. Moreover, if this were not so it is inconceivable that most of the enabling institutions of society, including, as Menger elsewhere notes, money, would be (intentionally and/or unintentionally) perpetually reproduced in the manner that we experience. Thus, the error of this particular idealising project in economics is clear. A failure to recognise the real import of certain epistemologically significant situations of science, i.e. those wherein (approximate) event regularities are brought about, means the wrong conclusion is drawn. Specifically, instead of appreciating that, in such situations, science is primarily concerned with empirically identifying and understanding non-empirical mechanisms, it is © 1997 Tony Lawson
erroneously supposed that any event regularity brought about, or which in such conditions could be brought about, or is at least attributed, is itself the scientific object. This misapprehension, coupled with both a recognition that the sort of event regularities considered to be significant in science are rarely observed in an exact form, and an awareness that the scientific knowledge associated with their elaboration does somehow bear general insight for illuminating reality, encourages erroneous ways of proceeding in social research. Indeed, from the perspective in question, one lacking a clear conception of (or a role for) a structured intransitive domain, there is no adequate basis for deterring the imputation of (strict) event regularities to more or less any old counterfactual conditions and constructions, including, notably, conceptions of the human agent acknowledged by everyone as being quite fictitious. Once more, it seems, an attempted defence of the deductivist project in economics cannot in the end be sustained.
THEORISING AS A PRELIMINARY TO SUCCESSFUL EXPLANATION A second broad form of defence of such acknowledged fictions as abound in economics is to argue that they represent not so much a one-sided alternative to explanatorily successful models, as a preliminary stand-in on the path to achieving them. Once more many different strategies might be (and are) devised. But, generally speaking, an unrealistic idealisation is to be interpreted as a heuristic device in a stepwise process of moving from simplified or ideal conceptions to others of greater complexity and so, it is supposed, realisticness.
THE METHOD OF SUCCESSIVE APPROXIMATION Such a process of gradually complicating the picture with the aim of making it more realistic is often referred to as the method of successive approximation. A statement of it is provided by Musgrave (1981): [A scientist] may wish to develop...a theory in two stages: in the first stage he takes no account of factor F, or ‘assumes’ that it is negligible; in the second stage he takes account of it and says what difference it makes to his results. Here the ‘assumption’ that factor F is negligible is merely a heuristic device, a way of simplifying the logical development of the theory. Let us call such assumptions heuristic assumptions. (Musgrave, 1981:383)
© 1997 Tony Lawson
Musgrave illustrates the method of successive approximation with the example of inter-planetary motion: When NEWTON sought to discover what his theory predicted about the solar system, he first neglected inter-planetary gravitational forces by ‘assuming’ that there was only one planet orbiting the sun. He proved that, if his theory was correct, the planet would move in an ellipse with the sun at one of its foci. This assumption was not a negligibility assumption: NEWTON knew that planets would sometimes have detectable gravitational effects on one another. Nor was it a domain assumption: NEWTON was not saying that his theory only applied to one-planet solar systems. You miss the point if you object that NEWTON’S assumption is false, because our system has more than one planet. You also miss the point, though less obviously, if you object that the consequence of NEWTON’S assumption was false, because planets do not move exactly in ellipses. The consequences drawn from heuristic assumptions do not represent the precise predictions of the theory in question; rather, they are steps towards such precise predictions. (ibid.: 383) Within economics this sort of procedure is often attributed to Marshall.13 Even Robbins, who suggests that economic generalisations are known by ‘immediate acquaintance’ expresses support for this sort of approach.14 More recently the method has been employed in modern econometrics, both implicitly in the traditional ‘statistical estimation’ approach in response to prediction failure (see Chapter 7), and explicitly in some of the more recent ‘calibration’ approaches. An equation is said to be calibrated when its parameters are quantified either in such a manner as to accord with some prior theory held by the modeller, or on the basis of casual empiricism, or in the light of some feature of historical data that the modeller wishes the model to ‘mimic’. It is held that after such models have been formulated and quantified on the basis of the original criteria, they can gradually be rendered more complicated in a stepwise procedure of seeking a closer fit with historical data (see for example Kydland and Prescott, 1982, and complicating adaptations in Hansen 1985; Hansen and Sargent, 1988; King and Plosser, 1989; Kydland, 1984; Kydland and Prescott, 1988; and an assessment of the approach in Hoover, 1995a). In recent years the step-wise approach has been described and examined by economic methodologists and philosophers in a number of assessable contributions (for example Cartwright, 1989; Hoover, 1994; Mäki, 1992a; Nowak, 1980; Rosenberg, 1978). In consequence, I will not dwell upon the many specifics of the approach. Rather, I focus on two of its aspects which © 1997 Tony Lawson
seem to me to warrant more attention than they are usually given. Specifically, I want to draw attention to two conditions which, I argue, are essential to the method’s success. These conditions are (1) that the factors considered in ‘isolation’ be real causal factors, structures and/or transfactually acting mechanisms or tendencies; and (2) that the effects of the factors so considered in ‘isolation’ combine or interact mechanically. The necessity of satisfying both conditions has in effect already been established in the preceding examination of Menger’s position. Consider the first condition. In the discussion on Menger’s position it was found that it is only a conception of some tendency’s ‘isolation’ that can, with any legitimacy, be treated as falsely ideal, not the nature of the thing, agent, structure or power itself and/or its way of acting. For it is (at best) the former that can correspond to a real possibility. In the case of Newton’s analysis of interplanetary motion the method is successful, in part, because it is real (gravitational) tendencies that are so considered. By the same token, the lack of any notable explanatory successes in economics is not unrelated to the lack of any real structures, capacities or tendencies well expressed by such conceptions as omniscience or infallibility. Certainly, a process of analysis which starts by focusing upon conceptions of omniscient and infallible economic atoms, and attempts gradually to complicate the picture by introducing human frailties and limitations, could hardly achieve anything of insight or interest. Despite Robbins’ explicit suggestions to the contrary, perfect foresight and the like simply are not ‘tendencies which, in the world of reality, operate only in conjunction with many others’ (Robbins, 1932:94). They are not features of the ‘world of reality’ at all. In the main they are misleading fictions, pure and simple.15 The relevance of the second noted condition, i.e. of the requirement that causal factors considered in isolation interact in mechanical fashion, becomes apparent once we take note of the fact of emergent properties, novel kinds, internally related structures and other forms of complex objects which possess properties at their own level, irreducible to their component parts. For in these cases it makes little sense to attempt to understand their natures by looking at component bits or parts as if existing separately in isolation. Only if the individual factors operate the same way in isolation as they do in combination, and there are no powers of the more complex entity irreducible to those of its parts, i.e. if the various factors are atomistic in their interrelations, does it seem that the method can be expected to be of use. Notice that this second condition is again satisfied in the case of interplanetary motion; gravitational tendencies do appear to combine mechanically. It is because both of the noted conditions are satisfied in the case of the planetary system that Musgrave’s illustration of Newton’s procedure, in this example, works. But the conditions in question, and in particular the requirement that causes combine mechanically, do not hold in © 1997 Tony Lawson
general. Perhaps this is clearest in regard to chemical reactions and combinations. But mechanistic combining is hardly typical of social phenomena either. For example the network of social relations so central to social life cannot meaningfully be broken down into parts with some bits treated as though existing in isolation before others are eventually added back in. It makes no sense at all to treat any feature in isolation from another to which it is essentially related (a situation discussed at length in Chapter 12 below). In studying family behaviour, say, it is clearly quite irrelevant to study conceptions of parental mechanisms apart from conceptions of the nature, including needs, of children or the mutual relationship in which parents and children stand. Equally, it is incoherent to consider the situation of landlords in isolation from (conceptions of) tenants, employers from employees and so on. This seems rather obvious. Yet examples of the error in question abound in economics. It is necessary to think only of the many contributions which as a ‘first step’ affect to model trading nations such as the UK as if existing in isolation, as closed systems with no external countries to trade with. Without the concept of, and without considering the nature of a country’s relationships to, the rest of the world, that country’s intrinsic social, economic and political structures, processes and state of affairs do not really make any sense. Any hypothesised mechanism, theory of growth, conception of policy options, and so forth, elaborated as if a country such as the UK were physically isolated from the rest of the world, can only be a quite irrelevant idealised fiction. The same is true of models which assume away governmental sectors, the labour process, and so on. The feasible alternative to such a step-wise procedure for analysing a nonatomistic world is to treat structured entities at their own level of being. If factors combine organically, or if at some level powers exist that are irreducible to ‘parts’ or to the materials out of which the bearer of powers emerged, then analysis must take place at this higher level directly. Of course, higher order agencies remain subject to laws operating at the lower levels. Conscious human beings, for example, cannot escape the physical laws of matter. But human powers cannot be reduced to such physical laws. It follows that if we want to determine how various non-atomistic factors will combine or interact we may just have to ‘wait (or try) and see’. This conclusion applies not only to such situations as the mixing of specific chemicals for the first time or the trying out of an invented recipe but also to determining the effects of such social changes as the removing of the Berlin Wall, or the (il)legalising of Sunday trading, or the (il)liberalisation of a country’s exchange controls, or the expansion of the ‘world wide web’. Of course it may be possible (and necessary) to intervene continually with the aim of (and with some success in) affecting the process of change. But where organic combination, or transformation in social structure, or real emergence, © 1997 Tony Lawson
or some other form of novelty are involved, there may be especially little of relevance to go on for those intent on forecasting the outcome at the ‘beginning’. Of course, where the situation is complex the analytical focus taken will have to be (especially) partial. But the tenable procedure here is abstraction (a process discussed at length in Chapter 16) not imagined isolation. This involves viewing something (momentarily) under a subset of its aspects, not supposing that they exist in isolation. It involves determining the quality, say, of a new or experimental cake by tasting it after it is baked (i.e. under the aspect of its taste) or by examining its colour or texture or seeing if it sags in the middle. It is unlikely to be sufficient, and may be hardly revealing at all, to proceed by examining the flour and other ingredients in their isolation in the hope of building up successively a knowledge of the final outcome of some hitherto untried recipe. In short, while the usefulness of the method of successive approximation necessitates both that the factors which are examined in (usually imagined) ‘isolation’ be real structures or tendencies and that they combine mechanically or at least in some predictable fashion, in the social realm in particular (as we shall elaborate upon further in Chapters 12 and 16), the latter requirement may be hardly satisfied at all.
THE METHOD OF ISOLATION It will be apparent, in consequence, that although I find a good many interesting and innovative ideas in Mäki’s (1988, 1990a, 1992a, 1992b, 1993a, 1993b, 1994) contributions, and in many ways our projects are obviously similar, I am not especially in sympathy with the ‘method of isolation’ as laid out in Mäki (1992a), at least not if this is to be interpreted as a method that economists are being recommended to adopt. According to Mäki this is a method ‘whereby a set of elements is theoretically removed from the influence of other elements in a given situation’ (1992a:318). Mäki adds that ‘In an isolation, something, a set X of entities, is “sealed off” from the involvement or influence of everything else, a set of Y entities; together X and Y comprise the universe’ (ibid.: 321). Of course, even in experimentation no such isolation occurs literally, only physical re-arrangement. But the aspect of all this that I find most problematic is Mäki’s notion of theoretical or ideal isolation, an apparently ‘traditional forceful procedure…in economics’ wherein no material re-arrangement is involved at all. Rather ‘a system, relation, process, or feature…is closed from the involvement or impact of some other features of the situation’ by way of ‘an intellectual operation in constructing a concept, model or theory’ (ibid.: 325). In fact, Mäki goes further and distinguishes ‘internal’ and ‘external isolation’: ‘In an internal isolation, © 1997 Tony Lawson
one isolates a system from influences coming from within the system, while external isolation closes a system from influences that have sources which are external to the system.’ Mäki adds that both ‘internal and external isolation are relevant in economics’ (ibid.: 326). Now it should be clear that to assert the generalised validity of such isolationist moves would be to commit something like the fallacies of atomism and isolationism identified in Chapters 7 and 8. Not surprisingly, in following through such isolationist moves, and specifically through making reference to certain ‘techniques of isolation’, namely ‘omission and idealization’ (or, in other words, fictionalising through knowingly leaving things out and distorting what is left in), Mäki is able to accommodate a good deal of contemporary orthodox economics (ibid.: 328–34). It is also within this framework that Mäki is able to reinterpret, and imply qualified support for, Nowak’s (1980) version of the method of successive approximation. However, if, as I am arguing (and will indicate more fully in Part III) the social world is not usefully viewed as crypto-atomistic and the various social phenomena cannot be combined mechanically, then, just as contemporary orthodox economics is largely irrelevant, so, as a social scientific device, must be any such ‘method of isolation’. In truth Mäki (1992a) does seem to acknowledge as much in his very last paragraph: This point relates to a major problem involved in the method of isolation as used in studying social and economic phenomena. This is the question of whether the causes of economic phenomena are combined ‘mechanically’ or ‘chemically’, to use J.S.Mill’s phrases. When causes combine ‘mechanically’, their effects can be ‘added up’ like vectors, and the outcome is an additive ‘sum’ or ‘resultant’ of the effects of those causes taken singly. On the other hand, when causes are combined ‘chemically’, some qualitatively novel, emergent outcomes ensue…. It is easier for the user of the method of isolation to deal with the domain of ‘mechanics’ than that of ‘chemistry’. No wonder, therefore, that standard neoclassical economists do their work most of the time as if economics were ‘mechanics’. The challenge they are requested to meet concerns the relative adequacy of the ‘mechanical’ versus the ‘chemical’ metaphysics and of the methods respectively supported by them in the study of the economy. (Mäki, 1992a:349) Perhaps, in consequence, it is best to interpret Mäki’s project here as being primarily descriptive of (mainstream) economists. That is, it may be that above all else Mäki is seeking to render explicit and to clarify the sorts of methods which economists are effectively acting upon, as others have already suggested (e.g. Boylan and O’Gorman, 1995:129). I note, moreover, that in a © 1997 Tony Lawson
very recent paper, Mäki (1994) puts more emphasis (albeit still under the head of the ‘method of isolation’) on forms of reasoning associated with abstraction as traditionally understood, the subject matter of Chapter 16 of this volume. In any case, the point to emphasise here is that although the ‘method of isolation’ as described above may well underpin versions of the method of successive approximation, it does not justify them. For this the social world would have to be crypto-atomistic. And this, I am arguing, is a social metaphysics which, as a generalisation anyway, cannot be sustained. To summarise, for the method of successive approximation to be successfully applied it seems essential that the factors considered separately in isolation be real structures and/or tendencies whose individual effects can be combined mechanically. In consequence, the fictions which pass as ‘economic theory’ postulates, which (we have already seen in the first part of the chapter) cannot be defended as providing an aspect of an alternative to explanation, cannot be justified by reference to this step-wise procedure either. For while the postulates in question rarely express real structures, mechanisms and their tendencies it is simply erroneous to suppose that complex things do typically combine in mechanistic fashion. It depends on the nature of the object in question; society, like nature, can only meaningfully be carved at the ‘joints’ (if and where they exist). Generally speaking the world, including society, is populated with structures of internal relations, their powers and tendencies, processes, mechanisms, and so forth, each operative at its own level. The aim of science is to uncover, identify and understand them. While the ‘economic theory’ concentration upon impossible causal hypotheses appears unable to contribute very much to this, any attempt to gain understanding of real structures by way of looking at component parts as if acting in isolation seems, certainly as a general procedure, likely only to mislead.
© 1997 Tony Lawson
10 SUBJECTIVISM
So far the focus of this part of the book has been on projects or contributions which have reacted to explanatory failure in economics in ways that do not challenge the philosophical perspective which underpins and informs the dominant or mainstream approach. What, though, of those economists who have responded by attempting to provide an alternative approach through challenging explicitly the philosophical presuppositions of the orthodox project? Of particular significance and interest here are those who have distanced themselves from the mainstream approach but without accepting an alternative along the lines of that systematised so far as transcendental realism. Specifically, what are we to make of subjectivism, an increasingly influential anti-naturalist stream of thought in economics? Hayek is the most significant figure here, the subjectivist who has had the most sustained impact. More than fifty years ago Hayek rejected the path of orthodoxy and set about attempting to elaborate an approach of greater relevance. He criticised in particular the reliance of mainstream economists upon positivist methods and procedures borrowed, as he saw it, from the natural sciences, methods which he assessed as contributing ‘scarcely anything to our understanding of social phenomena’ (1942–4:21). And he spent many years attempting to charter a more viable, specifically social scientific alternative project. With apparently increasing numbers of economists reacting to the failure of the modern discipline by turning to the subjectivist tradition in general, and to Hayek’s non-scientistic alternative in particular, it is on the latter that I want to focus here. As is well known, Hayek’s first significant attempt at elaborating a nonpositivist perspective on social theorising is found in his famously influential ‘Scientism and the Study of Society’ essay, published in the early 1940s (Hayek, 1942–4). Equally familiar is the fact that the motivation for the ‘scientism essay’ is indicated in his earlier essay, ‘Economics and Knowledge’, published in 1937. Here, Hayek observes that for the realisation of an
© 1997 Tony Lawson
equilibrium as conceptualised in orthodox theory, there must be a coincidence of the ‘objective real facts’ of the economy and the ‘things known to persons whose behaviour we try to explain’.1 This is a prerequisite which Hayek concludes to be quite obviously unobtainable. And with this rejection of the orthodox premise of universal correct knowledge Hayek turns, in his ‘scientism essay’, to enquire how it is that despite limitations in human knowledge, some kind of order in society comes about. He also investigates the question of whether a science of society is in fact possible. For obvious reasons it is the scientism essay that I now consider at length. The reason I do so here in Part II (where the broad objective is to highlight the errors of contemporary economics which stem from positivistic influence) is that although Hayek aims explicitly to throw off the shackles of positivism, the outcome actually achieved will be seen to be not so much a transcendence of positivism and its errors as a sideways transposition or displacement. In other words, the result is a reproduction of certain errors of positivism albeit in a modified, and specifically a subjectivised, form. By considering Hayek’s account here, the limitations and consequences of failing to develop a structured ontology in particular are once more indicated, albeit in a novel form. One further point warrants immediate emphasis. Although the 1940s’ scientism essay has proven to be particularly influential upon the subjectivist, anti-naturalist, reaction to orthodoxy in economics, it can plausibly be argued2 that Hayek’s social theorising evolved during the 1950s and 1960s into a perspective with some affinities to the account systematised here as transcendental realism (see e.g. Fleetwood, 1995, 1996a, 1996b; Lawson, 1994a; Peacock, 1993). To the extent that this is so and yet this insight remains unrecognised, unfortunate consequences follow. For, where this transformation in Hayek’s thought goes unrecognised, it is common practice to run together passages from Hayek’s writings of the 1960s onwards with those from the earlier scientism essay. In this way, apparently sustainable insights of his later thinking are used to give credence to erroneous, if well known and influential, methodological stipulations found in his earlier essay. This is a practice that can only sustain the confused state of the contemporary discipline. Only when positivism and its errors have been fully transcended, as perhaps in the sort of position that Hayek elaborates later on in his life, can economists free themselves from a reliance upon, or, as Hayek puts it, escape fully the ‘tyranny’ of, methods and ways of proceeding that can contribute so little to social explanation and understanding.
© 1997 Tony Lawson
HAYEK’S THEORY OF KNOWLEDGE AND NATURAL SCIENCE Hayek’s social theory as set out in his scientism essay is first and foremost a theory of knowledge, its nature and limits. An obvious matter to pursue here, then, is the sense, if any, in which it is a significantly different account to the positivist theory, supporting a different conception of science. And I think it must be acknowledged straightaway that Hayek does, to a quite significant extent, cede at least natural science to positivism. Throughout his discussion of the natural sciences Hayek allows for the existence only of ‘events’ or ‘sense-experience’ or ‘sense qualities’ and the like, referring explicitly to the ‘sense qualities which most of us are inclined to regard as the ultimate reality’ (1942–4:31). Any mention of entities not given in experience is sparse and couched solely in the language of ‘constructs’ or ‘creations’ or something similar (see Fleetwood 1995; Lawson, 1994a). Science then is founded on the facts of experience. Specifically, with an empiricist ontology apparently accepted (i.e. with reality restricted to the events, observable aspects of things, and other phenomena of possible experience) natural science necessarily takes the form of elaborating (presumed) constant conjunctions of actual phenomena. Hayek’s particular contribution here, his avoidance of a complete positivist orientation, is to argue that in the process of elaborating these regularities (natural) science provides a classification of ‘external stimuli’ different to that immediately provided by the senses. For example, while we might first classify all white powders together precisely because they are each white and powdery, natural science comes to reclassify the substances in question according to the sorts of events that ensue when these substances are placed in alternative states or conditions. It orders them according to the event regularities (or natural scientific ‘objective facts’) with which they are thought to be associated. Hayek writes: the persistent effort of modern Science has been to get down to ‘objective facts’. (1942–4:29) [Science’s] main task became to revise and reconstruct the concepts formed from ordinary experience on the basis of a systematic testing of the phenomena, so as to be better able to recognize the particular as an instance of a general rule. (ibid.) Science...begins with the realization that things which appear to us the same do not always behave in the same manner, and that things which appear different to us sometimes prove in all other respects to behave in © 1997 Tony Lawson
the same way; and it proceeds from this experience to substitute for the classification of events which our senses provide a new one which groups together not what appears alike but what proves to behave in the same manner in similar circumstances. (ibid.: 31) This process of reclassifying ‘objects’ which our senses have already classified in one way, of substituting for the ‘secondary’ qualities in which our senses arrange external stimuli a new classification based on consciously established relations between classes of events is, perhaps, the most characteristic aspect of the procedure of the natural sciences. (ibid.: 32–3)
HAYEK’S THEORY OF KNOWLEDGE AND SOCIAL SCIENCE In elaborating a social science which is different to natural science, which is Hayek’s stated intention, the theory of knowledge described above is necessarily, if somewhat implicitly, extended. Specifically, in turning his attention to the social realm Hayek allows that human beings, and explicitly social scientists, are able to grasp the contents of other people’s minds, including their attitudes, opinions and beliefs. These latter features constitute the ‘data’ or ‘objective facts’ of the social sciences: the facts of the social sciences are merely opinions, views held by the people whose actions we study. They differ from the facts of the physical sciences in being beliefs or opinions held by particular people, beliefs which as such are out data, irrespective of whether they are true or false... (ibid.: 47) If these ‘facts of the social sciences’ are phenomena which are knowable but not given in direct experience, their knowability is nevertheless closely bound up with experience. For, according to Hayek, it is a fact of experience that we all experience the same thing in the same way. And this is so even if or when we do not classify things as they really are (i.e. according to the classification provided by natural science). We thus recognise the opinions, beliefs and attitudes of others from their actions and doings, and precisely because, or so Hayek suggests, these other individuals experience everything the same way as we do; because we classify all external phenomena in a similar manner: The...fact that different men do perceive different things in a similar manner which does not correspond to any known relation between these © 1997 Tony Lawson
things in the external world, must be regarded as a significant datum of experience which must be the starting point in any discussion of human behaviour. (ibid.: 37) We know, in other words, that in his conscious decisions man classifies external stimuli in a way which we know solely from our own subjective experience of this kind of classification. We take it for granted that other men treat various things as alike or unlike just as we do, although no objective test, no knowledge of the relations of these things to other parts of the external world justifies this. Our procedure is based on the experience that other people as a rule (though not always—for example, not if they are colour-blind or mad) classify their sense impressions as we do. But we not only know this. It would be impossible to explain or understand human action without making use of this knowledge. (ibid.: 43–4) So long as it was naively assumed that all the sense qualities (or their relations) which different men had in common were properties of the external world, it could be argued that our knowledge of other minds is no more than our common knowledge of the external world. But once we have learned that our senses make things appear to us alike or different which prove to be alike or different in none of their relations between themselves, but only in the way in which they affect our senses, this fact that men classify external stimuli in a particular way becomes a significant fact of experience. (ibid.: 48) Hayek’s theory of social scientific knowledge thus turns on the ability of social scientists to grasp the opinions, beliefs and attitudes of other individuals. The implicit empiricist theory of ontology presupposed in positivism is now augmented by the ‘mental phenomena’ noted. We can all access the beliefs and attitudes of others, and we know this to be so from experience. And the reason it is so is the (equally empirical) fact that we all have minds of a similar structure: the facts of the social sciences are merely opinions, views held by the people whose actions we study...beliefs which as such are our data, irrespective of whether they are true or false, and which, moreover, we cannot directly observe in the minds of the people but which we can recognize from what they do and say merely because we have ourselves a mind similar to theirs. © 1997 Tony Lawson
In the sense in which we here use the contrast between the subjectivist approach of the social sciences and the objectivist approach of the natural sciences it says little more than what is commonly expressed by saying that the former deal in the first instance with the phenomena of individual minds, or mental phenomena, and not directly with material phenomena. They deal with phenomena which can be understood only because the object of our study has a mind of a structure similar to our own. That this is so is no less an empirical fact than our knowledge of the external world. (ibid.: 47–8)
SOCIAL STRUCTURE What are the implications for social science? Is there in fact anything more to social science than the grasping of agent conceptions? According to Hayek there is. Specifically, it is held that the social sciences, are concerned with human conscious or reflected actions, where the aim is to trace out their unintended or undesigned results. Before exploring this conception further, however, it is useful to reflect first upon what cannot be allowed, given Hayek’s position. In particular, Hayek’s theory of knowledge does not acknowledge a reality of social structures existing apart from their being conceptualised in action. If something like transcendental realism is eventually embraced by Hayek many years later, a structured ontology is not recognised in the influential scientism essay here in question.3 It is true that in his earlier piece Hayek sometimes (if infrequently) uses the language of ‘structure’ and ‘relations’ and thereby encourages the view that a social theory of the sort that I am here defending is in place in the 1940s. But such an inference cannot be sustained. The passage in the scientism essay that most easily lends itself to, and is most often given,4 something like a transcendental realist interpretation, runs as follows: If the social structure can remain the same although different individuals succeed each other at particular points, this is not because the individuals which succeed each other are completely identical, but because they succeed each other in particular relations, in particular attitudes they take toward other people and as the objects of particular views held by other people about them. The individuals are merely the foci in the network of relationships. (1942–4:58–9, emphasis added) But even this passage is consistent with the reading I am defending. I have added emphasis to part of the quoted passage to highlight the fact that, for © 1997 Tony Lawson
Hayek at this time, relations are the same thing as attitudes; they are particular attitudes that people take towards each other and things.5 Thus, relations here are not to be interpreted in the manner of the (individually as well as scientifically) intransitive real existents as elaborated within transcendental realism, aspects of social structure which, as we shall discuss in Chapter 12 especially, may be poorly, quite inadequately and perhaps falsely comprehended. Even the term ‘structure’ in the above passage means something quite different to its interpretation within transcendental realism. It is, in fact, nothing more than a generic term for the concepts and views that people have formed of each other and things. Indeed, in the passage which immediately precedes the one noted, as in the one which follows on from it, Hayek is quite explicit that this is so. These passages are as follows: The structure of men’s minds, the common principle on which they classify external events, provides us with the knowledge of the recurrent elements of which different social structures are built up and in terms of which we can alone describe and explain them. While concepts or ideas can, of course, exist only in individual minds, and while, in particular, it is only in individual minds that different ideas can act upon another, it is not the whole of the individual minds in all their complexity, but the individual concepts, the views people have formed of each other and of the things, which form the true elements of the social structure. (ibid.: 58) ...and it is the various attitudes of the individuals toward each other (or their similar or different attitudes toward physical objects) which form the recurrent, recognizable and familiar elements of the structure. (ibid.: 59) In short, if Hayek eventually embraces a social theory which is broadly consistent with transcendental realism this is not achieved in the course of his influential 1940s’ work deriding the scientism of mainstream economics. We have seen that Hayek’s social scientific ontology is not restricted to phenomena given in direct experience, but also includes the opinions, beliefs and attitudes of others. However, it does not extend to structures at once both social yet irreducible to individual conceptions, and perhaps inadequately or even falsely comprehended. With this in mind we can now examine Hayek’s conception of the possibilities for social science. Specifically, we can explore how social science is thought to proceed in a manner different to natural science.
© 1997 Tony Lawson
HAYEK’S HERMENEUTICISED SOCIAL SCIENCE The most significant implication of Hayek’s position is that neither human conceptions nor actions are matters to be explained or further analysed in social science,6 but merely items to be grasped. It is this implication which makes Hayek’s position subjectivist or, perhaps better, hermeneuticist. With a theory of knowledge, and so an implicit theory of ontology, that does not recognise structures of determination, conscious action is implicitly, if not explicitly, reduced to the individual’s conscious aims or desires. For, just as there can be no unacknowledged (including inadequately understood) social structures or tacit skills conditioning what takes place, so physical conditions, being once more reducible to (unstructured) events and states of affairs, can have little bearing upon action except in the manner in which they are conceptualised. The last point is significant. For if it is an insight of the subjectivist or hermeneutic orientation which Hayek takes, that social life is conceptdependent, it is its characteristic error, which Hayek also makes, that social life is concept-determined, that it is exhausted by its conceptual aspect. Thus, for social life, for human action and its consequences including social science, the essential features of physical events, states of affairs and other physical phenomena are the ways in which they are conceptualised by the individuals in their actions: In fact, most of the objects of social or human action are not ‘objective facts’ in the special narrow sense in which this term is used by the Sciences and contrasted to ‘opinions’, and they cannot at all be defined in physical terms. So far as human actions are concerned the things are what the acting people think they are. (ibid.: 44) The point is that [definitions of instruments such as hammers] are abstractions from all the physical attributes of the things in question and that their definitions must run entirely [my emphasis—T.L.] in terms of mental attitudes of men toward the things. (ibid.: 46) What is relevant in the study of society is not whether these laws of nature are true in any objective sense, but solely [my emphasis again— T.L.] whether they are believed and acted upon by the people. (ibid.: 51) This essential concern of social science with the manner in which things are conceptualised is held by Hayek to be ‘even more true’ of relations between people ‘which for the purposes of social study cannot be defined in the © 1997 Tony Lawson
objective terms of the physical sciences but only [my emphasis—T.L.] in terms of human beliefs’ (ibid.: 52). And the same sorts of observations are made in regard to such items as human needs. These, indeed, are held to be just what individuals believe they are, in effect, merely subjective preferences. Thus, for Hayek, when other ‘objectivist’ economists make their ‘frequent statements about the objective needs of the people...[the term]...objective is merely a name for somebody’s views about what the people ought to want’ (ibid.: 92). Society then is considered to be conceptual in nature, consisting, ultimately, in the opinions, beliefs and attitudes of individual agents. Hayek allows that lay conceptions include fallible accounts of the natural world, or at least of how things ought to be classified. But he is unable to allow this fallibility to carry over, without strain, to social science. For purposes of social science things just are what people believe them to be. Or rather the fundamental objects of social science just are these beliefs; the scientist is concerned with individual beliefs and attitudes only and the aim is not to explain such beliefs but merely to grasp them.7 Nor is the fact that human beings may wrongly categorise the objects of experience an obstacle to social science so conceived. For whether lay conceptions are right or wrong (according to the classifications provided by natural science) it is always possible for the social scientist to access them and indeed to interpret the phenomena of the world in identical fashion to lay agents, due to the common structure of human minds. The only issue of any conceivable explanatory interest at the level of action is why people see, or perceive, any given thing or phenomenon in the way they do, and in particular in the same way as each other (even if they do not conceive of it as it really is). But such matters, to the extent they can be addressed at all, are delegated to psychology. For the social sciences including economics, the beliefs, desires and so conscious actions of individuals are matters not to be explained but to be grasped. They constitute data that, once accessed, are then only to be arranged; the sole task for the social scientist is to use such data to trace out the undesigned results of the sum total of such conscious actions: It is important to observe that in all this the various types of individual beliefs or attitudes are not themselves the object of our explanation, but merely the elements from which we build up the structure of possible relationships between individuals. Insofar as we analyze individual thought in the social sciences the purpose is not to explain that thought but merely to distinguish the possible types of elements with which we shall have to reckon in the construction of different patterns of social relationships. It is a mistake, to which careless expressions by social scientists often give countenance, to believe that their aim is to explain conscious action. This, if it can be done at all, is a different task, the task of psychology. For the social sciences the types of conscious action © 1997 Tony Lawson
are data and all they have to do with regard to these data is to arrange them in such orderly fashion that they can be effectively used for their task. (ibid.: 68) Moreover, because Hayek so closely identifies conscious actions with individual conceptions, he talks of being able to determine the undesigned effects of human action, of coming to know society, by building it out of the concepts and ideas that people hold: Not only man’s action toward external objects but also all the relations between men and all the social institutions can be understood only by what men think about them. Society as we know it is, as it were, built up from the concepts and ideas held by the people; and social phenomena can be recognized by us and have meaning to us only as they are reflected in the minds of men. (ibid.: 57–8)
SOCIAL ‘WHOLES’ To appreciate the limitations of what Hayek inevitably concludes concerning the undesigned results of human action, results which are frequently referred to as ‘wholes’, it is useful briefly to consider as a bench-mark the insights that transcendental realism can sustain. From this latter perspective social wholes such as society and the economy are not given in experience but presupposed by it. Social wholes on this understanding are obviously not merely a mass of separable events and sequences. But equally clearly, neither are they constituted by the concepts that agents hold. Rather (as I shall elaborate further in Chapter 12) they are typically complex, irreducible, causally efficacious totalities, often inadequately or falsely understood, and continually, if unintentionally, reproduced and transformed in practice. As objects of study, such social wholes can be neither read straight off of a given world nor reconstructed from our subjective experiences. It is the latter result that is significant here. For if Hayek’s hermeneuticism leads him to concur with the former realisation, i.e. that social wholes are not given in experience, and so cannot just be read off, it prevents him from grasping, without strain, the latter insight. That is, given his failure or inability to elaborate a conception of intransitive social structures (including meanings as well as social rules, etc.), i.e. a conception which can sustain the possibility of scientific explanation and so critique, Hayek has little option but to conclude that social wholes are constituted by the concepts that agents © 1997 Tony Lawson
hold. It follows that any such wholes can only be reconstructed from our subjective experiences: While most people will readily admit that in this field there may exist special difficulties in recognising definite wholes because we have never many specimens of a kind before us and therefore cannot readily distinguish their constant from their merely accidental attributes, few are aware that there is a much more fundamental obstacle: that the wholes as such are never given to our observation but are without exception constructions of our mind. They are not ‘given facts’, objective data of a similar kind which we spontaneously recognize as similar by their common physical attributes. They cannot be perceived at all apart from a mental scheme that shows the connection between some of the many individual facts which we can observe... [Social collectives or wholes] do not stand for definite things or classes of things (if we understand the term thing in any material or concrete sense) but for a pattern or order in which different things may be related to each other— an order which is not a spatial or temporal order but can be defined only in terms of relations which are intelligible human attitudes. This order or pattern is as little perceptible as a physical fact as these relations themselves; and it can be studied only by following up the implications of the particular combination of relationships. In other words, the wholes about which we speak exist only if, and to the extent to which, the theory is correct which we have formed about the connection of the parts which they imply, and which we can explicitly state only in the form of a model built from those relationships. The social sciences, thus, do not deal with ‘given’ wholes but their task is to constitute those wholes by constructing models from the familiar elements—models which reproduce the structure of relationships between some of the many phenomena which we always simultaneously observe in real life. (ibid.: 96, 97, 98) Later in the essay Hayek also notes: These ‘wholes’, in other words, do not exist for us apart from the theory by which we constitute them, apart from the mental technique by which we can reconstruct the connections between the observed elements and follow up the implications of this particular combination. (ibid.: 125) In sum, if in reacting against positivist influence in social science Hayek eventually came to subscribe to something like the perspective systematised above as transcendental realism, the latter is not a position sustained in his most influential 1940s’ philosophical essay, ‘Scientism and the Study of © 1997 Tony Lawson
Society’. Instead, the position there achieved is a rather cumbersome and suspect hermeneuticism. And I now want to argue that this currently influential early position of Hayek’s does not so much escape the characteristic errors of positivism as reproduce them in a subjectivised form, in the fashion of hermeneuticist accounts generally. In particular, Hayek’s position must be recognised as a yet further, if now a hermeneutic version of, epistemological foundationalism. In addressing this issue it proves useful to make reference to a recent interpretation of Hayek’s scientism essay which, though sharing some common ground with the present one, nevertheless reaches more or less opposite conclusions. By referring to this alternative account, and specifically through indicating why, in the face of it, I maintain the interpretation that I do, certain essential aspects of my argument can be brought that much more clearly into view.
FOUNDATIONALISM AND HAYEK’S HERMENEUTICISM In a recent contribution Madison (1990) draws out the overtly hermeneutical nature of the position which Hayek sustains in his scientism essay. However, Madison interprets this position as standing in opposition to any (unsound) foundationalism, and in particular to one rooted in atomistic individualism. In himself subscribing to an orientation that emphasises a hermeneutical priority of understanding over explanation, a position which apparently accepts that in ‘economics one is never dealing with “brute facts”’ (Madison, 1990:49) and which rejects ‘the conception of reality operative in objectivistic, foundationalist thinking’ (ibid.: 37), Madison embraces Hayek’s contribution as an early example of such social theorising. In this, it seems, Hayek’s explicit acceptance of the label ‘methodological individualism’ is not what it might first appear. Specifically, ‘far from standing for some kind of atomistic individualism (as it does in the case of many economists) Hayek’s methodological individualism is, in effect, a defence of what might be called the hermeneutical priority of the “social”’. Madison continues: This is fully apparent when one considers that for Hayek the ‘compositive’ approach of methodological individualism is to be contrasted to the ‘resolutive’, analytical approach which according to him characterizes the natural sciences. The latter approach is one which, in accordance with the traditional mechanistic-empiricist-positivist way of proceeding, decomposes complex wholes into simple parts and which seeks to explain the whole (e.g. an organism, an economy) in terms solely of the interactions between its (supposed) constituent parts—the ontological prejudice at work here being that the ‘parts’ are more real © 1997 Tony Lawson
than the whole. From a hermeneutical point of view, it could be said that the whole point of methodological individualism is, or should be, not to reduce the whole to the sum of its parts but to remind us that these irreducible ‘wholes’ are nevertheless not things—to be explained causally—but are, rather, interpreted objects and are not under-standable apart from the categories of human understanding and agency (it goes without saying, of course, that only individuals understand and act). (Madison, 1990:42) It is true that in Hayek’s assessment the analytical approach which is taken to characterise the natural sciences starts out with, and is used to decompose, something referred to as, if not ‘complex wholes’ exactly, then ‘comparatively complex phenomena’ (Hayek, 1940–2:66). But these ‘complex phenomena’ are quite different to the social ‘wholes’ which social science is supposed to elaborate. Indeed, the complex phenomena that Hayek takes to be the starting point for the analytical methods of the natural sciences are precisely the constant complexes of sense qualities that we can simultaneously perceive; they are the events or states of things as given in experience. The social wholes are certainly not phenomena which in Hayek’s view are accessible to direct experience. Moreover, Hayek never suggests, nor does he suppose that positivists claim, that the entities postulated in the resolutive or analytical approach are more real than the more complex phenomena that we may directly experience. On one occasion Hayek does state that ‘the physical sciences necessarily began with the complex phenomena of nature and work backward to infer the elements from which they are composed’ (ibid.: 65–6). More usually, however, Hayek avoids attributing reality to entities in terms of which the phenomena of experience are analysed. Instead, he uses the language of ‘constructs’ or something similar in discussing them. Such conceptual entities (which are not really ‘parts’ at all) are utilised because they facilitate regularities or rules connecting perceptible phenomena, allowing insights concerning such perceptible phenomena that otherwise could not be obtained: The new world which man thus creates in his mind, and which consists entirely of entities which cannot be perceived by our senses, is yet in a definite way related to the world of our senses. It serves, indeed, to explain the world of our senses. The world of Science might in fact be described as no more than a set of rules which enables us to trace the connections between different complexes of sense perceptions. But the point is that the attempts to establish such uniform rules which the perceptible phenomena obey have been unsuccessful so long as we accepted as natural units, given entities, such constant complexes of sense qualities as we can simultaneously perceive. In their place new © 1997 Tony Lawson
entities ‘constructs’ are created which can be defined only in terms of sense perceptions obtained of the ‘same’ thing in different circumstances and at different times—a procedure which implies the postulate that the thing has in some sense remained the same although all its perceptible attributes may have changed. In other words, although the theories of physical science at the stage which has now been reached can no longer be stated in terms of sense qualities, their significance is due to the fact that we possess rules, a ‘key’, which enables us to translate them into statements about perceptible phenomena. (ibid.: 33–4) The main point here, though, is that the phenomena which Hayek describes as the ‘data of the social sciences’, i.e. the concepts, attitudes and views of individual agents, are, as far as social science is concerned anyway, indeed in effect atomistic. They constitute the base elements upon which social wholes are founded and out of which they are, or can be, composed or built up. Notice, moreover, that, for Hayek, these base elements are, despite Madison’s criticisms of any such conception, quite ‘objectivistic’ in the sense of existing ‘out there’ independently of any particular individual. Hayek acknowledges as much explicitly: We can do no better when describing this difference between the approach of the natural and that of the social sciences than to call the former ‘objective’ and the latter ‘subjective.’ Yet these terms are ambiguous and might prove misleading without further explanation. While for the natural scientist the contrast between objective facts and subjective opinions is a simple one, the distinction cannot as readily be applied to the object of the social sciences. The reason for this is that the object or the ‘facts’ of the social sciences are also opinions—not opinions of the student of the social phenomena, of course, but opinions of those whose actions produce the object of the social scientist. In one sense his facts are thus as little ‘subjective’ as those of the natural sciences, because they are independent of the particular observer; what he studies is not determined by his fancy or imagination but is in the same manner given to the observation by different people. (ibid.: 46–7)
HAYEK’S ATOMISIM The peculiarity of the compositive method of the social sciences which Hayek recommends, then, is precisely the condition that the elements which form the © 1997 Tony Lawson
starting point of analysis are not to be explained or further analysed. Rather they are to be treated as the atomistic constituents out of which the more complex phenomena, the social wholes, are to be built up or composed. Such a position is indeed a foundationalism of an atomistic, individualist, sort: A few more remarks must be added about the specific theoretical method which corresponds to the systematic subjectivism and individualism of the social sciences. From the fact that it is the concepts and views held by individuals which are directly known to us and which form the elements from which we must build up, as it were, the more complex phenomena, follows another important difference between the method of the social disciplines and the natural sciences. While in the former it is the attitudes of individuals which are the familiar elements and by the combination of which we try to reproduce the complex phenomena, the results of individual actions, which are much less known—a procedure which often leads to the discovery of principles of structural coherence of the complex phenomena which had not been (and perhaps could not be) established by direct observation—the physical sciences necessarily begin with the complex phenomena of nature and work backward to infer the elements from which they are composed. The place where the human individual stands in the order of things brings it about that in one direction what he perceives are the comparatively complex phenomena which he analyzes, while in the other direction what are given to him are elements from which those more complex phenomena are composed that he cannot observe as wholes. (ibid.: 65–6) It is perhaps also worth noting here that when Hayek considers how one might convey the problems of the social sciences to a physicist, an explicitly atomistic ontology, or anyway one that is somehow analogous to it, is presumed. Notice too that according to Hayek’s understanding any problem of prediction is one not of principle, but of practicality, arising from the paucity of data in our possession: The physicist who wishes to understand the problems of the social sciences with the help of an analogy from his own field would have to imagine a world in which he knew by direct observation the inside of the atoms and had neither the possibility of making experiments with lumps of matter nor the opportunity to observe more than the interactions of a comparatively few atoms during a limited period. From his knowledge of the different kinds of atoms he could build up models of all the various ways in which they could combine into larger units and make these models more and more closely reproduce all the features © 1997 Tony Lawson
of the few instances in which he was able to observe more complex phenomena. But the laws of the macrocosm which he could derive from his knowledge of the microcosm would always remain ‘deductive’; they would, because of his limited knowledge of the data of the complex situation, scarcely ever enable him to predict the precise outcome of a particular situation; and he could never confirm them by controlled experiment—although they might be disproved by the observation of events which according to his theory are impossible. (ibid.: 72–3) In short, Hayek’s achievement in his scientism essay is not so much a transcendence of positivism and its errors as a sideways shift of it all; a move towards a subjectivised version. In place of the ‘brute facts’ of positivism (that Madison rightly criticises) we find, in effect, the brute opinions, beliefs and attitudes of hermeneutical foundationalism. The self-evidence of the empirical world in positivism is reflected in the self-characterisation of the social world in hermeneuticism, the empiricist foundations of the positivistic natural sciences are translated into the conceptual foundations of Hayek’s hermeneuticised scientism. In short, most of the errors of positivism noted by Madison are not so much overcome by Hayek as repackaged in a subjective or hermeneuticised fashion. The hermeneutic tradition is correct, of course, to point out that social science deals with a pre-interpreted reality, a reality already brought under concepts by individual agents. It is also correct to insist, as Hayek does, that the methodological implications of this be elaborated and acted upon in science. Its error, however, lies in reducing social reality to these concepts, and so social science to little more than their unprob-lematic grasping.
THE MATERIAL EMBEDDEDNESS AND INTRANSITIVITY OF SOCIETY Not only does such a project neglect the physical side, including the material embeddedness, of society and economy, it fails to comprehend that aspects of reality can be both social and yet inadequately conceptualised. These are matters that will be of explicit concern to us in Chapter 12. But to anticipate briefly here, we intervene in, we both manipulate and are conditioned by, nature in all our causal interactions with the world, including those with other human agents. The social realm is not a dislocated redescription (or mis-specification) of nature. Rather it is embedded within, and in continuous dynamic causal interaction with, the rest of nature—as the incidence of worldwide poverty on the one hand and human aided ecological disaster on the © 1997 Tony Lawson
other clearly testify. On the conceptual aspect, human action, though not taking place independently of some concept, can and does occur independently of its adequate concept. By this I do not mean only that agents’ conceptions of ‘external’ or physical reality are fallible, a point that Hayek openly admits. I also mean that the immediate own intentions of agents and the meaning of their own acts may be opaque to themselves, frequently at the level of everyday interaction, and systematically at the level of the explanations and descriptions of the reasons motivating their contributions to such interaction. Thus, social science may reveal, as a necessary real condition of some actual social activity, a level or aspect of social reality which, although not existing independently of agents’ conceptions, may itself be quite unknown, or at least inadequately, including perhaps misleadingly, comprehended. Such a feature or level of reality may consist in real relations, or processes, or structural complexes that are really productive of social life but unavailable to direct access by the senses. It may also consist in an aspect of knowledge or motivation that is not, in fact, consciously reflected upon. The source of Hayek’s errors lies in a continued commitment to an essentially empiricist ontology albeit one that is augmented by the concepts and beliefs of others. The solution comes (as arguably Hayek himself eventually realised) with an acceptance of something like the ontology of transcendental realism, and specifically with acknowledging the reality of structure irreducible to events, including social structures that are dependent upon, but irreducible to, the concepts and actions of human agents. In contrast to hermeneuticism such a position is able not only to sustain the existence of beliefs, meanings and attitudes independently of their analysis, but also to situate the possibility of, and indeed facilitate, their scientific explanation and critique. In sum, although Hayek recognises the need to create a social science that is distinct from, and irreducible to, natural science, he fails in formulating his subjectivist alternative to transcend many of the errors of positivism. Lacking a conception of an intransitive social realm which includes structured social objects of knowledge Hayek ultimately fashions an alternative foundationalism. That is, in parallel with the positivist reduction of physical reality to experience, Hayek succeeds only in reducing human society to conceptions. More specifically, although (a) recognising the hermeneutic insight that human conceptions are in some sense productive of society and of concern to social science, and (b) rejecting the Humean premise that subjective beliefs and ‘objective facts’ coincide, that impressions and their (implicit) objects are identified, but (c) lacking an adequate theory of ontology, and specifically of the structured intransitive objects of knowledge, Hayek was more or less bound to conclude that human conceptions and their possible © 1997 Tony Lawson
connections are the sole constituents of society and the exclusive objects of social science. Once, however, a structured ontology is recognised, and it is acknowledged that science properly employs a causal, in addition to a perceptual, criterion of existence, the basis is laid for the errors of positivism to be not merely displaced but effectively transcended. I do not wish to imply that Hayek’s scientism essay is without insight; the hermeneutic moment in science must be acknowledged in particular. Nor, of course, do I wish to deny that there are differences between the natural and social sciences. My argument, rather, is that by maintaining in effect a flat atomistic ontology Hayek is prevented from establishing the real significance of the subjectivist/hermeneuticist insights, while, by essentially abandoning natural science to positivism, he is prevented from locating the real natural/social scientific distinctions. The warranted response, then, is to build on Hayek’s early insights but to elaborate a social ontology and account of social science that transcends rather than displaces positivism’s errors.
© 1997 Tony Lawson
11 THE LIMITS OF CONTEMPORARY ECONOMICS
In the last few chapters, in examining the responses of various mainstream sub-projects to explanatory failure, I have identified, in effect, the broad limits of mainstream economics (as well as examined the questionable nature of the sort of theorising which takes place within these limits). Now the finding that there are indeed limits to what can be pursued by the mainstream project may come as a surprise to some. For a common complaint of its critics is that the project appears to be sufficiently ‘arbitrary’, ‘elastic’, ‘malleable’ or ‘slippery’ that, if the incentive is strong enough, it can be made to accommodate more or less anything. In particular, if any assumption frequently made in orthodox contributions is singled out and criticised for being particularly unrealistic or otherwise untenable, proponents of the project seem always able to replace, adapt, amend or short-circuit the assumption. The nature of the orthodox project appears to be sufficiently arbitrary that its proponents can always come up with some alternative.
CONSEQUENCES OF A MISCONCEPTION OF SCIENCE This latter assessment, then, must now be severely revised. It is at this point clear just how restricted in options the orthodox project is after all. Certainly a variety of specific substantive claims can be entertained. But significant inflexibility derives from the project’s unquestioning adherence to an untenable, if often implicit, conception of science and explanation. For while this (deductivist) orientation allows a variety of substantive hypotheses, the variability in the sort of specifications that are possible is very restricted indeed. In essence the only specifications feasible are those which, in one guise or another, reduce economics to the study of an isolated and stable system of interacting crypto-atoms. In short, whatever flexibility is manifest in terms of its formalistic constructions the fact of the matter is that the © 1997 Tony Lawson
orthodox project has severe problems in seriously sustaining any claims that are to a significant degree realistic. I commenced Part II of the book with an examination of econometrics as traditionally practised because this activity is most easily recognisable as a product of deductivist, and specifically positivistic influence. From the vantage point now achieved, however, each strand of mainstream economics is seen to rest upon the same basic set of influences, including most notably a misconception of the structure of science and its results. At this point then it is possible, and I think insightful, to consider in a rather more systematic way the options open to mainstream economics in the face of its problems and repeated explanatory/predictive failures.
A TYPOLOGY OF RESPONSES TO FAILURE Broadly speaking the continued failure of the deductivist project in economics really allows only three basic types of response (short of escaping positivistic influence entirely and embracing something like transcendental realism). The first is to abandon the requirement of achieving explanatory successes altogether, whether for all science (Menger) or at least for economics (Hayek). However, to the extent that positivistic presuppositions of some kind remain in such endeavours then erroneous moves, oversights and/or unexplained/ unjustified claims inevitably result, as we have seen in the previous two chapters. For those economists (currently the majority) who are not enamoured by such a seemingly unpromising path and are determined to hold out the hope of explanatory successes, only two further options appear open (and are inevitably grasped). The first is to adapt, relax, or otherwise modify the deductivist’s philosophical criterion as to what constitutes a satisfactory explanation. The second is to seize upon some particularity of the social realm that, temporarily at least, necessitates some distance between philosophical presupposition and actual substantive analysis. The former of these two resorts invites a turn to probabilistic forms of reasoning, to the substitution of inductive-probability for deductive-certainty in explanation. In economics this sort of move is well illustrated by the development of econometrics—motivated, as we saw in Chapter 7, by Haavelmo’s perception that ‘economics, so far, has not led to very accurate and universal laws like those obtaining in the natural sciences’ (1944:15). However, while such extensions or modifications of the deductivist approach bring internal problems of their own, ultimately they do not resolve any difficulty. They merely serve to replace the goal of deductive certainty of some (explanadum) event by the (high) probability of that event, or by the © 1997 Tony Lawson
deductive certainty of an average outcome in a larger set of trials or a series of realisations, or some such. The problem with such modifications is that they do not provide any real escape from the requirement of the ubiquity of spontaneous closed systems that characterises deductivism. Specifically, any form of reason-ing that is premised on a ubiquity of spontaneous stochastic closures is equally untenable when confronted with the phenomenon of open systems. In the end, therefore, the only path that appears to hold out any hope for mainstream economics is that of emphasising some extenuating feature of the social world that prevents the deductivist model from meeting with expected success for the time being. This is a response inevitably taken up by inductivestatistical and deductive-nomological approaches alike. Here one or both of two sub-options appear, at first sight, to be available: to emphasise that the invariant event regularities which supposedly characterise social life either are so complex as to be undetected so far, or operate at a level that is in some sense more fundamental or basic than that so far achieved. In other words, failure so far must be put down to difficulties of satisfying, respectively, the conditions for extrinsic and intrinsic closure. These stances seem to capture and render intelligible most developments that are currently under way in contemporary economics. If it is the former that characterises the most familiar response in (traditional) econometrics and now in pure theory (with its turn of attention to computer simulations) of elaborating ever larger systems, it is the latter that explains the recent rational expectationist search for deep invariances and Haavelmo’s quest for autonomous equations, and so on. The method of successive approximation and related approaches might encourage activities on both these fronts simultaneously. Of course, the precise nature of any responses, ‘solutions’ and regresses enacted will depend upon the historical/contextual conditions of their development. But the sorts of moves that are or could be made do appear to be captured by this reasoning. And the problems to which such responses are oriented all stem from a stubborn perseverance with certain erroneous results of positivism, including most fundamentally of all an uncritical insistence upon wielding methods that presuppose a closure, in a forlorn attempt to illuminate what is in fact a quintessen-tially open social system. The way forward, to repeat yet once more, is to abandon the whole misleading positivistic perspective and its results. It is time to fashion an alternative that recognises the reality of an open social system. It is opportune to develop a perspective on the way that social reality is, rather than merely to assume under the sway of the epistemic fallacy that it must conform to the sort of a priori, typically formalistic, methods of which economists are currently, if largely unthinkingly, enamoured.
© 1997 Tony Lawson
Part III
HUMAN AGENCY AND SOCIETY
© 1997 Tony Lawson
12 SOCIETY AND ECONOMY AS REPRODUCED INTERDEPENDENCIES
Although the argument of Part I is (necessarily) pitched at a rather high level of abstraction it has been found to be sufficient to undermine the positivist conception in which the practices of mainstream economists are rooted. Before it is possible to pass commentary on alternative approaches to economics and to assess the real possibilities for economic conduct, however, it is necessary to develop a theory of social ontology at a somewhat more concrete level. I have already accepted as criterial for the social a ‘dependency upon human intentional agency’. Of course, not everyone even accepts that people are intentional beings. This recognition must serve to remind us of the particularity (along with the fallibility and probable transience) of the account that is set out below. Yet without some theory of social ontology any assessment of the possibilities for economic conduct is bound to be more or less arbitrary or, at best, merely conventional.
CRITICAL REALISM In the remainder of this book the general theory of science which has been systematised above as transcendental realism along with the specific theory of social ontology to be developed below (or more generally a philosophy of social science which encompasses it) are collected together under the heading of critical realism. In adopting this heading I am following Bhaskar once more, although various commentators have proposed it previously (for example Sellars, 1916; Drake, 1920; Hicks G.D., 1938). It is a label that I am not entirely comfortable with. But it is now well ingrained in the fast growing related literature (see for example Arestis, 1996; Baert, 1996; Collier, 1994; Dow, 1996; Fleetwood, 1995, 1996a, 1996b; Foss, 1994; Kanth, 1991; C.Lawson, 1994; Outhwaite, 1987; Peacock, 1993; Pratt, 1995; Pratten, 1993; © 1997 Tony Lawson
Rattaggi, 1995; Rotheim, 1993; Runde, 1996; Sofianou, 1995) and this seems a good enough reason to persevere with it, at least for the time being. However, it warrants emphasis here that while the (specific) transcendental realist account so far defended may stand even if aspects of the social theory to be set out below are eventually found to be untenable, alternative social theories (which are consistent with the more general transcendental realist theory of science) may themselves qualify as ‘critical realisms’. The latter possibility arises because the main reason to attach the term critical to the realist theory of social science developed below turns on its recognition of the human agency-dependent nature of social structure. Specifically, because social structure is dependent upon human agency, it is (in a manner to be expanded upon in due course) open to transformation through changing human practices which in turn can be affected by criticising the conceptions and understandings on which people act. The sciences along with philosophy are themselves social structures and practices which are unavoidably susceptible to change through critique. Thus, any emergent science of economics will be part of its own field of study and thereby sensitive to, and, like society at large, ultimately dependent on, social criticism. The point, then, is that if it is a recognition of this situation that legitimises the label in question, it is at least conceivable that competing social theories will emerge which both accept the general transcendental realist account of science and also recognise the dependency of social life on human conceptions (without being reducible to them) and so on critical reason. In fact, it will become clear in Chapter 15 below that my own assessment of the possibilities for social science differs from Bhaskar’s in certain ways. I persist with the label of critical realism, in short, in full recognition that it is an ascription for which numerous conceptions may eventually equally qualify.
ABANDONING SOCIAL ATOMISM What sort of social theory, though, can be put in place of the actualistic, atomistic, conception that characterises mainstream economics? It is clear that the resort to macro- (as opposed to micro-) economics does not provide any real alternative, at least as it is currently conceived. We have seen that mainstream economists, including those who adopt a ‘macro’ orientation, regard it as sufficient to characterise the economy as constituted by the interactions of numerous pre-given, independent, atomistic, economic individuals and the actual states of affairs (structures of preferences and beliefs, asset distributions, inputs and outputs, prices and quantities) which their interactions presuppose and/or are said to bring about. The central question for such economists is whether to proceed by looking at the behaviour of given individuals and their particular situations and explaining aggregate © 1997 Tony Lawson
outcomes on that basis, or by searching at the aggregate level directly, by looking for ‘statistical’ or other regularities in aggregate behaviour. The latter route, which is usually motivated by the practical consideration that ‘data’ at the level of individuals are unavailable, is either left theoretically unjustified or rationalised by such hypotheses that stability is most likely to emerge at the level of average behaviour, and so revealed through summing (averaging) over (presumed well-defined stable distributions of) random atomistic behaviour, or some such. Either way, the basic ontology of social atomism remains along with its epistemological manifestation as a form of reductionism, basically as methodological individualism. That is, economic events and states of affairs are always to be explained by deducing them from, and only from, conjectured principles governing the behaviour of actual individuals, or, as in macro economics, of idealised ‘average’ individuals, and ‘descriptions’ of their situations. In this manner methodological individualism (including its macro-foundationalist variant or inversion) stipulates the material conditions for adequate explanation in economics to accommodate the formal ones necessitated by the latter’s implicit attachment to a positivistic conception of the structure of science, and as laid down more explicitly in its universal acceptance of the deductive-nomological model. Against the social atomism that is entailed by economic orthodoxy’s acceptance of the epistemological claims of positivism I shall, in maintaining the transcendental realist perspective, argue first, that human individuals (of course) cannot be conceptualised as passive and autonomised in experience; and second, that the subject matter of social science including economics cannot be reduced to principles governing the behaviour of human individuals (however conceived) and descriptions of their situations. In particular, in addition to acknowledging the centrality of human beings, human activities and contexts of action to all social life, I want to develop the argument of Part I aimed at establishing the reality (while also arguing for the complexity) of social structures (rules, relations, positions) that are irreducible to people. The conception of the social world to be sustained is of a network of continually reproduced inter-dependencies. That is, social reality is conceived as intrinsically dynamic and complexly structured, consisting in human agency, structures and contexts of action, none of which are given or fixed, and where each presupposes each other without being reducible to, identifiable with, or explicable completely in terms of, any other.
THE ROUTINISATION OF SOCIAL LIFE As a starting point for the argument I focus on the phenomenon that human life is highly routinised. This is a useful entry point in that it is fairly © 1997 Tony Lawson
incontrovertible that the noted phenomenon is a relatively generalised feature of experience. By a routine I mean simply a regular course or manner of proceeding or going on, a recurrent performance of particular acts. Human beings, I am suggesting, are continually engaged in the performance of routines so understood. Moreover, it is apparent that any one routine is often carried out or performed by a wide community, and this can happen over wide expanses of time and space. These observations are frequently recorded of tribal communities but they are equally characteristic of all, including western industrial, societies. To appreciate this it is necessary only to reflect upon forms of communication (greeting, turntaking in conversation); transport (driving on a given side of the road, stopping at red lights); economy (queuing or haggling in the market place, the exchange of coins and notes for commodities at agreed rates of exchange, the enclosure of commodities sold in paper or plastic bags); and so on. Now if it is indeed a generalised feature of experience that social life is characterised by a pervasiveness of repeated (across people, time and space) social practices or routines, something must account for it. That is, it is appropriate to ask the retroductive or transcendental question: ‘What must be the case for such a routinisation of social life to be possible?’ And the intelligibility of the situation presupposes, if amongst other things, the existence of widely known or shared generalised procedures of action, procedures that can be referred to here as social rules. I shall argue that rules (as an aspect of social structure more generally) must be conceived of as something different not only from human beings, but also from the actions in which they engage. That is, although rules are a condition of routinised behaviour they are not the same (sort of) thing.
THE IRREDUCIBILITY OF SOCIAL STRUCTURE But why not conceive of rules (and social structure more widely) merely as the general, or perhaps the structured, aspects of actions or routines? In the UK, by and large, motorists drive on the left-hand side of the road. Why not conceive of the relevant rule just as that generalised aspect of the behaviour of UK motorists that they drive on the left-hand side of roads? On such a conception rules, etc., would remain real and still constitute proper objects of social scientific study. Note, first of all, that people not only conform in their behaviour but some actively rebel. If, for example, rules (however understood) of dress exist for particular events (a party, wedding, memorial service, opening of parliament) then knowledge of these rules allows the possibility of radical disconformity. It facilitates the possibility of intentionally upsetting, insulting, or otherwise © 1997 Tony Lawson
annoying a host or some other party, or at least of making some rebellious sort of statement. Clearly, a knowledge of the (set of) rule(s) is just as pertinent to (intentional) rebelliousness as it is to (intentional) conformity. But why not conclude merely that a rebel is reacting against generalised forms of behaviour? The point is that rule breaking or intentional nonconformity does not merely occur in situations where the majority conform. It is evident, for example, that few drivers on UK motorways travel much of the time at, or below, 70 miles per hour. Yet, as soon as a police car appears on the scene (and frequently when police cameras are evident on parts of the motorway) there is a noticeable decline in speed to below 70 miles per hour— indicating an awareness of a legal limit to (a rule governing) the speed at which a car can legitimately be driven on a motorway, and one that is clearly irreducible to the generalised features of driving behaviour.1 Similarly, if the majority of workers were to ‘work to rule’ most of the time then it would be difficult to render intelligible the ‘threats’ to do so often made at times of overt industrial disputes. Equally the contestability of rules, whether in courts of law or on the shop floor, would presumably be rather more easy to settle if it was always only a matter of deciding on the generalised aspects of actual behaviour. The intelligibility of such situations necessitates that rules be recognised as something other than action, but which condition, and are drawn upon in, action (including being reacted against as well as contested). There is an affinity between the sort of reasoning employed here, and the form of argument utilised in Part I to demonstrate that the natural world is structured and open. In the latter case the motivating observation, the premise of the argument, is that event regularities in the natural sphere do sometimes occur but usually in conditions of experimental control, while results obtained in experimental situations are somehow successfully utilised outside the experimental situation. In the case of social rules, the corresponding observation is that routines do occur, that regular patterns of behaviour are repeated over people, time and space, and yet the understandings implicated in such practices are somehow also successfully utilised outside the (or perhaps implicated in more than one) routinised mode of life—in rebelling, knowing how to break the law without being caught, in destabilising (production), and so on. In each case the noted phenomena can be rendered intelligible only by acknowledging an ontology of structures at least some of the time out of phase with the phenomena (events or practices) which they order and facilitate, but which in certain situations can be brought in phase, giving rise to regularities (constant conjunctions of events or routines) at the level of observable outcomes. Of course, there are also notable differences between the two situations. Significantly, when natural events are out of phase with mechanisms that condition them the effect of the latter tends to be in a ‘consistent direction’. Thus the effect of gravity near the earth’s surface (i.e. © 1997 Tony Lawson
attributable to the mutual attractions of any object to be found there and the earth itself) is always to ‘pull’ any object in question towards the ground. In the case of social rules radical disconformity, for example, can lead to action totally opposed to that explicitly stipulated. But however strange this may seem there is nothing here that negates that general feature of social structures that they are manifest in, but irreducible to, the events or actions they facilitate. The actions of (intentional) conformers and (intentional) rebels alike presuppose the pre-existence (and so relative autonomy), and causal efficacy (and so reality), of social rules as something irreducible to human agency and action. To repeat, rules can be out of phase with the human activities which they facilitate. Rules are drawn upon in action, but are not simply to be identified with it. In their effects rules are always mediated by human agency and any action that results can never be reduced simply to the sets of rules by which it was conditioned. This is so whether rules are explicitly formulated as statutary legislation and backed with formal sanctions or exist as conventions, i.e. maintained via (implicit) general agreement. If, however, rules do tend to be followed regularly, if routinisation is a feature of social life (and it clearly is), such conformity is itself a phenomenon that needs to be explained, an issue I return to in Chapter 13.
SOCIAL RULES Exactly what, though, are social rules? As Thompson (1989) has acknowledged the literature on this is remarkably vague. In the light of the preceding considerations, however, it seems that they are best conceived of as generalised procedures of action that, under a suitable transformation at least, can be expressed as injunctions of the form: ‘if x do y under conditions z’. For example, ‘if driving, keep to the left-hand side of the road, when in twentiethcentury Britain’. The stipulation ‘under conditions z’ will often be dropped in any explicit formulation but will always be implicated albeit unacknowledged. All action, for example, takes place over limited regions of time and space. Note that this formulation is quite general and intended to apply equally to semantic, moral, constitutive, regulative, etc., forms, or aspects, of rules alike. The ‘do y’ in other words is to be interpreted widely and to include such injunctions as ‘interpret...as’, ‘count...as’ ‘take...to mean’, and so on. It is the normative or legitimating or facilitating aspect of rules captured by this formulation that renders them distinct from representations of event regularities as found in the experimental situation. That is, the significant results of science that do take the event regularity form can be formulated as © 1997 Tony Lawson
‘if event x then event y, as long as conditions e hold’—where e typically denotes a well controlled experimental situation. The point here is that if there does appear to be a similarity of form between the structure of social rules and the noted manner of reporting results in natural science, a significant difference is that social rules, typically, carry only normative or legitimatory (constitutive/regulative/moral/semantic) force. A social rule, in other words, and as already indicated, is a formulation of action that, under specified conditions, must, should, or can usefully, legitimately, meaningfully, or advisedly, etc., be carried out, rather than a prediction or observation of an action. It is a (possibly contested) directive, code, or understanding about how an act could or should be performed; it is not per se a prediction or claim that the performance so indicated in fact proceeds. For the sets of rules in fact to be utilised, whether in conformity or otherwise, or in a manner contested or uncontested, human collusion and mediation must always be involved. But all of this is just to repeat in an alternative form the recognition that social rules govern, condition, limit, facilitate and can be drawn upon in, but cannot be reduced to or identified with, action.
SOCIAL RELATIONS AND SOCIETAL PRACTICES Social life, then, is highly rule-governed. The routinised nature of social life testifies to this. But this is not all there is to social structure as conceived here. In particular, social life must be recognised as also highly relational and, in part, constituted by positions which people occupy. The fact and significance of social relations and positions are seen to follow once we take note (and enquire into the conditions) of a further general feature of experience: that there is a systematic disparity across individuals regarding the practices which are, and apparently can be, followed. Although most rules can be utilised by a wide group of people it by no means follows that all rules are available, or apply equally, to everyone. To the contrary, society is highly segmented in terms of the obligations and prerogatives that are on offer. Teachers, for example, are allowed and expected to follow different practices from students, government ministers to follow different ones from lay-people, employers from employees, men from women, landlords from tenants, and so on. Rules as resources are not equally available, or do not apply equally, to each member of the population at large. What, then, explains the differentiated ascription of obligations, prerogatives, privileges and responsibilities? This question directs attention to the wider one of how human beings and social structure such as rules come together in the first place. If social structure is a different sort of thing to human beings, human agency and even action, what is the point of contact © 1997 Tony Lawson
between human agency and structure? How do they inter-connect? In particular how do they come together in such a manner that different agents achieve different responsibilities and obligations and thereby call on, or are conditioned in their actions by, different social rules and so structures of power? If it is clearly the case that teachers have different responsibilities, obligations and prerogatives from students, and government ministers face different ones from the rest of us, it is equally apparent that these obligations and prerogatives exist independently of the particular individuals who happen, currently, to be teachers, students or ministers. If I, as a university teacher, were to move on tomorrow, someone else would take over my teaching responsibilities and enjoy the same obligations and prerogatives as I currently do. In short, society is constituted in large part by a set of positions, each associated with numerous obligations, rights and duties, and into which agents, as it were, slot. We can determine something more about this system of societal positions if we take note of the additional observation that practices routinely followed by an occupant of any position tend to be oriented towards some other group(s). The rights, tasks and obligations of teachers, for example, are oriented towards their interactions with students (and vice-versa), towards research funding bodies or governing institutions, and so forth. Similarly the rights and obligations of landlords are oriented towards their interactions with, or in respect of, tenants, and so on. Such considerations clearly indicate a causal role for certain forms of relation. Two types of relation must be distinguished: external and internal. Two objects or aspects are said to be externally related if neither is constituted by the relationship in which it stands to the other. Bread and butter, coffee and milk, barking dog and post-person, two passing strangers, provide examples. In contrast, and developing the discussion of Chapter 6, two objects are said to be internally related if they are what they are by virtue of the relationship in which they stand to one other. Landlord and tenant, employer and employee, teacher and student, magnet and its field are examples that spring easily to mind. In each case it is not possible to have the one without the other; each, in part, is what it is, and does what it does, by virtue of the relation in which it stands to the other. Now the intelligibility of rule-governed and the rule-differentiated social situation noted above requires that we recognise first the internal relationality of social life, so that the atomistic conception underpinning mainstream economics must, as a generalisation certainly, be rejected; and second that the internal relationality in question is primarily not of individuals per se but of social positions. It is the positions that are defined in relation to others, say of teachers to students. The picture that emerges, in other words, is of a set, or network, of positions characterised by the rules and so the practices associated © 1997 Tony Lawson
with them, where the latter are determined in relation to other positions and their associated rules and practices. On this conception the basic building blocks of society are positions, involving, depending upon, or constituted according to, social rules and associated tasks, obligations, and prerogatives, along with the practices they govern, where such positions are both defined in relation to other positions and are immediately occupied by individuals.
SYSTEMS AND COLLECTIVITIES I have so far focused upon such aspects of social structure as rules, practices, relationships and positions without any explicit attention to notions of social systems or collectivities. These, however, can now be elaborated quite straightforwardly. For the conception of social systems and collectivities that is supported in this framework is precisely of an ensemble of networked, internally-related, positions with their associated rules and practices. All the familiar social systems, collectivities and organisations—the economy, the state, international and national companies, trades unions, households, schools and hospitals—can be recognised as depending upon, presupposing, or consisting in, internally related position-rule systems of this form. Sub-distinctions can be drawn. While a social system is best conceived of as a structured process of interaction, and perhaps an institution2 is most usefully viewed as a social system that has been found to be (relatively) enduring, a social group or collectivity can be understood as consisting in, or depending upon, or as a set of people distinguishable by, their current occupancy of a specific set of social positions. Notice that at any one time a particular individual will occupy any number of positions. That is, the same person may be a parent, child, worker/ boss, teacher/student, immigrant/native, male/ female, old/young, member of religious or political or community organisations and so on. The resulting conception then is one that renders intelligible the often noted, but reputedly difficult to sustain, sense of a group or collective interest and thus the basis for a theory of collective action; and yet allows the possibility of a conflict of interest at the level of individuals. Put differently, on this relational conception any specific collectivity can be understood in terms both of its relations to other groups, especially those against which it is defined and/or is opposed, as well as of the complex of internal relationships within the collectivity itself. Amongst the many advantages of this conception is the feature that, in stark contrast to mainstream economics, it allows a meaningful focus not only upon production and exchange activities but also upon a range of distributional issues as well, such as resources to groups as well as people to positions. The general conception being argued for, then, is that of society as a network of internally-related positions and associated rules and so practices. © 1997 Tony Lawson
But there is an obvious lacuna still to be filled in here. For, it is clearly a presumption of the above account that many of the societal positions which individuals occupy and rules associated with them are relatively enduring. Indeed, it is clear that they must be articulated across sometimes significant stretches of time (as well as space). However, the whole question of the basis of any fixity of social structure, and, more generally, of the relationship of social structure to human agency, has yet to be broached in any detail. These are issues that must be addressed. Before turning to such matters explicitly though, one quick cautionary remark concerning the picture sketched above is necessary. It is clear that corresponding to the notions of external and internal relations are two polar conceptions of the nature of social reality, conceptions that can be labelled atomism and organicism respectively. According to the former conception, reality is made up entirely of externally related entities so that all things exist and act in ways that are quite independent of any relationships in which they stand. According to the second stereotype, reality is made up solely of internally related entities so that essential, or constitutive, aspects of any particular entity can only be determined from a knowledge of the relationships in which it stands. At the level of ontological analysis per se there are no a priori reasons to follow either the (Humean) path (which characterises contemporary mainstream economics) of assuming that all relations are external or the (Hegelian) route of insisting that all are internal. The relational situation in any context can be determined only through substantive research. Ex posteriori, social life does appear to be highly internally related, as the elaboration of the position-rule system above seems to testify. Even so it must be illegitimate to generalise from these considerations to all forms and aspects of social situations. Moreover, even where structures do combine or cohere in an internally related complex the latter may contain both atomic and organic connections. In fact, whenever structures cohere as a totality, causal interaction between any given pair may, in principle, be both external and identity constituting. Finally, it is necessary to recognise that although any two entities or aspects may be internally related and so existentially on par, this does not entail that they must carry equal causal force. A prime minister or parent or employer may respectively dominate various ministers or children or employees while the latter nevertheless remain essential to the existence, powers and activities of the former.
THE AGENCY STRUCTURE RELATIONSHIP The conception of society so far elaborated is intelligible on the condition that positions, rules and relationships do to some extent endure over stretches © 1997 Tony Lawson
of time and space. Ex posteriori, of course, it is clear that they do. Although nothing is fixed and unchanging it is difficult to deny that positions such as those of teachers and students, landlords and tenants, or, say, rule structures implicated in languages or highway codes, are enduring to some extent. The persistent but differentiated patterns of routinisation indicate this. The interesting question, then, is how does social structure endure (and change) to the extent that it does? Is it an external given, operating, as it were, behind the backs of individuals, or a perpetual creation of individuals, open perhaps to extensive manipulation, or something else? Indeed, what exactly is the nature of the relationship between human agency and social structure that is here being argued for? If the point of contact between agency and structure tends to be societal positions, what is the manner of their interrelation? It is essential to address the last question if we are to understand how certain structures endure. Now it is not, I think, an exaggeration to suggest that most accounts in economics that explicitly focus upon the agency-structure relation veer towards one or the other or both of the two just noted poles. That is, either (1) structure is reduced to (is conceptualised as the mere creation of) individuals, or (2) agency is reduced to (is conceived of as being totally determined by) external, coercive, structure. Voluntarism or determinism. These are the usual competing hypotheses. Even the more insightful of non-orthodox economists frequently fail to avoid some such reductionism. Thus, while Hayek, amongst others, reduces society to human conceptions (at least in his 1940s’ scientism essay), Veblen, as perhaps the best of the (old) institutionalists, is not alone in sometimes reducing agency to its conditions, in what amounts to a technological determinism. And just as when economists seek to apportion blame or emphasise policy mistakes they tend towards the voluntarist extreme, so when they formulate their econometric models the deterministic tendency comes to dominate. Both polar conceptions must be rejected as untenable. First of all, social structure such as rules, positions and relations are a precondition for intentional action. Because social structure is drawn upon in, and so presupposed by, action it must pre-exist it. Because it makes a difference to action and so to the states of affairs brought about, it must be real. And the point here is that because human intentional agency presupposes already existing social structure the latter cannot be regarded as the mere creation of individuals. The subjectivist/voluntarist reduction of structure to agency then must be rejected. At the same time, if the human race were to disappear tomorrow social structure would disappear along with it. Social structure depends upon human beings. And human beings are able to act intentionally, to exercise choice. Thus, structure can hardly be regarded as fixed, as externally coercive, and © 1997 Tony Lawson
so reified. The objectivist reduction of agency to structure must therefore also be ruled out. Note moreover, that a solution is not achieved merely by combining these errors of voluntarism and reification/determinism. It makes little sense to posit at one and the same time a voluntaristic idealism whereby individuals create the social structure and a mechanical determinism whereby people are essentially the product of their situation. Yet this, I think, is a danger that is encouraged by the recent mainstream or ‘new institutionalist’ concern to focus more on structure. It comprises an attempt to combine, on the one hand, something like situational analysis and the ‘single exit model’ approach in order to provide a deterministic account of individual behaviour, with, on the other hand, an approach aimed at repeatedly creating social structures out of beliefs and/or actions of individuals.3 If individuals are reduced to social structure and social structure is reduced to individuals all that is achieved is an inappropriate, quasi-dialectical, account which fails to sustain the insight that structure and agency are distinct, albeit highly inter-dependent, things. But what is the nature of the relationship now in question? If Hayek’s 1940s’ hermeneutic thesis that society is concept-determined is unsustainable, his insight that society is nevertheless dependent upon human agency including human conceptions must be retained. But if society, and social structure more generally, depend on things other than human conceptions what, if anything, can be said about their manner of existence? In the light of the discussion of relationships, positions, and social totalities above, it must by now be clear, as arguably Hayek himself eventually acknowledged (i.e. twenty years on from the scientism essay), that social structures can exist without being adequately conceptualised by agents, or without being conceptualised at all. In what sense then do they depend on human agency? The answer is that social structures also depend upon human doings, that is, human activities or praxis. It is through human actions that specific social structures come about and endure (when they do) whether or not individual agents have an awareness that, or precisely how, this is so. This is true not only for structures that are not, or are poorly, or falsely, understood, but equally for matters known at best only tacitly (an issue developed in Chapter 13). Thus, irrespective of whether or not individuals are, for example, discursively aware of (i.e. could explicitly elaborate) the rules of grammar of their language, the existence and any endurability of such rules is undoubtedly dependent upon speech acts in which people engage. The consequence of all this is that a switch of emphasis in social analysis is necessitated, away from the traditional conceptions of creation and determinism to notions of reproduction and transformation. Human intentional activity does not create social structure for, to repeat, the latter is © 1997 Tony Lawson
presupposed by such activity. Instead individual agents draw upon social structure as a condition of acting, and through the action of individuals taken in total, social structure is reproduced or (in part at least) transformed. Equally, though, social structure cannot be reified. Only at the moment of acting can it be interpreted as given to any individual. Through individuals drawing upon it in (always potentially transformative) action, structure is continually reproduced or modified in form. In their daily activities human beings draw upon social structure which, in turn, is reproduced or transformed through human action taken in total. Although human acts may sometimes be performed with the intention of (1) reproducing structure (speaking to a child with the intention of imparting knowledge of language) or (2) transforming structure (collective attempts to change some feature of the current economic or legal system), it is likely that most structural reproduction and/or transformation arises as an unintended product, whether or not desired or even recognised. Of course, if the reproduction/transforma-tion of social structure is only rarely recognised by individuals as their reason for acting in the way they do, individuals usually have some motivation for, and conception of what they are doing in, their activity. Human doings are mostly intentional under some description (see Chapter 13 on human agency). Even if most speakers of English, say, are not intending, in their individual speech acts, to reproduce that language, its reproduction nevertheless is the sum result of the speech acts in which English speakers engage, just as the speech acts in which individual agents engage always have their own intended objectives. If the reproduction/transformation of social structure is rarely an intended project, it is equally the case that the individual agents are not always aware, certainly not discursively or self-consciously so, of the structures (such as language rules) upon which they are drawing. The picture that emerges, then, is one of largely unmotivated and only partially grasped social reproduction. Individuals draw upon existing social structure as a typically unacknowledged condition for acting, and through the action of all individuals taken in total, social structure is typically unintentionally reproduced. Social structure in general is neither created by, nor independent of, human agency, but rather is the unmotivated condition of all our motivated productions, the non-created but drawn upon and reproduced/transformed condition for our daily economic/social activities. If the dual feature of social structure that it is both condition and consequence of action is, following Giddens, aptly termed the duality of structure, then, following Bhaskar, that dual feature of action that it constitutes both motivated production and unmotivated (structural) reproduction, is appropriately labelled the duality of praxis. The term duality here, as opposed to ‘dualism’ with its usual connotations of complete © 1997 Tony Lawson
separation and insulation, denotes how structure and action can each be viewed from two contrasting perspectives. Structure is both condition and consequence, while the consequences of action (including inaction) are both motivated and unmotivated. And although in all this it is clear that structure and human action presuppose each other it remains the case that they are different things; neither can be completely reduced to, identified with, or explained completely in terms of, the other.
SOCIAL CHANGE It is important that I draw out the notion of change that is involved here and its relevance to everything under consideration. For by recognising the centrality of action and the processes of social reproduction/ transformation to all aspects of human life including society and economy, an intrinsically dynamic perspective for social science is seen to be unavoidable. Moreover, from the perspective sustained, but not from that of atomistic empirical realism, it is possible to distinguish different kinds of change, including accidental from necessary. It is not in the nature of a puppy to break its leg but it is in its nature to grow into a dog. The latter is not, but the former is, an accident. Growing into a dog is a necessary tendency governed by the animal’s genetic code. Breaking the leg is an event due at least in part, to (or is seen from a particular perspective to be a result of) some non-necessary external factor. Empirical realism, resting as it does on the category of events or states of affairs, has no basis for distinguishing the accidental from the necessary. Any event is merely an accident in the passing of time. Events may be held to be constantly conjoined. But without an account of natural necessity, of a mechanism that accounts for such a reputed empirical regularity, the latter itself can be entertained only as an accident. Perseverance with it, basing prediction upon it and so forth, can be accounted for, or ‘justified’, merely as convention. Empirical realism is characterised by the blanket category of events where change and repetition alike are only and always merely accidents. Notice that change in the sense of the puppy developing into a dog is part of what being a dog is. In similar fashion change is integral to everything else including social or economic objects, structures and systems. We have seen that although social structures are a condition for individual acts it is only through the actions (including inactions) of individuals in total that any social structure endures when it does, that it is reproduced or transformed. Social structures, in consequence, can not be regarded as somehow fixed; they can never be reified. There is always fluidity and movement. Even when, over a restricted region of time and space, a structure might be held to have been © 1997 Tony Lawson
reproduced ‘intact’ as it were (and, of course, given the holistic nature of much social structure all reproduction doubtless entails some change, just as change will rarely be total), this is always on the basis of intrinsically dynamic and always potentially transformative, human practice. In the event, any structure, whatever its level, is only ever relatively enduring, being geohistorically grounded. Like everything else it is spatially situated and becomes and begoes in time. Change, then, (as with continuity) is endemic to social life. Systems and structures, from language and education systems to relationships between men and women, or young and old, are constantly evolving. In consequence, a social object such as the market is not usefully regarded as something that exists and as something that experiences change, as if its existence and change were two entirely separate aspects of it. Similarly, change cannot be reduced merely to contrasting states of affairs such as structures of market prices on different days, or voting patterns in succeeding general elections. Rather, social items such as markets, political systems and households must be understood as processes, as reproduced structures of interaction, with change recognised not as (or not only as) an external happening, the result of an external or exogenous shock, but as an integral part of what the system or object in question is. It follows from this transformational conception of social activity that if change, or at least the potential for change, is always present, the analysis of change per se is of no greater (or less) significance to social explanation than the understanding of continuity and reproduction. Both (relative) change and (relative) continuity need to be addressed and accounted for in science.4 Ex post, to be sure, certain structures are found to exhibit greater continuity than others. But this does not license any view to the effect that the relatively more enduring structures are somehow more natural, normal or given.5 Put differently, there is no ontological asymmetry between relative continuity and relative change. It is a mistake to regard any social structure or system as something that would somehow reproduce itself unless obstructed from doing so. In consequence, there is equally no epistemic or logical asymmetry between the explanation of relative continuity and relative change. Tradition or fashion, routine or novelty, habit or deliberation, imitation or invention, preservation or innovation, conservation or subversion, lock-in or lock-out, are all in principle open to, and require, social explanation. In the social world, change and continuity are on par as real possibilities and as phenomena that can and should be accounted for in the laborious social practice of science.6
© 1997 Tony Lawson
THE RELATIONSHIP OF SOCIAL RULES TO POSITIONS Finally, in the light of the perspective now established on the positionality and dynamism of society, some further comments on the nature of social rules ought at this point to be made. I have suggested above that social rules can be interpreted as generalised procedures of action which, under certain transformations, can be represented as (implicit) injunctions of the form ‘if x do y, under conditions z’ (although, I should emphasise, I have not suggested that agents acting on social rules typically represent them in this manner or indeed hold any explicit representation at all—see Chapter 13). I have also suggested that rules are often associated with positions. Now there is an ambiguity in the literature as to whether or not any positions referred to are best conceived of as part of the rule formulation, i.e. as part of the conditional ‘if x’, or part of the ‘under conditions z’, clauses. For example, is it the case that the rule is ‘in twentieth-century Britain drivers must keep to the left’, or that, for such motorists, the rule is ‘keep to the left’. Is the rule about the position or attached to it? Do given rules affect different people differently (in virtue of the positions people occupy) or do different people (in different positions) follow different rules? I am not sure that anything hangs on the distinction in practice. Indeed, as the referents of any such expressions are aspects of social structure which are neither reducible to their representations nor typically explicitly formulated, each form of representation may well be equally valid. In terms of presentation, though, each expression has its advantages and it is the insights of each that I want to emphasise here. The second formulation (according to which rules are attached to positions) serves most straightforwardly to remind us of the segmenta-tion of social life, of its inequalities of opportunities, of power, of rights, etc., of the different obligations and prerogatives built into it. The first formulation (according to which rules are about positions) serves to remind us that social rules associated with specific positions are not merely reproduced and transformed by the immediate occupants of the positions in question but by the actions, inter-actions and inactions of a far wider grouping—basically all of those in a position to be affected by or to influence them. Sometimes the immediate occupants of a position in question may have very little ‘say’ at all in the matter. In the UK, for example, (although clearly not in many other countries) a convention that appears to be widely accepted by parents and TV companies is that children below a certain age should be in bed (or anyway not watching TV) after 9 p.m. Yet it is not at all clear that children themselves wish such a rule to be in place. A further ‘rule’ that is (still) often heard and perhaps even more often acted upon is that ‘a woman’s place is in the home’. This rule is certainly (and sometimes actively) reproduced by some women, but always with the covert aid of numerous © 1997 Tony Lawson
other parties, including many men, children’s story books, other media, and so forth. In short, it probably does not matter whether we envisage rules as attached to, or about, positions, or both. The point is to acknowledge the positionality of social life, and recognise that if rules are attached to positions, their conduct is affected by (and affects) the actions of people situated more widely. To sum up the discussion of this chapter, if we can agree with Marx that human beings ‘make their own history, but not in circumstances of their own choosing’, we must emphasise not only that the circumstances and history in question are not fully understood intentional products, but equally that social analysis must focus upon the making (upon action or praxis and change) as much as (the socially-positioned) makers. Society and economy are perpetually reproduced and/or transformed through praxis. The human agent, nevertheless, if to some extent now ‘decentred,’ cannot thereby be neglected. Rather, the human agent, along with human subjectivity, must be reconceptualised in the light of the emerging perspective. It is the topic of human being, subjectivity and agency that I turn to consider next.
© 1997 Tony Lawson
13 A SKETCH OF THE ACTING SUBJECT
What sort of conception of the human subject is entailed (or presupposed) by the foregoing discussion? In moving to address this question it may be useful if I first clarify my understanding of various central categories usually employed in discussions of this topic. After doing so I turn to focus upon relevant substantive issues. By an agent, I understand a possessor or bearer of powers, abilities and capabilities. By human agency I mean the specific powers and capabilities of human beings. By human acts or action I understand the intentional exercise of human agency, i.e. intentional human doings. By ascribing intentionality to actions I understand actions to be those human doings that are caused by reason(s), where reasons, in turn, are beliefs grounded in the practical interests of life (see below). Finally, I take the notion of choice to denote a power possessed by each individual whereby, in any situation, he or she could really have acted other than he or she did. These interpretations raise further issues which warrant elaboration. Before providing it, though, let me re-emphasise a feature of human agency which has been left undeveloped in the discussion so far, namely that the capacities of choice and intentionality involve much more than the transformational capacity of being able to ‘make a difference’. The latter transformational power is shared by inanimate (e.g. hydrochloric acid) and animate (e.g. bees) agents alike. As a faculty of human beings it does not in itself presuppose that, when exercised, the agent need choose, or intend, i.e. plan in any way, or be aware of, the events that eventually ensue. With the exception of certain ‘higher animals’ most examples of agency appear to be restricted to this transformational capacity, this essentially behavioural response to the interplay between intrinsic conditions and external circumstances. In contrast, the interventions of human beings are, to the extent that they are intentional, to a greater or lesser extent under the individual’s control, allowing that events really could have turned out differently.
© 1997 Tony Lawson
INTENTIONALITY Now an obvious question that arises is whether all human behaviour or doings can be correctly described as intentional? In other words, is human behaviour to be understood as synonymous with human action? Most human behaviour appears to be intentional under some description, although not all things done in an act, and certainly not all the consequences of it, need, or typically will be, intended. Thus we can recall the frequently referred to ‘accordion effect’: I pull the light cord intending to illuminate the room but not intending, yet nevertheless simultaneously managing to break the cord and to frighten away the prowler outside the window. Clearly, although such a sequence of events might be described in various ways, e.g. ‘I broke the light cord’, ‘I frightened the prowler’, there is usually a form which expresses an act; individual behaviour can mostly be understood as an action under some description. The interpretation of intentional behaviour put forward above posits reasons as causes. But of course any thinking or cognitive activity that takes action as its result or object presupposes the causal efficacy of reasons, in the sense of their making a potential difference to the physical or mental state that eventually obtains. Unless a reason could serve as a cause there would be little point in anyone (economists, market researchers, and policy makers included) assessing different beliefs or desires held for purposes of deciding how to, or how others most likely will, or how others might most easily be influenced to, act. If a reason could not count as a cause there would be nothing to be gained from its contemplation and little point, or even meaning, to thought at all.
IRREDUCIBILITY AND THE PHENOMENON OF EMERGENCE If intentionality and choice are causal and so real, there is also a case for supposing that they are not reducible (even in principle) to some more basic (e.g. physico-chemico or neurophysiological) set of mechanisms or principles of organisation. For if certain successful (or significant) human activities (such as experimental practices bringing about scientifically significant results), are to be rendered intelligible without resorting to explanations of the ‘cosmic coincidence’ sort, it must be accepted that individual actions, and so reasons, intentions and choices, are causally efficacious with respect to physical matter. And if certain determinate material or physiological states would not have come about but for the exercise of human powers, including powers of the mind, there is at least a prima facie case for supposing the latter to be irreducible to, if dependent upon, the former. © 1997 Tony Lawson
This argument mirrors in structure the one made in the last chapter and Chapter 3 to support the assertion that social structure is similarly irreducible. That is, I have accepted here that, because intentional human action is a necessary condition for certain definite states of the physical world, determining the conditions of applicability of the physical laws to which it is subject, the powers of human agency cannot be reduced to physical principles. Human agency cannot be reformulated merely in terms of physical laws. Previously I argued in similar fashion that because social structure is a precondition of intentional human action and interaction, facilitating the exercise of human powers to which they are subject, then the properties of social structure cannot be reduced to, or explained completely in terms of, the capabilities of the human agent. Let me expand on the conception that I am supporting here, a conception influenced once more by Bhaskar’s contributions, amongst others. I am suggesting that human agents (as well as social structure) have specific powers and dispositions which serve to differentiate them from the rest of reality— from other phenomena, things, kinds of things, and also other animals and non-human species. Certainly if this were not so the social sciences would, in principle at least, be reducible to some more basic level. I am arguing, however, that nature in fact is stratified. Now if this is indeed the case it by no means follows that any two levels or strata need be totally independent of one another. Specifically, human beings do have a material component and life does appear to have ‘evolved’ from inanimate matter. In other words, we must recognise the phenomenon of emergence. To recall the discussion of Chapter 6, an entity or aspect found at some level of organisation can be said to be emergent if there is a sense in which it has arisen out of some ‘lower’ level, being conditioned by and dependent upon, but not predictable from, the properties found at the lower level. In principle it is possible that the mechanism by which the higher order entity or aspect emerged can be reconstructed and explained in terms of principles operative at the lower level. But if powers located at the higher order level are genuinely emergent, their explanatory reduction is proscribed. It can be noted that any specific analysis of a genuinely emergent realm may commit either (or both) of two forms of ontological error. Not only is it possible to misinterpret emergent phenomena as reducible (in principal, if with time and effort) to certain elements and governing principles lying at some lower level of organisation, but equally it is possible to forget that higher-order phenomena are always dependent upon, and conditioned by, principles which operate at the lower level from which they emerged. And it is just as important to avoid the latter error as it is the former. Just as social structure cannot be understood independently of considerations of human powers, the natural order in which both the social and the psychological are © 1997 Tony Lawson
embedded, and upon which they can act back, must be recognised as a condition for social action and thus as an object of social analysis. Parenthetically, there is some urgency for the latter to be fully recognised by economists. Specifically, there is an increasingly apparent need for economists to address themselves in a non-superficial way to ecological/ environmental issues, to concern themselves with identifying and understanding primarily ecological/environmental mechanisms and conditions, as well as the nature of, and requirements for, our sociological being.1 Our natural environment is not a resource that can be constantly plundered; an attention to sustainable development is crucial.
TACIT KNOWLEDGE The human subject, then, is conceptualised here as a relatively autonomous causal agent, and human action is viewed as always purposeful (and so in this sense teleological). In other words, human action is governed by reasons so that it is always directed by some beliefs, or knowledge, and towards some end(s). It warrants emphasis, however, that this pursuit of ends and drawing upon of beliefs cannot be understood as some simple, unitary, always reflected upon, discursive affair. This conclusion follows from the briefest consideration of the sort of day to day (highly routinised) activities in which people engage. Notice, first, that human beings not only initiate change in purposeful ways but also monitor and control performances (and indeed monitor the monitoring of performances; we are aware of our own state of awareness during the course of action). Not only do we initiate our journeys but we monitor our driving performances as well as the performances of other motorists (and as well our assessment of their performances). Now it is clear that the social, including economic, activity that each agent reflexively monitors is an ongoing flow, a continuous stream. It includes the individual’s own acts, the doings of others, the socially constituted appropriateness of forms of conduct in specific settings, and so forth. In consequence, it is apparent that for the successful performance of many acts, such as driving on the motorway, the reflexive monitoring of activity must also be a continuous flow, rather than a discrete series of acts. But if the monitoring of conduct is, or can be, a continuous flow it must be largely tacitly carried out, i.e. performed at a subconscious level, a level of practical consciousness. For a continuous discursive commentary, a non-stop reflection upon each act of monitoring or assessment, appears infeasible. Only rarely is the discursive identification of acts of assessment (as with most acts) attempted, and then usually in response to explicit demands for clarification or assessment. The faculty of reflexive monitoring which we all continually exercise, in short, © 1997 Tony Lawson
presupposes not only a discursive consciousness but a level of tacit consciousness. If the reflexive monitoring of conduct is largely tacit so too must be the reasoning that takes place on a continuous basis. For human activity is itself continuous. Of course, the motives or needs which stimulate or prompt action and the overall plans or projects that may be formulated in line with them, tend to have direct bearing or purchase only in relatively unusual circumstances, such as moments of significant changes in possibilities and constraints. Such motives and needs, in short, bear on the reasons for particular actions rather than the modes in which they are chronically carried out. But much day to day conduct is concerned precisely with the question of possible modes in which relevant actions can be carried out and thus will be tacitly determined. It follows that much of the body of knowledge and skills drawn upon in action will be known only tacitly. If a discursive reflection upon each doing is infeasible then much of the knowledge upon which we draw must exist at the level of tacit consciousness; it constitutes tacitly held, or tacit, knowledge. There is a difference and a potential gap between what is said and what is done. It may be worth emphasising explicitly here that the relevance of this sketch of tacit knowledge is not limited to the context of discerning social rules. To the extent that tacit or practical knowledge is recognised at all in contemporary economics, it is usually discussed in the context of rule following, as though all and only social rules can be comprehended in this fashion. Of course, many of the social rules that we draw upon in ‘going on’ in life are tacitly held. This is undeniable. Conventions, i.e. procedures sustained by general (implicit) agreement—such as turn-taking in conversation—appear usually to be of this sort. But there is no hard and fast relationship here. For example, I am quite able to negotiate my way around wide swathes of the moorland next to where I grew up even when ‘lost in thought’ (while I am often quite unable to do this—and reach my chosen destination—without conscious reflection over a terrain that is barely familiar to me). Yet it is not clear that in doing so I am (solely) relying upon social, or any other kind of, rules. Moreover, even where social rules are tacitly held many of them were once discursively acquired. People are initially taught rules of (a possibly foreign) language, how to play a musical instrument or a game, how to hold cutlery or a pen, the highway code, and so forth, even if, with time, such rules become internalised and, temporarily at least, forgotten from discursive consciousness, being only tacitly known. Rules then, even for experts, may initially be discursively known. Moreover, they are usually recoverable, and are seen to be whenever a proficient performer is asked to explain what he or she is doing. Indeed, many expert performers of rule-based skills are, of course, teachers (including parents) who are repeatedly © 1997 Tony Lawson
formulating in propositional terms knowledge that is (also) held at the level of practical consciousness.
UNCONSCIOUSNESS There is at least a third ‘level’ of human subjectivity, that of unconsciousness, which must also be acknowledged here as bearing significantly upon human activity, so constituting a further type of conditioning factor of the social world. If it is the case that the line between discursive and tacit consciousness is fluctuating or permeable (in the sense that discursively acquired knowledge can, with time and experience, become tacit knowledge, just as tacit knowledge may, on reflection, be rendered discursive) there appears to be a significant barrier between discursive consciousness and unconsciousness. Specifically, the latter includes forms of cognition and dispositions which appear either to be inaccessible to discursive consciousness because they were formed before the development of linguistic competence, or to constitute items that are repressed from discursive consciousness, being inhibited from appearing there in anything other than (possibly significantly) distorted form. Thus while individual agents can nearly always report their intentions in, and/or reasons for, acting as they do they cannot always identify their motives, many of which are unconsciously held. This Freudian expansion of the realm of subjectivity to include unconsciousness is no longer particularly contentious. It is nevertheless of interest to see both that the outline sketched above can be (transcendentally) sustained, and that further relevant insights are obtained, if we consider once more the observation emphasised earlier that human beings chronically engage in the performance of routines. For it is clear that the latter phenomenon warrants an explanation not only in terms of what social reality must be like for widely shared routines to be possible (the question addressed in the previous chapter), but also in terms of what human beings must be like for such routinised behaviour to be so widely pursued. It is certainly true that part of the explanation for widespread routinised behaviour is that it is not merely an efficient, but usually the only, way of meeting day to day objectives. That is, much of social life is so closely structured by relatively enduring rules, especially conventions, that any agent necessarily draws upon them repeatedly, and in doing so acts in a routine fashion, in order to get things done. Think of any ‘consumer’ on a shopping expedition. If desired goods are to be acquired the consumer must routinely monitor prices, select currency in appropriate amounts, engage in the transfer of ownership rights (accepting the opportunities and responsibilities that purchase and ownership entail), and so on. I emphasise that the claim here is © 1997 Tony Lawson
not that agents engage in rule following and so act routinely because this is somehow more rational on average (or whatever) than continually acting on some alternative knowledge of the circumstances of time and place. It is not the case that rule following, though less efficient in some sense on a case by case basis, is somehow a more efficient or optimal way of acting overall (Vanberg, 1994). The point is that society is constituted by rules and practices, if amongst other things. If I, living in the UK, wish to consume a banana or coconut today I have no option (that is not likely to land me in jail) other than to engage in the performance of the noted, routinely performed, forms of action. People, then, engage in various routines in some part at least because doing so constitutes the only way of meeting day to day objectives. But there is more to the routinisation of human life than this. For human agents engage chronically in routinised forms of behaviour even where the practices involved are not widely followed by others, where they are not conditioned by (widely) shared social rules. Families are regularly observed to partake in routine activities ‘of their own’ both inside and outside the home. Even individuals living alone appear to develop their own habits or other routinely performed forms of conduct. In addition it is often observed that when individuals find themselves, even temporarily, constrained from engaging in routine patterns, stress and tension set in (Giddens, 1984). Human beings, in short, appear, and seem to need, to engage in routinised forms of behaviour whether or not there is any direct motivation to do so, whether or not there is a motivation rooted in the pragmatic, historically specific, goals of daily life.
SAMENESS AND CONTINUITY How is this apparently generalised feature of experience to be explained? To account for this situation it is necessary to acknowledge a role for something like unconscious motives which indirectly dispose human agents to engage in institutionalised or routinised forms of conduct. The explanation of the noted phenomena must be a human need for, or one that is satisfied by, a significant degree of continuity, stability and sameness in daily affairs, the avoidance of radical disruption. It would seem that at the level of the unconscious is a basic need for inner security grounding a generalised disposition towards the maintenance of trust in the natural world and the avoidance of anxiety, a disposition that is in practice fulfilled through, if amongst other things, the doing of familiar things routinely. In recent years Giddens (1984) has focused more closely than most on elaborating the nature and consequences of the system of inner security or, as he (following R.D.Laing) refers to it, the system of ‘ontological security’. © 1997 Tony Lawson
Drawing selectively upon the insights of Erikson (1963), Goffman (1963), Kardiner (1939) and Sullivan (1955), Giddens suggests that the: psychological origins of ontological security are to be found in basic anxiety-controlling mechanisms (as indicated by Erikson...), hierarchically ordered as components of personality. The generation of feelings of trust in others, as the deepest lying element of the basic security system, depends substantially upon predictable and caring routines established by parental figures. The infant is very early on both a giver as well as a receiver of trust. As he or she becomes more autonomous, however, the child learns the importance of what are in Goffman’s term ‘protective devices’, which sustain the mutuality implied in trust via tact and other formulae that preserve the face of others. Ontological security is protected by such devices but maintained in a more fundamental way by the very predictability of routine, something that is radically disrupted in critical situations. The swamping of habitual modes of activity by anxiety which cannot be adequately contained by the basic security system is specifically a feature of critical situations. (Giddens, 1984:50–1) According to this interpretation, because an orientation to the avoidance of anxiety occurs early on between the infant through his or her interaction with the closest guardian, typically the mother, and because it occurs before any significant acquisition of language, the mechanisms developed to counteract the development of anxiety are, in part at least, unconscious ones.2 As the infant grows into an adult, the controlling of anxiety remains a basic need or disposition. Giddens (1984:54) even suggests that this is perhaps ‘the most generalised motivational origin of human conduct’ (ibid.: 54). But whereas in infancy the experience of trust, stability, sameness, and continuity is achieved through the parental maintenance of predictable caring routines, in adult life it is obtained in the routine modes of conduct that facilitate ‘going on’ generally. The performance of routines, in other words, is not only essential to the reproduction of social structure but is equally fundamental to the production and reproduction of each individual personality. Supportive evidence for such a conception of a system of inner security, one grounding various needs and motivations originating at the level of the unconsciousness, is provided through examining human responses to situations in which continuity, sameness and trust are undermined, through observing human responses to situations wherein habitual modes of activity are swamped by anxiety which cannot be adequately contained by the basic security system. Giddens, for example, draws upon Bettleheim’s first-hand account of the profound effects of enforced de-routinisation on the experiences © 1997 Tony Lawson
of inmates in a Nazi concentration camp. As individuals lose any certitude from the reproduction of rules and predictable routines they lose any sense of autonomy in action. At the limit they lose their most basic sense of control over their own physical doings. Giddens summarises Bettle-heim’s observations as follows: The disruption and the deliberately sustained attack upon the ordinary routines of life produce a high degree of anxiety, a ‘stripping away’ of the socialised responses associated with the security of the management of the body and a predictable framework of social life. Such an upsurge of anxiety is expressed in regressive modes of behaviour, attacking the foundation of the basic security system grounded in trust manifested towards others. Those who are ill-equipped to face these pressures succumb and go under. Some are able to sustain a minimal sphere of control and self-esteem that allows them to survive for a longer period. But eventually, in most of the old prisoners at least, a process of ‘resocialisation’ takes place in which an attitude of trust (limited and highly ambivalent), involving identification with authority figures, is re-established. (Giddens, 1984:63–4)
STABILITY IN THE FACE OF UNCERTAINTY It seems clear that such considerations are significant in understanding economic activity, especially given the latter’s highly routinized nature. Nor are the conclusions drawn wholly novel in economics, including the assessment that sameness and continuity serve a deep seated psychological need or motive. Consider, for example, Keynes’ General Theory. In his Chapter 12, Keynes focuses in particular upon the irreducible uncertainty regarding future yields that faces financial investors.3 He notes that in such conditions, despite the fact of widespread speculative activity in highly liquid security markets, some appreciable stability is nevertheless experienced. Keynes attributes this stability to the ‘conventional’ practice adopted by many investors of basing their actions on a scenario obtained by projecting the existing situation (i.e. the assessment of investment prospects as represented by current market evaluations—see Runde, 1997) into the future and modifying it only to the extent that there exist more or less definite reasons for expecting a change, a practice which apparently reasserts itself even following periods of significant instability.4 But what explains the widespread reliance on this particular ‘convention’? The main causal factor suggested by Keynes is precisely the psychological need of individuals for continuity and sameness in all they are © 1997 Tony Lawson
doing. Keynes, indeed, takes the view that the practice in question is so deep rooted in our behaviour that ‘it continues to influence our minds even in those cases where we do have good reasons to expect a definite change’ (1973b: 125); that ‘the idea of the future being different from the present is so repugnant to our conventional modes of thought and behaviour that we, most of us, offer a great resistance to acting on it in practice’ (ibid.).5 Furthermore, while a maintenance of the convention allows that investors ‘need not lose...sleep’ (1973a:153), anxiety is never completely lulled or fully controlled. The existing conventional evaluation may at some moment give way. Because the structure of market evaluations at any point in time cannot be based on a significant knowledge of eventual yields (for such knowledge does not exist) and because at any time there may be changes in relevant news, the situation, even if stable, may simultaneously be highly precarious— a perceived potential for change due to the possible arrival of ‘news’, etc. will usually be present. If the ‘convention’ does give way, according to Keynes, the practice of calmness and immobility, of certainty and security, suddenly breaks down. New fears and hopes will, without warning, take charge of human conduct. The forces of disillusion may suddenly impose a new conventional basis of valuation.... At all times the vague panic fears and equally vague and unreasoned hopes are not really lulled, and lie but a little way below the surface. (Keynes, 1973b:114–15) It is such reasoning that enables Keynes to explain how it is that panic buying and selling is an often experienced phenomenon and why investors sometimes choose to hold (non-interest bearing) money even as a store of wealth: Because, partly on reasonable and partly on instinctive grounds, our desire to hold money as a store of wealth is a barometer of the degree of our distrust of our own calculations and conventions concerning the future. Even though this feeling about money is itself conventional or instinctive, it operates, so to speak, at a deeper level of our motivation. It takes charge at the moments when the higher, more precarious conventions have weakened. The possession of actual money lulls our disquietude; and the premium which we require to make us part with money is the measure of the degree of our disquietude. (1973b:116) To recap once more, the intentionality of human action presupposes not only discursive consciousness but also a tacit or practical level of consciousness as well as unconsciousness. Of the three it appears to be tacit consciousness that © 1997 Tony Lawson
is the most significant for understanding day to day conduct. For it is only tacit consciousness that is implicated continuously in daily social, including economic, life, which renders possible human action as a continuous process. The discursive identification of acts, in contrast, will usually only be attempted in response to explicit demands for clarification. Unconscious needs and motives, though also significant for understanding conduct, will usually bear on the reasons for kinds of action rather than the actual modes in which they are carried out.
INDIVIDUALITY Before drawing together the various threads of the discussion of these last two chapters it is opportune to focus explicitly upon one further aspect of human subjectivity, namely individuality. The sort of conception which can be supported here must already be apparent. But in these times of ‘post-modernist’ extravagances, where the emphasis is very much upon individual peculiarities and an almost a priori (if para-doxically uniform) opposition to uniformities at any level, any failure here to sketch explicitly a conception of individuality congruent with the above outline would no doubt be construed by some as an inability to sustain any such conception at all. It is the case that the account of social structure and human agency which has been developed in the previous two chapters has been formulated, by and large, without regard to the manifestations of particular social structure and human agents in any specific context. The social theory that has been sketched, in other words, is a series of insights into relatively trans-historical and transgeographical features of the sort of phenomena that constitute the subject matter of the social sciences. But if this is so it is not here being suggested that the items so far elaborated are ever expressed or manifest in anything other than more or less historically-specific and highly differentiated forms. Just as forms of social structure (such as, say, social rules) always take a specific form (as, say, a specific rule to drive on the left side of the road in the UK) so human nature and subjectivity are only ever actualised, or made manifest, in historically, geographically and culturally specific conditions and terms. In fact, the more contentious issue is whether there is indeed a common human nature, a nature common to all human beings. It is clear that, given the discussion so far, any conception of a common human nature that is supportable could not be ahistorical. But equally the discussion of the generalised need for the reproduction of conditions of sameness, stability, continuity and trust, suggests that the claim of a relatively enduring common human nature is tenable at some level at least. Indeed, it seems rather implausible to suppose that human beings do not possess various shared © 1997 Tony Lawson
characteristics (e.g. language capabilities) which both derive from a scientifically recognised common genetic structure and serve to differentiate us from other species. Human nature, then, when viewed from one aspect, or at a high level of abstraction, or so I am suggesting, can be accepted as a common human attribute, one grounded in our genetic constitution and manifest in certain species-wide needs (e.g. the control of anxiety) and capacities or powers (e.g. language use). But even a common human nature can only ever be expressed in inherently socialised, more or less historically, geographically, and culturally specific, and very highly differentiated, forms. In other words, when human nature is viewed under a different aspect to the above, and specifically at a lower level of abstraction, it can be understood as an historically relatively specific nature, the development of which has its origin at the time, place and conditions of an individual’s birth, and which is subsequently influenced by the class, gender, occupational positions, and so forth, in which the individual stands, along with his or her experiences generally (for example we cannot just speak in the abstract, we have to speak a specific geo-historically located language, etc). To the extent that numerous people throughout their lives are subject to identical or similar forms of determination, an historically quite definite nature may thus be held in common. In the limit, though, any individual will always be subject to a unique combination of experiences and modes of determination producing a particular personality, so that, from a third, rather more specific, perspective, or a yet lower level of abstraction, the nature of any given human being must be seen as a more or less unique individuality. In short, a person’s individuality is primarily constituted by his or her social peculiarity. Each individual is the product of his or her actions and experiences within the social relations and other modes of determination into which he or she is born and thereafter lives. An individual’s actions, or things which happen to him or her, are comprehensible in terms of the individual’s socially conditioned capacities, powers, liabilities and dispositions. The agency of each individual is thus conditioned by the relationships in which he or she stands just as these relations, as with social structure in general, are in turn dependent upon the sum total of human doings.
ECONOMIC RATIONALITY By way of linking together the various threads of this and the previous chapter it is useful to contrast the conception of human subjectivity proposed with the theory of the rational individual that dominates contemporary economic orthodoxy. As we saw in Parts I and II, the individual agent of contemporary © 1997 Tony Lawson
mainstream economics usually inhabits a world composed of such surface phenomena as (measurable) events and states of affairs, phenomena which, by and large, each agent is supposed unproblematically to perceive. The deductivist aim is then to specify each economic model in such a way as to guarantee that under given determinate conditions x a specific outcome y is guaranteed to follow. This goal is achieved, typically, by imputing to any ‘economic agent’ some unitary objective, a set of beliefs/knowledge of the measurable events and states of affairs which comprise the agent’s environment, and an ordering of some kind over the perceived potential satisfiers of the imputed objective, one that facilitates an ‘optimising decision’. Often the knowledge-set in question is fantastic; it is almost always, in the standard account, sufficient to facilitate an unambiguous optimising (maximising or minimising) response or decision on the part of the agent. In other words, this assumption of calculative optimising behaviour, i.e. of economic rationality, which is sustained throughout much of contemporary mainstream economics, is merely a gloss on the proceedings. It is imposed a priori to guarantee the dogma that under any given conditions x the same deducible event or outcome y always follows. It is apparent that the supposed universality of this orthodox axiom of rationality sits uneasily with the conception I am defending, immediately we take account of the capacity of human choice. I am not suggesting, however, that there is necessarily a logical incompatibility between the orthodox conception of economic rationality and human choice. For the orthodox rationality axiom really amounts ‘only’ to the stipulation that (1) if x and y are ‘feasible options’ of some kind, and (2) if x is in some well defined sense preferred to y, then (3) y will not be pursued. Clearly this axiom can, superficially at least, be rendered compatible with the intuition that agents have real choice. This is achieved, in other words, if—albeit only if—there is (always) within the set of feasible options a subset of possibilities x1, x2,... xn, say, which the agent is indifferent amongst or unable to rank, but where each possibility is considered preferable to any ‘feasible’ option not included within this subset. The problem with this latter scenario for mainstream economics is that the goal of the deductive prediction of specific events is left unattainable after all, while such a goal provides the only reason for insisting on the universal positing of the axiom of an ever calculating, only instrumentally rational, agent in the first place. Clearly, if the rationality axiom no longer serves its intended purpose the question arises as to why it should be imposed as an a priori premise. All barriers disappear to the less dogmatic view that instrumental rationality is at best a hypothesis to be empirically assessed in specific contexts.
© 1997 Tony Lawson
SITUATED RATIONALITY That having been said, if a recognition of the reality of human choice is sufficient to undermine the usual rationale for accepting calculative ‘economic rationality’ as a universal doctrine, it by no means follows that the account which I am proposing does not itself involve a conception of human rationality. Indeed, the concepts of choice and intentionality entail that individuals are able to formulate plans in line with beliefs and desires, and to act upon them. But while the image of the perpetually-calculating, optimising agent is one that few find realistic, the observation that individuals continually and skilfully negotiate their daily affairs is nevertheless a commonplace. In part, indeed, the theory elaborated here has been an attempt to render the latter observation intelligible. The conception of human rationality that emerges, though, and here of course rationality must be construed as a capacity rather than an actuality, is very much one of a situated sort. Indeed, the vision provided might be interpreted as a theory of situated rationality. Not only are individuals’ choices of actions conditioned by the situated options which they perceive, but also the individuals themselves, their expressions of their needs and motives, the manner in which their capacities and capabilities have been moulded, their values and interests and so forth, are conditioned by the context of their birth and development.6 At any given point in time any individual is situated in a range of positions, with associated, perhaps contradictory, real interests, as well as other needs and motives. Associated with the positions in which any individual is located will be a range of rules to draw upon, obligations to fulfil, structures of power to utilise and be influenced by. Many such social structures will be inadequately or falsely understood. Most of the skills utilised, modes of conduct performed, will be tacit. Action in such a context is a continuous stream, continuously monitored, and rarely rendered available to discourse. And yet it is undeniable that agents capably negotiate their daily affairs. In short, while there can be no doubting the skilful, capable, knowledgeable, and purposeful—in short rational—manner in which people usually act, the concept of rationality implicated in all this is a far cry from the ever calculating, ranking, optimiser of standard economic theory. Perhaps Oakeshott’s (1962:109) conception of rationality as ‘the certificate we give to any conduct which can maintain a place in the flow of sympathy, the coherence of activity, which composes a way of living’ captures this conception well. However that may be, the picture which emerges from the discussion of the last two chapters is one of a process in motion, of dynamic, structure and agency inter-dependency. Human practices or doings are the key to it all. © 1997 Tony Lawson
Human beings are the moving forces in history and it is upon human actions or doings that everything in the social world turns. Not only is social structure reproduced and transformed through human practice but so too is the personality of each individual. Social structure is the, typically unacknowledged, condition of action as well as its, usually unintended, consequence. Although individuals do not normally act with the goal of reproducing or transforming structure in the way that, in total, they nevertheless do, each agent normally acts capably and purposefully within the structures and modes of determination in which he or she is situated. And as individuals perform in routine fashion the tasks necessary for going on in life, the conditions in which they act influence the development of each human individuality every bit as much as these conditions are themselves perpetually reproduced, reformed and transformed through human doings (and nondoings) in total. This is the sort of perspective or vision on the nature of social reality that is most intelligibly sustained on the basis of (transcendental reasoning from) the generalised features of experience accepted throughout. And with this perspective achieved we can now assess its implications for the investigatory research of social phenomena. We can now turn to the specific task of drawing out implications for the conduct of economics.
© 1997 Tony Lawson
Part IV
ECONOMIC EXPLANATION
© 1997 Tony Lawson
14 BROAD OBJECTIVES AND POSSIBLE OBSTACLES
I turn, then, to examine how the understandings and conclusions reached so far bear upon the doing of economics. Now I must straightaway emphasise that it is impossible to provide here a package of detailed, specific rules which have only to be followed. Critical realism does not come like this. At the same time, of course, I am obviously not advocating an ‘anything goes’ project. Much of the discussion has been devoted to revealing the errors and limitations of applying various definite methods, aims and criteria in the context of social phenomena. But a fundamental component of my argument (and indeed a central plank in the critique of deductivism in economics) is the recognition that the subject and context of any analysis will bear on the analytical methods and criteria that are appropriate. Given the (unavoidably) high level of abstraction of the perspective systematised in the previous two chapters, with the focus being on relatively generalised features of social phenomena, any analytical implications that can be drawn will inevitably be correspondingly general in nature: basically taking the form of broad criteria and strategy. The insights gained from these last chapters do indicate a need for a basic re-orientation of the discipline, primarily a transformation in accepted objectives. And we will see that they do also indicate a good deal about the direction in which the discipline ought to move. But they cannot be used to contrive a list of detailed instructions which practising scientists have only to act upon. Before outlining such directional and other implications as can be drawn, let me at this point provide a brief overview of this fourth part of the book, to indicate something of the ground to be covered. The current chapter is concerned first of all with sketching the sorts of objectives for economics that the preceding account of social ontology can sustain. At this point I also examine whether peculiarities of social material other than its apparent insusceptibility to controlled experimentation, manifest inherent obstacles to social scientific successes. The question of whether the limited opportunity for experimental control in the social realm constitutes an unsurmountable barrier © 1997 Tony Lawson
to achieving a successful economic science is taken up in Chapter 15. In concluding that neither social material’s insusceptibility to experimental control nor its other peculiarities provide inherent obstacles to social scientific successes I also use Chapter 15 to lay out various broad strategies, modes of reasoning and very general criteria which economics can, with both justification and, I think, some expectation of success, accept. The procedure of abstraction is examined as an essential aid to achieving the scientific possibilities for economics in Chapter 16. In Chapter 17 attention is given to the epistemic status of the sorts of results, or knowledge, that economic science can hope to achieve. And in Chapter 18 an illustration is provided of certain key features of the sketch of the explanatory process developed in these next few chapters.
BROAD AIMS It has been established that the social world is structured. It follows that economic explanations are concerned in a significant way with identifying social structures and conditions which govern, facilitate, or in some way produce, actual social events and states of affairs of interest. While it is easy enough to conceive of specific natural events (such as falling leaves or movements in iron filings), and to acknowledge natural generative structures and mechanisms (such as gravity and magnetic fields), much of the previous two chapters has been concerned with the perhaps more contentious issue of identifying the nature of the social analogues. I begin with social events.
THE CENTRALITY OF HUMAN PRACTICE According to the social theory developed in the preceding chapters the social world depends upon human activities and practices. Human beings are the motivating forces in history, and it is through human doings that society and economy are reproduced. It follows that any social phenomena that we may wish to explain will either be the result of, or, perhaps, just consist in, some forms of human activity. If the latter is the case, that is, if our direct concern is with forms of human practices in themselves, the explanatory exercise straightforwardly turns upon identifying their facilitating conditions. If the former is the case, that is, if we are first of all interested in some consequences of human action, then part of the explanation will involve locating those practices most responsible for the phenomena in question. To take an example, if over a lengthy period of time productivity growth in a specific country is frequently lower than in most of © 1997 Tony Lawson
its ‘competitors’ (which has indeed been the experience of the UK over much of the last hundred years compared with other European countries), an obvious question to address at the outset is which, or whose, practices or activities are implicated. The practices in question could be those bearing upon the amount of investment that is undertaken, its type, the rate of diffusion of new technology, the manner of deployment of technology, including staffing-rates or machine running speeds, and so on. Whether it is of direct or indirect interest, in short, a fundamental step of any economic investigation will be to identify some set of human activities regarded as being of economic significance. Once a relevant set of practices has been singled out as central to our concerns it will be necessary to determine how and/or why they are undertaken, especially if any corresponding practices adopted elsewhere, or previously, are different in some significant way. In this the specific explanatory tasks will include uncovering the physical and social structural conditions of the practices in question, tacit skills drawn upon and possibly unconscious as well as conscious motivations. Many of these conditions may be unacknowledged by the individuals drawing upon them, and/or perhaps falsely or inadequately understood. For it is a conclusion of the previous two chapters that while society is the skilled accomplishment of active individuals, it remains to a degree opaque to the individuals upon whose activities it depends. In short, the broad aim of the social sciences, including economics, is to describe the structural conditions for some manifest social phenomenon (including some relevant set of practices or activities) to be possible, recognising that the situation in question may not be adequately comprehended by the individuals involved. Generally speaking, then, the explananda of the social realm, the phenomena to be explained, are the practices in which people engage, and the explanans are the physical, social and psychological conditions of the relevant actions. The qualification in the preceding sentence is required because we have also noted that the starting point of analysis may actually be manifest states of affairs, such as high rates of unemployment or inflation, rather than specific practices per se. The point here is that any such aspect of the social world is not only a possible condition but also a consequence of social activity, whether intended or (as is more often the case) not. It may thus be the consequences of actions which constitute the phenomena that, initially at least, are warranting of explanation. In many cases, though, the practices most responsible for economic states of interest will not be difficult to determine, and to the extent that, or where, this is so, the question of greater issue, the crucial phenomenon to explain, is precisely these human practices in question. It is worth observing that this last scenario is precisely what is denied by ‘methodological individualism’. According to the latter (or at least of any © 1997 Tony Lawson
formulation of it that warrants such a reductionist head), the goal is always to explain or describe social states or outcomes in terms only of individual agency (i.e. it is never the other way round). In some Austrian accounts, as we have seen in discussing Hayek’s 1940s’ position, this leaves social science with relatively little to do over and above the grasping of agents’ conceptions. Orthodox ‘theorists’ are also methodological individualists, of course. Here the objective has usually been to (attempt to) demonstrate that all eventualities can be rendered consistent with some typically quite irrelevant a priori, conception of calculative (optimising) behaviour, or some such. In any case, methodological individualism has been found above to be misguided. Once it is recognised that social structure is a real emergent property of human agency, which, as it were, acts back on the latter in making a difference to what people are able to do, it follows that a necessary, and possibly more demanding and interesting, aspect of social science is the uncovering of particular social structures of significance. In short, if some phenomenon singled out for the explanation is not itself a form of human activity then an essential step of the analysis will be to identify those practices (most) implicated in it. An ensuing task (or the immediate one if we are interested in certain human practices per se) is to explain those practices, to identify the conditions, including relations, positions, rules, skills and motivations, by virtue of which the activities in question are possible and indeed occur. Once an explanation has been achieved, it in turn becomes the object to be explained. This is no less true in natural science than it is in the science of society. However, in virtue of that peculiarity of social structure that it is dependent upon human agency, social forms cannot be expected to endure in the sense (or to the degree) that many natural structures and mechanisms do. In consequence, social explanation will usually need to include an explicit account of the agent-dependent manner of reproduction and/or transformation including the possible demise (and perhaps also of the genesis) of any set of social structures identified as explanatorily significant. Explanations of social phenomena, perhaps more so than those of natural phenomena, may prove to be insufficiently illuminating where explanatory causal structures themselves remain unexplained. Economic explanation is inevitably concerned with process.
INTRINSIC LIMITS TO SOCIAL SCIENCE If the above outline broadly captures the conception of explanatory goals of economics supported by the preceding analysis it is clearly also of interest to question whether anything can similarly be determined about how explanatory projects might, or perhaps are unable to, proceed. Now, in some, © 1997 Tony Lawson
but certainly not all, natural science contexts, controlled experimentation is feasible. In such cases, it may be possible to insulate stable mechanisms from other potentially interfering ones and to check their capacities for producing effects as postulated. The controlled experiment, however, is not a real possibility in social research. And this situation may seem to some to beg the question as to whether a successful social science is practically feasible at all. If universal constant conjunctions of events rarely occur outside the experimental situation, and if experimental control is infeasible in the social realm, there appear to be grounds for doubting whether the search for economic mechanisms can even be initiated, let alone prove successful. This is an issue that requires a good deal of detailed consideration. Before addressing it at length (the object of Chapter 15 below) however, I first want to dispel any idea that the possibility of a successful science of economics is already inevitably undermined by a second set of potential difficulties reflecting peculiarities of social material other than the latter’s insusceptibility to local closure. Specifically, it may appear that, whatever the possibilities for experimental control, insuperable complications arise for any would-be social science of economics due in particular to (1) the non-perceivable character of much of society and economy; (2) the internal relationality of the objects of social science; and (3) the concept-and practice-dependent nature of social structure underpinning its dynamic nature, and the fact that economics, through its investigations, can lead to a transformation in the objects it wishes to study. Ultimately, I do not think that any of these noted features compromise the possibility of successful social science, although some may indicate constraints on ways in which it can be done. Let me, though, briefly consider these potential difficulties and indicate why I take this view. First, social structure is necessarily non-perceivable; it cannot be identified directly, but only through its effects. It can only be known, and not seen, to exist. But in this respect it is no different from many objects of natural scientific enquiry, including gravitational forces (or curved space), magnetic fields, black holes, and so on. And as in natural science the validity of hypotheses of non-empirical structures and mechanisms can in principle be fully assessed via the empirical adequacy of any further hypotheses about observable phenomena that can be deduced or otherwise inferred from them. There are no special problems arising as a result of the non-perceivability of social mechanisms over and above any which are attributable to the infeasibility of experimental control. Second, although the subject matter of economics is structured by internal (as well as external) relations, there is again nothing which follows from this that necessarily undermines the possibility of successful social scientific explanation. Given any phenomenon to be explained, it is necessary to determine a hypothesis of a mechanism, or a set of mechanisms, capable of generating the phenomenon in question. If the conception which proves most © 1997 Tony Lawson
successful in this happens to refer to a complex totality (and I shall turn to the criteria for deciding which of competing conceptions are most successful in the following chapter), this does not raise any difficulties additional to the problem that the social sciences are unable to carry out controlled experiments. Notice, moreover, that although totalities are real, it is an epistemically contingent question as to whether we require a phenomenon to be understood as an aspect of a totality or not. For someone deciding whether or not to swim off the Kuwait coast in 1991 it was presumably enough to know that oil is being pumped into the sea. For someone wishing to understand why this was happening a good deal of human economic, political, geographical, social and religious history is required, revealing the social totalities, if amongst other things, at work. Third, there arises the question of whether anything follows from the insight that social structure is concept- and practice-dependent. In one respect this situation must actually aid scientific enquiry. For it follows from this condition that many of the objects with which social science is concerned will already have been identified under some description by the people whose activities are being explained. Of course, the understandings which individuals have of their situations are unlikely to be entirely adequate. Practices undertaken may be too partially, or even falsely, conceptualised by the participants, just as conditions drawn upon in action may exist independently of their appropriate conceptualisation (despite the failure of the hermeneutic tradition to contemplate such a possibility). But this does not mean that agents’ conceptions of their practices and their conditions (which, however inadequate, are as real as any other causally efficacious factor) are other than an essential entry point for social scientific research. It merely follows that one vital task of social science will be to produce appropriate transformations of the existing nominal conceptions of what is going on, before, or rather in the process of, retroducing (perhaps unacknowledged) structures and mechanisms, tacit skills and motivations, by virtue of which the practices in question are possible and take the form that they do. Of course, once such revised accounts of practices and hypotheses of causal mechanisms have been produced, they need to be assessed and legitimatised empirically (according to criteria to be discussed below). The question here, though, is whether any inherent limitations on the possibility of social science follow from the condition that society is intrinsically dynamic and even susceptible to transformation just by virtue of being an object of social scientific enquiry. In fact, a recognition of this condition mainly serves to remind us that social structures, like the human practices on which they depend, are inevitably geo-historically restricted. Indeed, they will frequently be internally related to highly localised conditions. It follows that an economic science will necessarily be historical © 1997 Tony Lawson
and geographical in character. But it is not obvious that any fundamental difficulties arise in consequence. Of course, at some level, space-time specificity is as much a feature of natural science as it is of social investigation. Water, say, just as much as any definite economic system, requires particular conditions for its existence, conditions which will be found only in certain space-time contexts. Not so long ago lasers did not exist on earth, and nor perhaps anywhere else in the universe. Currently, due to the intentional production of appropriate conditions, there are tens of thousands all around us. Even so, the contextual specificity of much social structure is likely to be that much greater, with all social structures being found only in restricted (and sometimes very highly restricted) regions of the globe and (human) history. But if this necessitates that economics is always context-sensitive, there are no obviously insuperable problems or obstacles here for any explanatory project of the sort being considered. It does follow, though, that economic enquiry can lead to a transformation in any social object that becomes the focus of study. Lay agents may in principle appropriate useful insights of social science and incorporate them in their activities, a response which may well constitute, or result in, a transformation of the object of study. The potentially troublesome question here is whether social objects that are open to transformation as a result of social enquiry can be regarded as truly intransitive objects of that enquiry. But of course they can. For, at the point that the social investigation is initiated the putative objects of that enquiry either do, or do not, exist, and, if the former, possess whatever properties they do quite independently of the process of investigation which eventually ensues. This is true, even if that investigation, in due course, informs a transformation in the object concerned. Indeed, the very objective of this current study is to help facilitate a rapid transformation in orthodox economics. Yet there can be no denying that, whatever its longevity, orthodox economics is a project that is currently real and highly efficacious. If strands of that project are only remotely concerned with achieving an empirically adequate expression of reality they are nevertheless currently an aspect of this reality (thus qualifying the way the term ‘and’ should be read in one possible interpretation of the title of the current book). In short, the social sciences are concerned to study intransitive objects just like the natural ones. Being, or existence, means the same in both contexts even if the forms of being are different, if the categorical features of the objects of two domains are to an extent distinct. This situation warrants emphasis. For, as increasing numbers of economists appear to be looking to abandon positivism in its most naive forms, it seems to be all too easy to overact and slide out to some alternative, hermeneutic, or post-modernist, or subjectivist, pole (or version) which erroneously attempts a dissolution of © 1997 Tony Lawson
intransitivity altogether. And one consequence of this, amongst many, as we have already seen and will explore more fully in Chapter 19, is the abandonment of the possibility of scientific critique. Yet it is upon the possibility of such critique that any project of intentional, knowledgable change, including emancipatory transformations (which many anti-positivists profess to aspire to), must be grounded. In sum, the noted peculiarities or characteristic features of social life do not appear to present any obvious obstacles to social scientific enquiry over and above any that may be engendered by the infeasibility of experimental control. They do, in part, explain why experimentation is not a viable social scientific strategy. But they do not in themselves appear to constitute or generate obvious further difficulties.
© 1997 Tony Lawson
15 ECONOMIC SCIENCE WITHOUT EXPERIMENTATION
The question I now must confront is how social scientific research can proceed in the absence of real possibilities of experimental control. How can the scientific enterprise even get under way without event regularities to account for? And how are competing theories to be assessed when the social sciences are generally denied the crucial test situation? Despite the lack of opportunities for controlled experimentation in the social sciences I remain optimistic about the social scientific prospects. In setting a context for developing my position I refer first, albeit extremely briefly, to relevant assessments of Bhaskar (1979) and Collier (1994). Each has questioned whether the potential for social scientific successes exists given the limited opportunities for meaningful experimentation in the social realm. And each has done so supporting a framework similar to that defended here. While Bhaskar sustains a degree of optimism Collier is rather pessimistic. A feature of social life that is fundamental to these opposed assessments is something which Bhaskar construes as a ‘compensator’ for the lack of experimental control in the social sciences. To the extent that Bhaskar is optimistic about the prospects for the social sciences this appears to turn on the existence of the suggested compensator. And it is through rejecting Bhaskar’s arguments on this point that Collier concludes that any optimism concerning the possibilities of a successful social science are misplaced. I too find problems with Bhaskar’s arguments, yet I do not share Collier’s degree of scepticism about the consequences. I now run briefly though the various issues involved, and indicate the basis of my optimism. My reason for proceeding in this contrastive fashion is that I think it helps both to convey the central significance of the question posed at the beginning of this chapter and also to bring certain fundamental features of the argument into greater relief. These latter aspects include some upon which rather a lot ultimately turns but which appear not to be emphasised sufficiently by either Bhaskar or Collier.
© 1997 Tony Lawson
THE HERMENEUTIC MOMENT IN SCIENCE Bhaskar’s compensator for the impossibility of experimentation in the social sciences turns on the insight that ‘social structures, unlike natural structures, do not exist independently of the agents’ conceptions of what they are doing in their activity’ (1979:38). This observation is certainly correct. And it does bear the already noted consequences (further developed in Chapter 19) that the results of social scientific research have the potential to influence the very objects of such enquiry. This is a potential that does not usually figure in the natural scientific context. Now a recognition of this capacity for social science thus to engender changes in the social world, including the practices of science itself, may well affect the sort of research that is encouraged, financed and accorded status. But, as we have seen, it is not the case that this undermines the possibility of social science. Social objects exist intransitively at the time any social scientific analysis of them is initiated, whatever the eventual effect upon them induced by such an enquiry. The question here, though, is whether the insight in question actually assists the cause of social science, and specifically whether a compensator for experimental activity is entailed. Bhaskar’s argument is that because ‘most of the phenomena which the social scientist has to deal with will already be identified, thanks to the concept-dependent nature of social activities under some descriptions’ (Bhaskar, 1979:49), social science is provided with a ready-made entry point for analysis. It is this that in some way compensates social science for its lack of experimental possibilities. Now it is certainly the case that this hermeneutic moment is fundamental for any social science. This has been acknowledged above. Indeed, without the agents’ own conceptions of what they are doing, social analysis could not easily get off the ground. Moreover, it is clearly the case that the concept-dependent nature of social structures serves to distinguish them from natural ones. But does the associated hermeneutic moment really constitute a distinct feature of the social sciences relative to the natural ones, a compensator for the impossibility of controlled experimentation? The point here is that while the concept-dependent nature of social activities under some description must be acknowledged, the concepts which people hold and act upon relate as much to natural structures as to social ones. We know a lot about both the social and the natural worlds by being agents within them. Activities such as cooking, farming, fishing and manufacture (of everything from baskets to boats, and shoes to shelter) have long depended upon the knowledge of the natural world that humans possess. Collier (1994) is correct to conclude that conceptualised practices such as these are informative about the natural world in exactly the same way that the ‘practices of commerce and state craft, conflict and cooperation, “knowing oneself” and “knowing others” are about © 1997 Tony Lawson
the human world’ (1994:247). In each case, the sources of information, the conceptions of lay agents, provide an entry point for scientific analysis, just as, in each case, the conceptions held are fallible and corrigible. If there is a difference between natural and social science with respect to the hermeneutic moment, it appears to be that the natural sciences have gone significantly beyond what the conceptions of lay agents can give (ibid.). But, however that may be (see below), it does not seem to follow that, in recognising the conceptdependent nature of social structures, a social science compensator for the lack of experimental possibilities is entailed. For Collier, the recognition that the hermeneutic moment occurs in the natural sciences in the same way as it does in the social sciences, and that it is as reliable and informative in the former as it is in the latter, encourages a pessimism with regard to social scientific prospects. For it follows, Collier suggests, that the prominence of the hermeneutic moment in the social sciences must be interpreted as a sign that the latter really have almost nothing more to go on. We have seen in Chapter 13 that the subjectivist tradition in economics (incorrectly) supposes that the grasping of agents’ conceptions is more or less all there is for social science to be concerned with. Although Collier rejects the subjectivist analysis (and acknowledges, of course, that the social world is structured and agents’ conceptions of social structures are fallible) he concludes that the impossibility of controlled experimentation in the social realm limits the possibilities for getting much beyond the hermeneutic moment in practice. If agent conceptions are erroneous, the inability to experiment places a constraint on uncovering significantly more adequate conceptions though science: [The human sciences’] input from agents’ conceptions is no more authoritative than such input is in regard to the natural world; but our capacity to correct, revise and add to the knowledge derived from agents’ conceptions is immeasurably more advanced in those sciences where experiments are possible. The teachability even of an experimental natural science doubtless presupposes our initial familiarity with (for example) heat, light and sound, push and pull, speed and weight; but before we have gone very far, we have redefined such concepts and left our homely understanding of them far behind. The hermeneutic moment is so prominent in the social sciences not because it is a more essential stage or a more reliable or informative source than in the natural sciences, but because, in the absence of experiments, we have so little else. As a result, we are also much more likely to get things wrong and much less likely to correct them in the human than in the experimental natural sciences... My conclusion from what we know about the ontology of the human © 1997 Tony Lawson
world is that it gives grounds for scepticism about the prospects of the human sciences. (Collier, 1994:248) The reason I do not share Collier’s scepticism stems from a different assessment of what follows from the inability of the social sciences to achieve meaningful experimental control. This certainly is the crux of the issue. In fact, it seems to me that the question of how social science can manage without the possibility of experimental control is the central question of methodology for those social scientists alive to the general irrelevance to social explanation of the deductivist model. The remainder of this chapter is mainly concerned with laying out in detail my position on this matter. Only at the end of the chapter do I reconsider Collier’s position as stated above.
EXPERIMENTAL AND NON-EXPERIMENTAL CONDITIONS CONTRASTED Before giving the details of my position, however, it is useful first to recall briefly certain features of the experimental situation which the discussion of Part I has indicated to be significant. We have seen, first, that experimentation is fundamentally about intervening in the world. It is about manipulating aspects of reality in order that certain causal mechanisms can be (more easily) identified and/or theories about them tested. It is a process of empirically identifying and/or assessing non-empirical structures and mechanisms through human contrivance. Second, experiments give us access to relatively enduring structures and powers and transfactually active mechanisms and their tendencies. For the intelligibility of experimental activity and results indicates that (at least some) mechanisms act though the flux of conditions that determine whether they are active and co-determine any actual outcome. That is to say, powers endure even when unexercised and, when exercised as mechanisms, act even where, as in open systems, there is no one-to-one relationship between their mode of operation and the particular sequence of events that occurs. In short, mechanisms act in their characteristic manner outside the closed conditions of their experimental identification. Such endurability, indeed, is equally a precondition (and so explanation) of ex posteriori successes in repeated trials, whether in a single location or in different laboratories. Third, the experiment is a situation in which the scientist must usually both enable or trigger the mechanism whose existence or nature is being invest-igated, to ensure that it is active, and also prevent any countervailing mechanism from interfering with the experimental outcome. The aim is to get a single mechanism going and to record its effects. Only if © 1997 Tony Lawson
a mechanism is active and the system so controlled can scientists reasonably expect, or base their evaluations upon, the occurrence of an empirical regularity. In short, the stimulus and enabling conditions for the mechanism must be satisfied, and the mechanism must be insulated and the flux of conditions held constant or otherwise controlled. Now while there clearly is a difference between the conditions of the natural and social realms the dividing line can get drawn in the wrong place. Collier appears to agree with Bhaskar (1979) that the ‘chief epistemological limit on naturalism’, i.e. the epistemological feature of the social sciences which qualifies the manner in which the thesis of naturalism can be sustained, is that social scientific objects ‘only ever manifest themselves in open systems’ (1979:45). The claim involved here is true as it stands. But I am not sure that it constitutes a limit to naturalism. At the very least something more needs to be said. First, by interpreting the inescapable openness of the social domain as a limit to naturalism there is a risk of encouraging the inference that the natural sciences can be reduced to the experimental natural sciences or astronomy, a conflation, I suspect, that neither Bhaskar nor Collier would accept. We understand the causes of earthquakes, the formation of mountains, and the spread of diseases without access to, or being able to engineer, appropriate closures. Indeed, some of the more promising developments in contemporary physics collected under the heading of superstring theory focus upon a postulated domain of reality where experimentation, at least for the foreseeable future, is completely infeasible (see e.g. Witten, 1988). Second, it remains the case that, even where controlled experimentation is possible in the natural sphere, the goal of a perfect closure, turning as it does upon system or mechanism insulation, constitutes an ideal scenario that cannot always adequately be engineered; indeed it may very rarely be. Problems of replicating the experimental results of others (or oneself), of ‘false alarms’, of inconclusive measurements and results, of unaccounted for disturbances, of equipment failures, of the often prohibitive costs of experimental equipment and so forth, are legion. Even at the conceptual level uncertainty will always remain as to whether the experimental design anticipates all the disturbing mechanisms that must somehow be held at bay. Third, and perhaps most significantly, these practicalities and/or limitations of the experimental sciences must be set against an adequate conception of any particular non-experimental contrast. It is certainly reasonable to doubt that controlled experimentation will ever be particularly meaningful in economics due to the impracticality of manipulating social structures and mechanisms in order more clearly to identify them. But just as this does not entail that relatively enduring social structures and mechanisms © 1997 Tony Lawson
do not occur, so it does not necessitate a priori that they cannot be detected. They cannot be intentionally activated, at least not by the investigative economist; so evidence of their operation must always be sought or discovered, not generated. Thus, the initiation of explanatory research is necessarily backward, not forward, looking. Nor can mechanisms of interest be insulated in the manner of experimental manipulation or the background conditions be purposefully held constant or otherwise controlled. Yet, this in itself does not rule out the possibility of mechanisms being detected. The fundamental point, here, is that it is not the case that the only conceivable alternative to well controlled experimentation, or more generally the production of a closure, of a strict event regularity, is a totally unsystematic, incoherent, random flux. Over restricted regions of time-space certain mechanisms may come to dominate others and/or shine through: nonspurious, rough and ready, partial regularities may be observed. Although the social world is open, dynamic and changing, certain mechanisms may, over restricted regions of time-space, be reproduced continuously and come to be (occasionally) apparent in their effects at the level of actual phenomena, giving rise to rough and ready generalities or partial regularities, holding to such a degree that prima facie an explanation is called for. If it is conceivable that a completely non-systematic flux could have dominated, ex posteriori this is evidently not the case. Thus, just as autumn leaves do fall to the ground much of the time, so woman are concentrated in secondary sectors of labour markets and productivity growth in the UK over the last century has frequently been slower than that of most other, otherwise comparable, industrial countries, and so on.
INTERPRETING PARTIAL EVENT REGULARITIES Let me refer to such partial regularities in the first instance as demiregularities or demi-laws.1 A demi-regularity, or demi-reg for short, is precisely a partial event regularity which prima facie indicates the occasional, but less than universal, actualization of a mechanism or tendency, over a definite region of time-space. The patterning observed will not be strict if countervailing factors sometimes dominate or frequently co-determine the outcomes in a variable manner. But where demi-regs are observed there is evidence of relatively enduring and identifiable tendencies in play. Now, a realisation that such demi-regs do occur, and the suggestion that they are even pervasive, may, on the face of things at least, seem rather remarkable. After all we have observed that the social world is perpetually changing, being continually transformed. Even supposing that certain structures and mechanisms of interest are often reproduced over significant © 1997 Tony Lawson
regions of time-space and that countervailing mechanisms are such that the primary mechanism often dominates and/or shines though them, the numerous background factors which constitute the social and natural environment nevertheless remain in play. In natural experiments any mechanism that is empirically identified is only isolated, or better insulated, in a relative sense; there are an uncountable number of factors held constant or controlled. How, then, can there be a pervasion of demi-laws or demi-regs in the intrinsically dynamic and open social world which lacks the possibility of experimental control?
EXPLAINING THE PREPONDERANCE OF DEMI-REGS To understand how demi-regs, as interpreted here, nevertheless are widely in evidence it is insightful to consider still further the typical conditions of experimental control. Not all experiments aim to hold background conditions constant. Such constancy seems essential if repeated trials are to be carried out. But, if this situation has no obvious parallel in social science, it is often, and perhaps usually, impossible in natural science too. Rather, experimental control frequently takes the form of comparing two different groups or populations with common or similar histories and shared (if non-constant) conditions, excepting that one group is ‘treated’ in some definite way that the second control group is not. Alternatively put, such experiments consist in observing two (or more) groups of ‘participants’ experiencing broadly similar conditions excepting (at least) one factor which is varied over the two groups in a controlled way. Experiments in plant breeding provide obvious examples. Often two sets of plants of the same type are grown in identical conditions except for (at least) one factor, perhaps a new fertiliser, which is varied over the two sets. Or, more typically, when various (similar but non-uniform) background factors such as soil composition and light are not directly controllable, it may be possible to divide the relevant land into a set of plots and then attempt to assign certain quantities of fertiliser to the various plots in a random way, with some plots receiving no fertiliser at all. Under such conditions the difference between the mean yield of the unfertilised plots is contrasted with that of fertilised plots to see if there is a systematic and significant difference, which can be attributed to the fertiliser. To repeat, a significant aspect of the experimental process is the existence of both a primary group and a second control group that acts as a foil. The aim is to link specific effects to a particular causal factor by having it operate in one of the two sets of situations but not the other. And it is such a relationship between aspects of a primary group and those of a relevant © 1997 Tony Lawson
contrast group which appears to explain the experience, nature and significance of most social demi-regs. That is, most social demi-regs capture reasonably systematic differences (or more generally patterns) at the level of actual outcomes between two groups whose causal histories are such that the outcomes in question might reasonably have been expected to be broadly the same, or at least to stand in some definite anticipated or plausible relationship which is systematically at odds with what we observe. We do not and could not explain the complete causal conditions of any social or other phenomenon. To do so would presumably mean accounting for everything back to the ‘big bang’ and beyond. Rather we aim to identify single sets of causal mechanisms and structures. And these are indicated where outcomes or features of different groups are such that, given the respective causal histories and conditions of these groups, their observed relation is other than might have been expected or at least imagined as a real possibility. Let me refer to empirical facts about relationships (typically differences) of this sort as contrastives. Although it is possible to argue that most explanation is concerned with such contrastives,2 not all contrastives necessarily rest upon the operation of enduring causes. (Thus, the immediate reason why Heather went through the door before Hannah is that Hannah paused to pick up some object she dropped on approaching it.) To the extent, then, that science is concerned to identify and understand relatively enduring structures and mechanisms the starting point will often be contrastives with at least some significant degree of space-time extension. Given that few scientifically interesting empirical contrasts involving social phenomena are likely to hold strictly or uniformly, the aim is to uncover those that at least hold to an extent that an enduring mechanism seems the likely explanation of the contrast involve. In short, the initiation of much social scientific research will inevitably depend upon the detection of contrastive social demi-regs.
CONTRASTIVE DEMI-REGS The world might have been such that partial regularities of interest were never, or perhaps only rarely, in evidence. In fact, it is clear that in the social sphere, as elsewhere, contrastive demi-regs are pervasive at all levels. For example: ‘women look after children more often than men do’; ‘a relatively small proportion of children from poor backgrounds in the UK continue into higher education’; ‘average unemployment rates in western industrial countries are higher in the 1990s than the 1960s’; ‘in the 1990s UK firms are externalising or “putting out” more parts of the production process than twenty years ago’; ‘in the late nineteenth century UK firms increasingly internalised parts of the production process’; ‘an increasing proportion of the © 1997 Tony Lawson
world’s population lives in cities’; ‘women in the UK usually wear brighter colours, use more make-up, but go alone to the pub less often than men do’; ‘the proportion of the UK public that reveals an intention to vote for the Conservative Party increases in run-ups to general elections’; ‘government spokespersons tell more lies in war-time’; ‘reported crime in the UK has increased steadily since the 1970s’; ‘Cubans currently spend more time in queues than the English, who in turn spend more time in queues than Italians’; and so on. Or, at a general level, the persistence of inflationary trends in certain economics but not others, of significant variations in rates of growth or decline of area-specific manufacturing sectors, of poverty in the midst of plenty, of production primarily for exchange rather than, as previously, for immediate use, provide examples of notable space-time patterns in economic phenomena. In each such case there is not an invariable relation but repetition of such a nature, or to such a degree, that an explanation seems required all the same. In each such case it is to be expected, or anyway is prima facie plausible, that there are systematic and identifiable mechanisms in play which social science can uncover. The point I want to draw out here is that these partial regularities or contrastive demi-regs facilitate social science properly conceived, even if they are inadequate to the (misguided) objectives and needs of orthodox economists. Let me develop each part of this assessment in turn.
CONTRASTIVE DEMI-REGS AND SCIENCE The most obvious role for observable rough patterns or demi-regs is that they can serve to direct social scientific investigations, through providing evidence that, and where, certain relatively enduring and potentially identifiable mechanisms have been in play. And by their nature, by the fact that they are patterned at all, they reveal something in turn of the nature of the tendency in play. Notice that this role for contrastive demi-regs in the initiation or directing of research is the same in social and natural (including experimental) sciences alike. Thus, the natural sciences often take their leave from signs of some recurring animal or plant disorder, evidence of environmental deterioration, unanticipated ‘side-effects’ of some important experiment, and so on. In both social and natural sciences researchers can and do start from empirical facts that do not invariably hold, yet hold in a sufficiently large number of cases that an explanation appears to be required. Of course, a recognition of the role of broad but less than strict regularities in the initiation of scientific investigation is hardly novel, even in economics. Kaldor, for example, explicitly argues the need to initiate explanatory research in this way. In detailing certain ‘stylised facts’ Kaldor remarks that © 1997 Tony Lawson
we do not imply that any of these ‘facts’ are invariably true in every conceivable instance but that they are true in the broad majority of observed cases—in a sufficient number of cases to call for an explanation that would account for them. (Kaldor, 1985:9) I should add that although previously I have adopted the terminology of stylised facts myself, I now consider this to be strategically unwise. For I take it that usage of the term stylised in this context means3 something like ‘to cause to conform to a style of expression often extreme in character rather than the appearance of nature’ (Webster’s Third New International Dictionary). In other words, a supposed ‘stylised fact’ is intended to express a partial regularity reformulated as a strict one, in the form of a law. Kaldor, of course, saw this as a matter of mere presentation as we have seen, and always emphasised the role of such ‘facts’ as suggestive of a phenomenon in need of an explanation. Increasingly, however, we can see mainstream deductivist modellers attempting to legitimise their ‘whenever this then that’ formalisations as stylised facts interpreted as Kaldorian. In other words, the terminology, or rather its lineage, is being used to attempt to justify some fictitious set of idealisations, with the aim, in turn, of facilitating nothing more than model tractability, or some such. It is in order to avoid further encouraging this trend that I have preferred the terminology of demi-regs. It serves as a continual reminder that most empirical facts do not take the form of strict regularities, that they express phenomena to be explained, not the end-points of research or mere devices to be built into formal systems.
THE DETECTION OF INTERESTING DEMI-REGS It is easy enough to see how demi-regs may be identified. I have already indicated that the important category here is that of contrast or comparison and especially difference. The significance of patterns collected under the heading of demi-regs usually turns upon comparisons, and in particular upon differences: between men and women; or old and young; events or states of affairs (rates of unemployment, inflation or profits, etc.) in the UK today compared with twenty (or a hundred) years ago, or in modern Britain compared say with the continent of Europe; and so forth. In other words, we notice the effects of sets of structures through detecting relatively systematic differences in the outcomes of prima facie comparable types of activities (or perhaps similar outcomes of prima facie different activities) in different spacetime locations, or differences in types of position-related activities in comparable space-time locations, and so forth. © 1997 Tony Lawson
Moments of social upheavals, crises and disruption may be especially revealing in this respect. Bhaskar (1979) explicitly remarks that moments of social transformation ‘provide a partial analogue to the role played by experimentation in natural science’ (p. 48). I think they do. But such moments should be recognised as particular, albeit highly significant, instances of the wide range of occasions where mechanisms or processes become visible. In the case of social upheavals the contrast is largely temporal. And, indeed, generative mechanisms become that much more accessible at any geohistorical turning point. But such situations provide insight into underlying tendencies in just the same manner as the contrasting positioned activities (and non-activities) of, say, men and women, or of immigrant or native, old and young, of people in different geographical regions or cultural contexts. In the social realm such contrastives all tend to indicate something of the positioned structures that people are acting on. In short, any patterning, any ‘standing-out’, of phenomena which turns upon differences or unanticipated or surprising or implausible relationships of some kind, whether primarily social, historical or geographical, can serve to alert us to the existence or way of acting of some item previously unknown, unrecognised or perhaps known only implicity, in some taken for granted way.
SCIENCE AND SCIENTIFIC INTERESTS Let me briefly make explicit certain implications of this conception. I am suggesting that scientific explanation is inherently contrastive. We have seen that in accounting for some social phenomenon the aim could not be to provide its complete causal history. Rather, we can only aim to identify one (set of) causal mechanism(s). And to this end the obvious strategy is to seek out two (or more) situations where the outcomes might have been expected to be related in some manner other than turns out to be the case, and to attempt to determine the reason(s). Typically, this will involve identifying at least one mechanism that operates, or does so in a particular fashion, in the one (set of) situation(s) only. In consequence, we can now see even more clearly than hitherto that explanatory projects are inherently dependent on the interests of those involved. For it is now evident that the interests of the investigator influence not only the choice of phenomenon to be explained, but also, by selecting the contrast, the particular explanatory mechanism to be researched. For example, suppose that the primary phenomenon that we wish to explain is the average reported level of a certain crime in the UK in the mid-1990s. Then it
© 1997 Tony Lawson
is clear that the explanation provided, the causal factor identified, will vary according to whether the contrast, reflecting our interests, is, for example, 1 the situation that prevailed in the UK twenty years previously when the corresponding figure was significantly lower; 2 the current situation today in country x, say the US, where the corresponding figure is perceived to be significantly higher; or even 3 a desired situation of zero reported crime. In each case, an explanation considered to be satisfactory will identify at least one systematic difference between the causal history of the primary component and that of the chosen contrast, or which would appear to be essential for the contrast if, like zero crime, it is only an imagined situation. In each case, the set of causal factors responsible is likely to be different. Now for an explanation to be warranted it does appear that any contrast that is only imagined must be recognised as at least a real possibility (or perhaps once having been a real possibility). Perhaps few people living in modern western conditions conceive zero reported crime to be even remotely achievable. To the extent then that zero crime appears compatible with our species-being but unimaginable in practice, any explanation of the existence of crime presumably turns upon why or how certain types of societies or human relationships arose, and are reproduced, rather than others. But if this situation seems all too speculative it is easy enough to observe all sorts of quite plausible alternatives to the events and states of affairs, including persistent ones, that have actually come about. The aim then is to identify a causal factor (including perhaps an absence) which contributed to the state of affairs actually pertaining, but which would not have facilitated the imagined or expected, or did not condition an actual, alternative. Whatever else may be characteristic of the positioned social practice which is science, it is highly interest-dependent.
INCONSISTENCY, SURPRISE AND CRITICISM I turn at this point to a rather significant feature of the approach I am developing, one that has been implicit throughout this chapter and is bound up with the interest-dependent nature of all scientific practice, but which, mainly for ease of exposition, I have yet to elaborate upon explicitly. I refer to the fact that enquiry will usually be initiated not just by any partial regularities and/or contrasts, but those that, along with other beliefs, occasion certain contradictions, inconsistencies, experiences of surprise and ultimately doubt. I have indicated that contrastive social demi-regs are easy enough to © 1997 Tony Lawson
come by. Often, though, they will be quite uninteresting in the sense of generating little surprise from the point of current understandings. For example, in the UK today ‘on average more people go to church on a Sunday than on Tuesday’; and ‘people with smaller incomes on average spend less than those with larger incomes’. While these observations, like all beliefs, are fallible, they seem unlikely to conflict with the current standpoints, and in particular the deeply held beliefs or theories, of social scientists or indeed people in general. That is, not only are these observations unlikely to pose problems of theory doubt for any individual but equally they appear unlikely to condition significantly different, competing explanations, a situation which would itself be a stimulus to further enquiry.4 In contrast, reported observations that ‘productivity growth in the UK has usually been slower than that of other major industrial countries’; or that ‘children in single sex schools in the UK perform significantly better academically than children in mixed schools’, have occasioned surprise in many and been met with an array of competing explanations. Consider also the attention recently given in economics (David, 1985, 1986) to studying and explaining the widespread employment of computer keyboards where the top row spells QWERTYUIOP (an example discussed further in Chapter 18). Presum-ably, if the letter arrangement had been something like ABCDEFGHIJ then we would not so readily identify it as something in particular need of being explained, at least in areas where the Roman alphabet is dominant. In short, theoretical explanatory enquiry is likely to be initiated or further stimulated where contrasrive demi-regs occasion a sense of surprise, doubt or inconsistency, either between the observed phenomenon and a set of prior beliefs, or between competing explanations of it, and so forth. This part of my argument should not itself occasion any surprise. For an emphasis on the sort of inconsistencies in question has not only been implicit throughout this chapter but explicit throughout the book. The opening chapters were oriented to elaborating and resolving various theory/practice inconsistencies that pervade the subject; in Chapter 5 it was argued that the central role for philosophy is immanent critique (exposing internal inconsistencies in beliefs implicit in practices, or demonstrating how beliefs held cannot accommodate practices actually achieved); and in the same chapter we saw a specific form of immanent critique, namely determinate negation, wherein only a transcendental realist conception is capable of sustaining certain practices which competing philosophies of science sponsor. Science is like philosophy, then, in that both are stimulated by puzzles, contradictions and inconsistencies. And science is also like philosophy in that both are concerned fundamentally with transforming, including negating aspects of, some set of beliefs. In brief, according to the account which I am arguing for, science, like philosophy, is inherently critical. There are various reasons for the critical in ‘critical realism’. © 1997 Tony Lawson
CAUSAL HYPOTHESES The detection of contrastive social demi-regs of interest, then, can be used to initiate the investigation of causal factors. And the objective can be recognised as determining one or more factors directly responsible for any contrastive demi-reg identified, to identify factors that were operative in one set of conditions but not the other that helped produce or facilitate the contrast in question. But is there anything to be said about the process of reasoning by which causal hypotheses are obtained? We saw in Part I that the central mode of inference is neither deduction or induction. Rather it is retroduction. The aim is not to cover a phenomenon under a generalisation (this metal expands when heated because all metals do) but to identify a factor responsible for it, that helped produce, or at least facilitated, it. The goal is to posit a mechanism (typically at a different level to the phenomenon being explained) which, if it existed and acted in the postulated manner, could account for the phenomenon singled out for explanation. Not much can be said about this process of retroduction independent of context other than it is likely to operate under a logic of analogy or metaphor and to draw heavily on the investigator’s perspective, beliefs and experience. Notice, parenthetically, that this process of retroducing explanations of contrastive demi-regs is once more as significant in philosophy as in science. Bhaskar’s observation that outside astronomy most event regularities of interest in science are brought about in conditions of experimental control is itself a contrastive demi-reg. It is this contrastive demi-reg which conditions Bhaskar’s transcendental arguments for a structured ontology. Thus, just as substantive contrastive demi-regs as well as any experimentally confined specific regularities serve as bases for retroductive inferences to hypotheses about particular causal mechanisms, so the contrastive demi-reg of experimentally confined regularities serves as basis for the transcendental argument to transcendental realism. Particular differentiations of the world to hypotheses about specific mechanisms; generalised differentiations to philosophical ontologies. Transcendental reasoning is thus but a special case of the retroduction.
EXPLANATORY POWER If an obvious role for contrastive demi-regs in economics, one which they also fulfil in the experimental sciences, lies in the directing of investigative research, a second role is as an aid in assessing causal hypotheses once they have been formulated. We have seen that economics and the social sciences generally are denied the crucial test situation. However, the consequences of © 1997 Tony Lawson
this for the process of theory assessment is merely that event-predictive accuracy cannot be the criterion of theory selection. Rather the appropriate criterion outside of the controlled-experimental (or any fortuitously spontaneously closed) situation must be explanatory power. Theories can be assessed according to their abilities to illuminate a wide range of empirical phenomena. And typically this will entail accommodating precisely such contrastive demi-regs as are recorded or can be found. Several aspects to the process of assessing a theory’s explanatory power can be anticipated. The first relies on deduction. The point is to deduce from any retroduced hypotheses those consequences or effects which would follow if the hypothesis were true and the mechanism operative. The second involves checking out the various deduced consequences empirically. With a permanent possibility of countervailing factors there can be no guarantee that any such effects will be straightforwardly manifest. But the aim must be to try and identify conditions where, in the light of all that is known about the situation, the effects ought in some way to be in evidence. A third aspect to the process involves explaining the explanation. It includes identifying the conditions of any explanatory mechanism and checking they are or were operative. Let me expand upon these themes. The possibility of empirically assessing theories in the absence of crucial test conditions is perhaps clearest in the (most common) situation where the objective is to select amongst two or more hypotheses. For in this case we are ultimately interested in the relative performance of hypotheses whatever the relevant selection criteria. It is thus straightforwardly reasonable to search out that theory whose consequences appear mostly born out and which illuminates the widest range of empirical phenomena including any intersection upon which all competing theories have some possible bearing. Relative explanatory power is likely to be sufficient as a well as appropriate here. Of course, whether or not the explanatory power of a theory is easily assessed necessarily depends on context. Ultimately assessments of the empirical adequacy of hypotheses always do. 6 Consider an example formulated by Leamer (1983). Econometricians, according to Leamer, like to project themselves in the image of agricultural experimenters who, in a manner similar to that already attributed to plant breeders above, sub-divide farm land into a set of plots and then assign random quantities of fertiliser to each plot, with some plots receiving no fertiliser at all. We have already noted that under these conditions it may be supposed that the difference between the mean yield of the fertilised plots and the mean yield of the unfertilised plots is a measure of the effect of fertiliser on agricultural yields. The task of the agricultural experimenter is to determine whether the difference, assuming it is positive, is large enough to suggest a real effect of © 1997 Tony Lawson
the fertiliser (or so small that it could plausibly be attributed to random fluctuations). Against this preferred scenario, one approximating the situation referred to as a stochastic closure in Chapter 7, Leamer’s assessment is that the econometrician’s predicament is rather different: This image of the applied econometrician’s art is grossly misleading. I would like to suggest a more accurate one. The applied econometrician is like a farmer who notices that the yield is somewhat higher under the trees where birds roost, and he uses this as evidence that bird droppings increase yield. However, when he presents this finding at the annual meeting of the American Ecological Association, another farmer in the audience objects that he used the same data but came up with the conclusion that moderate amounts of shade increase yields. A bright chap in the back of the room then observes that these two hypotheses are indistinguishable, given the available data. He mentions the phrase ‘identification problem,’ which, though no one knows quite what he means, is said with such authority that it is totally convincing. (Leamer, 1983:31) Notice first how the imagined investigation is initiated by an observed contrastive demi-reg of interest. Yields are higher under trees than elsewhere. The problem is that two competing explanatory conjectures both seem consistent with the evidence. The obvious response of course, albeit one that econometricians occupied with fitting a line to given sets of data rarely contemplate, is to add to the ‘available data’. Specifically the aim must be to draw consequences for, and seek out observations on, actual phenomena which allow the causal factor responsible to be identified. If, for example, bird droppings is a relevant causal factor then we could expect higher yields wherever birds roost. Perhaps there is a telegraph wire that crosses the field which is heavily populated by roosting birds, but which provides only negligible shade in comparison (and shade which moves significantly as the earth revolves around the sun). Perhaps too there is a plot of land somewhere close to the farm house which is shaded by a protruding iron roof, but which birds avoid because of a patrolling cat. And so on. It all depends. The fact that it is not possible to state categorically at this abstract level the precise conditions under which substantive theories can be selected amongst, i.e. without knowing the contents of the theories themselves or the nature or context of the conditions upon which they bear, is an unfortunate fact of all science. Science is a messy business. It requires an abundance of ingenuity,
© 1997 Tony Lawson
as well as patience, along with skills that may need to be developed on the job.7
PROBLEMS IN DISCRIMINATING BETWEEN THEORIES It must be admitted that there will be occasions when, even with a range of types of data, it still proves difficult to discriminate between competing hypotheses. But this is the experience of all the sciences. The strategy in such circumstances can only be to continue to search out conditions for which the competing hypotheses bear different implications with regard to empirical phenomena and then to check out which hypothesis proves to be the most empirically adequate in those conditions. Any such difficulties of discrimination do not imply problems for the philosophy of science perspective I am defending, of course, they merely indicate situations in which there are practical difficulties in science. If two or more hypotheses do really appear to command the same degree of empirical adequacy then, ceteris paribus, the correct epistemic attitude,8 whatever the science, must be to attach the same degree of belief to each, or to suspend judgement. Notice that the possibility of difficulties in discriminating between theories does not turn upon whether or not the hypothesised causal mechanism is observable or otherwise. Although Boylan and O’Gorman (1995) invoke Quine’s under-determination thesis to suggest that there may be a multiplicity of empirically (or ‘descriptively’) adequate theories and explanations in terms of non-observable items, there is no reason to suppose that this applies any less (or more) to hypotheses concerning observables-in-principle.9 Is the perpetrator of the crime always the only suspect to fit the available empirical facts, and is the guilty person always the one convicted? Perhaps Leamer’s agricultural econometrician does not after all find a mid-field telegraph wire or a protruding farm house roof, while other data-consistent hypotheses come to mind: the concentration of nutritious leaf-mould near the trees, the windbreak or partial rain (or air-borne pollution) shelter provided by the trees, the proximity of the trees to a previously unconsidered brook, and so on. More to the point, if to elaborate upon the argument of the last chapter, unobservable natural forces as well as structures of a language or rules of a game can be, and frequently are, described and known. Thus, while the unobservable wind can be known to be responsible for the flying hat or the swirling leaves, the rules of grammar facilitating the speech acts of some group, or the secret code used by children to send messages to each other, may each be quite unproblematically retro-ducible and/or describable. By demonstrating the empirical adequacy of theories of gravitational and magnetic fields, social rules and relations, structures of languages and so © 1997 Tony Lawson
forth, these items can be known whether or not they can be directly observed. I have been discussing a situation of competing hypotheses. But it is not just in the comparing of hypotheses that demi-regs can be used in the manner described. Even a single maintained hypothesis can be continually assessed by examining the range of phenomena it bears upon. In other words, the empirical adequacy of any hypothesis can be progressively checked-out by deducing such implications as follow, with regard to any domain for which such implications hold and can in practice be ‘observed’. Where it is argued, for example, that gender relations in many societies are such that mechanisms are reproduced which serve to discriminate against women in favour of men, this assessment can be checked out against detectable patterns occurring in, for example, factories, homes, schools and universities, and any other place where the mechanism will conceivably reveal its effects.
RESPONSES TO EXPLANATORY FAILURE Needless to say, there can be no expectation that each stage of the proceedings will always be clear cut, even in those cases where the evidence available appears not to support a hypothesis in contention. For example, if definite mechanisms are hypothesised to be operative in certain situations where their anticipated effects do not at first sight appear to be in evidence, this outcome may warrant examining further, where possible. For, if there is reason to suppose that in such conditions a specific mechanism is nevertheless operating transfactually, it should be possible to identify offsetting countervailing factors. If, to continue the last example, it is thought that mechanisms which work to discriminate against women operate throughout the economy, yet it is discovered that in a particular factory women do not predominate in the positions regarded as secondary, further investigation may reveal that this is because the least desired jobs are given to (possibly illegal) immigrants, most of whom happen to be male, or to some other locally discriminated against group, and so forth. One obvious response to such a finding, then, is to examine situations where any such suspected countervailing tendencies are unlikely to be in operation. Of course, this practice of examining whether (perhaps the most notable) exceptions to patterns can be explained in terms of countervailing mechanisms is not confined to social science. Recall, for example, the way in which Neptune was discovered. On recording observations of the motion of the planet Uranus, nineteenth-century scientists found that its orbit did not conform to predictions derived using Newton’s theory of gravity. Instead of rejecting that theory, however, they looked for countervailing factors which © 1997 Tony Lawson
could account for the noted discrepancies. Specifically, Adams in England and Leverrier in France conjectured the existence of a previously unknown planet somewhere in the vicinity of Uranus. In response Galle turned to focus his telescope on the appropriate region of the skies and came to discover the planet we now know as Neptune. In natural and social science alike, then, exceptions to an observed pattern do not necessitate an automatic rejection of a given theory. Rather they constitute specific contrastive demi-regs which mark sites where further investigative work can fruitfully be undertaken. The ultimate outcome necessarily depends on context.
ASSESSING THE REALITY OF A HYPOTHESISED MECHANISM A further component of the process of assessing the explanatory power of some hypothesis is checking the reality of any mechanism postulated. This may involve checking the ‘triggering’ conditions (if there are any) to see if the mechanism is in play. It is not enough to know that a suspect had the motive and know-how for a particular crime if he or she could not have been at the relevant scene when the crime took place. It is not good enough to argue, like Friedman (1953), that the hypothesis of mobile leaves moving about the branches of the tree searching out the light, explains the distribution of leaves on the tree, when we know the hypothesis of mobile leaves to be false. It is not good enough because, unlike Friedman, we have accepted (through argument and evidence) that the explanatory goal is to identify mechanisms, etc., really productive of any identified phenomenon of interest. Thus any hypothesis couched in terms of some mechanism known not to exist or to be in play cannot be said to be explanatory in the requisite sense at all. It is for this reason that the assessing of the reality of some hypothesised mechanism can be subsumed under the head of assessing that hypothesis’s explanatory power. In the social realm, however, if structures and mechanisms endure over stretches of time-space it can only be by way of human action. If then we wish to explain some relatively enduring contrastive demi-reg, a full understanding of the situation requires that the mode of reproduction of the identified causal mechanism be itself investigated. In other words, it is necessary that the conditions governing the reproduction (and perhaps ‘initial’ emergence) of any identified causal mechanisms be accounted for, that the explanation be explained. Again this may entail the explaining of a contrast: why and how certain structures or mechanisms have emerged and been continuously reproduced in one situation but not another. When an explanation of an explanation is successfully achieved, of course, we are likely to have greater confidence in each part of the overall © 1997 Tony Lawson
explanation of the original phenomena. Moreover, conceptions of the practices which led to research being undertaken in the first place may themselves require reinterpretation once an overall understanding is achieved, i.e. in the light of the broader explanatory picture. Needless to say, there will often be competing explanations of the original explanation; there may be several accounts of why or how certain mechanisms have (or have not) emerged, and, where they have, why they have been reproduced and/or transformed in particular ways. Where this is so the criterion of theory assessment, once more, can only be relative explanatory power in the sense described above. Notice, finally, that the process of economic explanation here is aided by the fact that many of the relevant social phenomena in any situation are already conceptualised under some description by the agents involved. In addition, the set of activities being explained may already have been redescribed by observers and/or other interpreters. The realisation that this feature does not represent a ‘compensator’ for the infeasibility of experimental control does not thereby undermine its generalised value to science.
THE REQUIREMENTS OF ORTHODOX ECONOMICS If contrastive demi-regs, as conceived above, are seen to be fundamental to social science properly interpreted and facilitative of likely successes, it is easy enough to see that they are nevertheless inadequate to the (misconceived) requirements of mainstream economics. The term demi-regs, we can recall, is used here to denote patterns or regularities of sorts, regularities that are recognisable as such despite being something rather less than strict.10 Now we have seen that there are essentially two sets of reasons why a recognisable patterning of events will usually not be strict, both of which were discussed at length in Chapters 7 and 8 in connection with the theses of regularity stochasticism and regularity determinism. First, the environment in which any mechanism acts need not be sufficiently homogeneous. In the social realm, indeed, there will usually be a potentially very large number of countervailing factors acting at any one time and/or sporadically over time,11 and possibly each with varying strength.12 This means not only that where mechanisms ‘shine through’ they are unlikely to do so in a continuous, unimpeded, clear-cut fashion, but equally that operative mechanisms need not always ‘shine through’ in any recognisable manner at all. The implication of the latter situation, as we have also previously noted, is that the usual starting point for research into the nature of social mechanisms will invariably be conditions where the effect of mechanisms have in some way already been detected. We start from situations where, fortuitously, relatively stable © 1997 Tony Lawson
tendencies are revealed. In this sense social scientific explanation is inherently backward looking. The second reason for the absence of strict regularities is that the mechanisms or processes which are being identified are themselves likely to be unstable to a degree over time and space. I see no a priori reason to suppose that any relatively enduring, transfactually acting, social mechanism need be particularly constant in the way it operates over time and space; nor am I aware of any evidence which indicates that any are. Indeed, given the fact of the dependence of social mechanisms upon inherently transformative human agency, where human beings choose their courses of action (and so could always have acted otherwise), strict constancy seems a quite unlikely eventuality. Just as there is always some continuity in social change, so there is usually some (and often quite substantial) change in social continuity. In short, if it is a special situation of the open world that certain mechanisms (whether natural or social) reveal themselves in rough and ready patterns, it is as a special case of this special situation that the patterns produced correspond to strict event regularities, including any consistent with well-defined probability ‘laws’. The latter can really be expected only where (a set of) intrinsically constant and separable causal factors act in (relative) isolation. While such special cases of special cases are obviously of value to science where they occur or are feasible, it can now be seen that they are not essential to it if (non-spurious) demi-regs are in evidence. But if strict event regularities are inessential to science in its endeavour to uncover causal mechanisms and to discriminate between competing hypotheses, they are essential to the requirements of orthodox economics. That is, they are essential to the preoccupation with predicting future states of the economy and/or to any hoped for deductivist ‘explanation’ of economic actualities, including events and states of affairs. In other words, mainstream economics models itself on, or aims to achieve, the one set of natural science possibilities which is not available to it. Lacking the opportunities for controlled experimentation, yet mistakenly supposing the strict event regularities which are sought after in such set-ups to be essential to the scientific enterprise, orthodox economists react by assuming that social phenomena are after all but the results of an experiment, one that just happens to be well-controlled by ‘Nature’, or at least that can usefully be treated as such (see e.g. Haavelmo, 1944:9, 14). Armed with this conception of social reality and with deductivist (predictive) goals and criteria, mainstream economists, and perhaps this group alone, conclude that such patterns as can be detected at the level of actual phenomena are never strict enough, with the result that such contrastive demi-regs as abound have far too little (or the wrong sort of) influence upon the theories that economists hold.
© 1997 Tony Lawson
PURE AND APPLIED EXPLANATION Now if contrastive demi-regs abound they too nevertheless constitute a special case of a more general situation. It is to be expected that many aspects of social events are relatively unique occurrences, being the conjoint effects of numerous mechanisms acting simultaneously. In other words, there is a possibility, already noted, of a continuum of pattern outcomes stretching from closed systems of constant conjunctions of events to an inchoate random flux, with contrastive demi-regs lying between these extremes. In consequence, the question arises as to whether it is possible to explain the range of pattern outcomes. For it will often be of interest to understand quite novel or unusual social phenomena, whether spectacular ones such as stock market crashes or major increases in OPEC prices, or less dramatic ones such as the demise of local markets or large changes in local prices. In order to account for the range of actual phenomena it is necessary that economic explanation be divided into two relatively distinct movements or separate modes of activity. In fact, most of the discussion so far has implicitly concerned a mode of inference that should really be termed pure or abstract or theoretical explanation, the identification of underlying structures, powers, mechanisms and their tendencies. A necessary condition for this explanatory activity is that certain relatively stable and enduring mechanisms do at some time and place come to be reproduced, to endure, and do to some extent dominate, or shine through, others—at least to a degree that rough or partial event regularities of sorts are discernible. However, it is equally apparent that a second mode or type of inference, which is appropriately termed an applied or concrete or practical explanation, is also called for. For, to the extent that (relevant features of) concrete phenomena of experience are relatively unique or novel, being conjunctures of numerous countervailing tendencies, their explanation entails drawing upon antecedently established knowledge of relatively enduring structures and mechanisms (rather than revealing them), and investigating the manner of their joint articulation in the production of the novel event in question.13 Again this is a situation that holds just as much in natural science. Seismologists know the mechanism by which earthquakes are produced but can only explain particular manifestations after the event. Similarly, meteorologists do not always forecast the weather very well, yet after the event they convincingly draw on their (well grounded) knowledge of relevant physical principles to explain whatever took place. For sake of clarity let me briefly contrast the modes of reasoning employed in theoretical and applied explanation. Clearly neither of the identified modes of explanation is primarily deductive or inductive in form. Rather, in each case the aim is to redescribe some phenomenon under a new scheme of © 1997 Tony Lawson
concepts designating the structures, mechanisms or agents that are to some degree responsible for it. Theoretical explanations, we have seen, are characteristically analogical (scientists first searched, albeit unsuccessfully, for a virus responsible for ‘mad cow disease’ because viruses have so frequently been found to be responsible for disorders in animals previously) and retroductive (positing mechanisms which, if they were to exist and act in the postulated matter, would account for phenomena singled out for explanation). In short, theoretical explanations entail transforming existing cognitive resources into plausible theories of the mechanisms responsible for identified (typically less than strict) patterns of phenomena. These theories are then empirically assessed, of course, and, when found to be empirically adequate, themselves explained in turn, in the continuous unfolding of explanatory knowledge. Applied explanations, in contrast, are characteristically resolutive and retrodictive. They entail, first of all, the resolution of conjunctions or complexes, and the redescription of their components. This is followed by the determination (retrodiction) of possible antecedents of these components, and the empirical elimination of possible causes. For example, if we attribute weather pattern x to a particular combination of (already understood) causal mechanisms y, it is necessary to determine (retrodict) the conditions for y and then to check empirically whether these conditions actually obtained.
ECONOMICS AS AN EMPIRICAL AND ABSTRACT SCIENCE If it is by now clear that the explanatory project that I am arguing for is contrastive, interest laden and critical, it should also be apparent that it necessarily contains a significant empirical component. The measuring and recording of states of affairs, the collection, tabulation, transformation and graphing of statistics about the economy, all have an essential (if usually nonstraightforward) role to play. So do detailed case studies, oral reporting, including interviews, biographies, and so on. Indeed, I suggest it is precisely to such indispensible activities that the heading of econometrics is properly attributed. It is the case, furthermore, that the detection of non-spurious patterns will often require a good deal of more specialised, perhaps rather technical, knowledge and understanding of relevant situations. This is likely to be so whether the contrastive demi-regs are of the more mundane sort: ‘sportsperson x has not been performing well for the last n games’, or of a more unfamiliar and complicated variety: ‘the usual rough correspondence between the type of product market in which a firm operates and the employee © 1997 Tony Lawson
conditions it sustains (itself a demi-reg) has been systematically contravened over the last n years by the conditions obtaining in firm x’.14 Certainly, the ability to observe or detect patterns will necessitate looking at situations not only in the light of current understandings but also from different angles, at varying levels of generality, under varying space-time extensions, and so forth. Once more, the need to discern or detect patterns which may not be immediately obvious is a requirement that does not fall on social science alone. Currently, for example, symmetry principles play a prominent role in the development of elementary particle physics principally because they have been found to reveal patterns in the properties of prima facie quite distinct particles. Once these patterns are observed it frequently proves possible to explain them in terms of, i.e. to identify, the underlying forces (see e.g. Green, 1988). A further, and well known, example of the use of symmetry patterns in natural science is provided by nineteenth-century chemistry, with Mendeleyev’s discovery that the chemical elements could be arranged into groups with certain properties in common, i.e. with his elaboration of the periodic table. This led to an understanding of how elements are made of atoms and how, in consequence, many dozens of elements could be grouped in the suggested manner. It is evident, then, that there is a need for attention to be given to the processes of elaborating any such patterns. Clearly, they involve looking at phenomena in a one-sided way, focusing on some attributes to the neglect of others. I have already rejected any isolationist strategy that treats definite features as though they exist cut off from everything else. Rather, the relevant procedure here is abstraction. The failure to consider this procedure in any explicit fashion so far is an omission I turn to rectify in Chapter 16. Abstraction, though, is an essential feature of all cognitive enterprise; it is no more fundamental to science than any other social practice. And if the need for abstraction is more apparent when we consider seemingly complex patterns such as noted above, I must emphasise that it is nevertheless inevitably employed in all acts of apprehension, no matter how apparently simple the item involved (see, especially, Whitehead, 1926). Within science it is also fundamental to all aspects of the endeavour, not merely to the description of patterns in actual phenomena. Even so, one specific and important task for it lies in the delineation of such rough and ready patterns as arise, conceptions which provide, amongst other things, a marker where science might with reason, and some expectation of achieving illumination, continue its work.
© 1997 Tony Lawson
THE APPARENT FAILURES OF SOCIAL RESEARCH INCLUDING ECONOMICS Before considering such matters explicitly, however, there is one loose end to tidy up. There remains the question of what to make of Collier’s observation of the prominence of the hermeneutic moment in the social sciences and his belief that this indicates their relative failure. Now it is essential to exercise care here in interpreting any claim that the social sciences have failed to emulate the successes of natural science. Of course the naive attempts of orthodox economists to extrapolate forced correlations and to ‘explain’ on the basis of the deductivist model, have contributed very little to understanding the world in which we live. Nevertheless, endeavours at elucidating the nature of aspects of society and economy on the part of others (and even by mainstream academic economists when forsaking their official postures) may have been more fruitful than is usually imagined. On the one hand there is always the possibility, already noted, that insights obtained will have been appropriated by lay agents and incorporated into their activities. This can have the effect that knowledge which proved to be revelatory when it was obtained, eventually takes on the appearance of the banal or of common sense. On the other hand, there are various further considerations that are relevant here arising out of the situation that social structures and mechanisms seem usually to be more highly space-time specific than natural ones. A natural scientist working on gravitational fields or the molecular structure of copper may be confident that any illumination obtained will be as relevant in, say, contemporary Cambridge as in Beijing in a hundred years time (or a hundred years ago). But an economist examining the nature of industrial relations currently in place in a Cambridge firm may not even find them in place in a second Cambridge firm along the road, let alone in Beijing in a hundred years time or one hundred years past. In short, social scientific knowledge, when gained, will rarely have as wide a sphere, or scope, of relevance as natural scientific findings. But this, by itself, does not entail that the processes of social study are necessarily any less revelatory or scientific in nature, or less successful in illuminating their objects of study in fact. Perhaps this latter assessment requires further elaboration. Bhaskar, in his original contribution on the matter, even argues that any acceptable version of naturalism is qualified by the fact ‘that social structures, unlike natural ones, may be only relatively enduring (so that the tendencies they ground may not be universal in the sense of space-time invariant)’ (Bhaskar, 1979:38). But this statement seems incorrect in its characterisation of the objects of natural science. Certainly any social tendency in play will be dependent upon certain social structures being in place. But the operations of any natural © 1997 Tony Lawson
tendencies will be conditional upon natural structures in exactly the same way. Thus, just as any inherent tendency for (capitalist) profit rates to fall will depend upon capitalist structures (or other appropriate conditions) being sustained, so the disposition for, say, water to dissolve sugar will only be exercised where water exists. Both structures are currently present on planet Earth but (apparently) absent on planet Venus. Now it does appear to be the case that some natural mechanisms do not depend upon their time-space location. Newton’s laws, including the inverse square law of gravity, for example, are usually interpreted in this light; the same is true of electromagnetic laws and quantum mechanics. But even these examples can appear questionable once we turn to cosmology and explorations into the origins of the universe and the way in which it developed. From this perspective it seems at least feasible that such ‘laws change absolutely with time; that gravity for instance varies with time and that this inverse square law has a strength which depends on how long it is since the beginning of time’ (Feynman, 1988:206). In truth, in a (1989) postscript to his original work Bhaskar acknowledges that the suggested ontological limit to naturalism ‘only marks a necessary limit in relation to standard philosophy conceptions of physics and chemistry’ (Bhaskar 1979 [1989]:175). But a fundamental objective here, and most certainly of Bhaskar’s own contributions, is precisely to criticise and transcend these standard, if influential, conceptions. It is such conceptions that account for the widespread but erroneous impression that practices can be scientific only if they are concerned to uncover unconditional universalities. This insistence requires correction just as much as does the belief that scientific results necessarily take the form of constant event conjunctions. Science, it must now be recognised, is concerned to identify and understand structures and their mechanisms which govern actual events, but which themselves are only relatively enduring. To reiterate, the claim that the natural sciences are more successful than the social sciences requires careful interpretation. Certainly, some allowance must be made for the fact that the natural sciences, by and large (but not exclusively), concern themselves with improving their theories of a relative unchanging (or only slowly changing) reality while the social sciences concern themselves with understanding, and with continually improving their understanding, of a relatively fast changing, i.e. highly space-time specific, world. To the extent that certain strands of the natural sciences also concern themselves with a relatively fast changing reality, it is perhaps with the achievements of these strands that the successes of the social sciences can most easily be compared. In any case, any contrast drawn should be a considered one. I suspect, indeed, that when this is the case it will be found that social science has not fared quite so badly after all. © 1997 Tony Lawson
However that may be, it also follows that the faster nature, or greater space-time specificity, of social structures and mechanisms is itself sufficient for the hermeneutic moment to arise more frequently in social science. It is this, I think, which mainly explains Collier’s observation that the hermeneutic moment is always so prominent in social science. As structures and actions are continually transformed the social scientist will frequently need to reinvestigate what is going on, to keep abreast of the inherently non-predictable developments regularly taking place, including the transformations of human concepts. For this reason by itself the hermeneutic moment will usually be prominent in serious social research. But if this marks a distinguishing feature of social scientific practice, it no more follows by this token alone that social science is failing than it does that the possibility of naturalism is curtailed. (It does, though, mark a specificity of social science research that contemporary orthodox approaches rooted in positivistic injunctions appear to be totally unable to comprehend.) Now despite all such considerations, in the end it cannot be denied that, for the last fifty years or so especially, it is difficult to identify any obvious successes (explanatorily powerful, revelatory, hypotheses) of mainstream academic economics, let alone find results that can be held up to the achievements of the sciences of nature. Why then has mainstream academic economics in particular fared so badly? The explanation is precisely the mechanism that this book is attempting to counteract. That is, contemporary academic economists, for whatever reason, and no doubt under the influence of certain spectacular results of the Enlightenment, continue to labour under the apprehension that all scientific work is, or necessitates, or relates to, seeking out constant event conjunctions, that the goal just is deductivist modelling. In other words, I do not think that the failure of mainstream economics is a consequence of an unfathomable, overly complex and dynamic social reality, nor even of the infeasibility of meaningful experimental control. Rather, the continuing failure of the discipline must be put down to the often quite irrelevant, typically formalistic, methods and techniques which economists naively and unthinkingly wield in a forlorn hope of thereby gaining illumination of a social world that they do not ‘fit’. Of course, a failure to appreciate this, coupled with the corresponding persistent failure of the project to achieve any result of interest leads, in its turn, to the development of immunising strategies, ever increasing levels of technical complexity of models, the resort to computer simulations, and so forth. It leads, in short, to any response that is consistent with the misconceived, taken for granted, yet hardly fruitful, standard conception of science being sustained. The intention here is to indicate that all such efforts are likely to be beside the point, and how, despite everything, a successful social science of economics remains a viable option. For this latter possibility © 1997 Tony Lawson
to be realised, however, it is essential that economists abandon their preconceived (and upon examination clearly untenable) positivist conceptions of the structure of science and focus instead upon fashioning their methods to available insights bearing upon the nature of social reality. Through doing so, though, economists may yet find themselves explaining social phenomena in just the sense that natural scientists in fact, and successfully, explain the various phenomena of nature.
© 1997 Tony Lawson
16 ABSTRACTION
I have outlined broad objectives and modes of explanatory reasoning that are sponsored by the theory of social ontology sketched earlier. Although in so doing I have drawn some inferences about general procedures of enquiry that are often likely to be involved, there remains one procedure to which I have hardly referred explicitly, yet which is vital to any cognitive enterprise. I refer to the method or process of abstraction. Despite the limited explicit attention given to it in modern social theorising (including, curiously, the contributions of Bhaskar), abstraction is an indispensable method in science. Moreover, given the often found assertion that orthodox ‘economic modelling’ is itself based upon abstraction, it seems vital that I indicate why the procedure to which I refer does not at all reduce to the activities of the ‘modelling’ project in question. I interpret abstraction, here, according to its traditional meaning of focusing upon certain aspects of something to the (momentary) neglect of others. It is a process of focusing on some feature(s) of some thing(s) while others remain in the background. For example, in considering the ability of copper to conduct electricity well I may focus upon its atomic structure and thereby abstract from its colour, texture, malleability, and so on. It follows that there is always something which is abstracted from. Indeed, if this implication were always born in mind the mis-uses of the term that abound in mainstream economics (which we encounter below) might more frequently be avoided. That which is abstracted from is the concrete. The point of abstraction is to individuate one or more aspects, components, or attributes and their relationships in order to understand them better. Once this has been achieved it may be possible to combine or synthesise the various separate understandings into a unity that reconstitutes, or provides a better understanding of, the concrete. Thus, a comprehension of the various properties or aspects of copper may reveal its suitability for a particular technological application. Ultimately, the concrete can be understood as a synthesis.1 © 1997 Tony Lawson
ABSTRACTION AND CRITICAL REALISM It follows that abstraction, as interpreted here, can be both appropriately and inappropriately applied. There is nothing intrinsic to the method that determines, independently of considerations of the (type of) concrete object(s) of study, or the required focus, etc. that an abstraction is necessarily relevant or insightful. In particular, this understanding of abstraction as taking a particular focus or emphasis immediately raises the question of the appropriate vantage point, the level of generality, the space-time extension or scope involved. That is, the boundary setting and bringing into focus that characterises the process of abstraction simultaneously achieves a specific vantage point, a level of generality and a definite extension to any analysis. And it is important to recognise that the appropriate choice of these parameters cannot be determined independently of other considerations. Ultimately, the context of the analysis is crucial. Some things, though, can be said about the consequences which follow on acknowledging the critical realist perspective, despite its rather abstract nature. That is, it is possible to anticipate various definite and central tasks for abstraction associated with uncovering and understanding the powers, structures and tendencies that produce the actual course of events and states of affairs of the differentiated and open, natural and social worlds. In particular, abstraction will figure fundamentally both in the initial analysis of the phenomenon to be explained and in the attempt to illuminate the mechanisms that give rise to them. If some economic phenomenon of interest is a conjunction it can be resolved into its causal components, that is resolved into the different effects of various causal mechanisms. To focus upon one or a few such components is to abstract from the original phenomenon. The autumn leaf, for example, may be viewed under its aspect of moving to the ground, or under its aspect of ‘fluttering’ in the breeze, or even just in terms of its reddish colour. Economic structure too can be abstracted from by focusing upon certain powers or tendencies possessed amongst others. This is particularly clear if we focus upon particular capacities of the human individual. These claims can be expanded upon if we examine briefly how the theory of ontology systematised under the head of critical realism bears specific implications for the various aspects of determining a focus just noted, i.e. in determining an abstraction’s vantage point, level of generality, and scope (or extension).
THE VANTAGE POINT The selection of any specific phenomenon for explanation will necessarily reflect the vantage point of the enquirer not only because any individual’s © 1997 Tony Lawson
material and other interests depend upon the relationships in which he or she stands to others, but also in virtue of the fact that the possibility for direct experience of any individual is always highly relative. The identification of the particular phenomenon to be explained always depends upon the position, perspective and understanding of the viewer involved. Someone not trained in medical research, for example, may, on viewing an x-ray sheet, be absorbed in explaining the lighter and darker marks in terms of the human anatomy. A person with some medical training, in contrast, may immediately observe a mark or pattern that should not be there (in the sense that most people’s x-rays do not manifest this phenomenon), and as a result ponder upon the explanation. But if, say, an experienced surgeon observes the same x-ray sheet and ‘mark’, he or she may immediately ‘recognise’ the unwanted agent and wonder how it came to reside in the patient in question. For each person, the contrast, the phenomenon to be explained, and so the object of explanation, is different.2 To take a more familiar example: a first year economics undergraduate examining the results of running a statistical estimation programme may be mystified by all the diagnostics shown and wonder what explains them, whereas an experienced econometrician will be able to tell at a glance whether a preferred equation has performed well according to various conventional criteria. We have also seen that it is not merely the choice of phenomenon to be explained that reflects our knowledge, understandings, values and interests; in the end the latter bear upon the particular set of causal factors pursued as well. Of course, we could not explain the entire causal history of any phenomenon. I observed in the previous chapter that if the universe originated with the ‘big bang’ then this is presumably part of the causal history of any present day social phenomenon. Hence it is part of its explanation. At the same time, contemporary social life mostly depends on, and so is also in part explained by, the action of gravitational forces. Yet few social scientists make explicit reference to any such factors when offering an account of social phenomena. And we saw in Chapter 15 that the main reason why this does not matter is that explanation is made inherently contrastive. When we look to explain phenomenon x we essentially pose the question ‘why x rather than y?’. The objective is not the elaboration of the complete causal history of some phenomenon but the identification of at least one significant difference between the latter’s causal history and that of the chosen contrast. This feature of causal explanation seems particularly transparent in the social realm where the question of why some agency has performed differently from another frequently arises. And while it is the nature of the contrast that determines how far back or how far afield we must look to find a satisfactory explanation, the contrast chosen reflects our interests and understandings. © 1997 Tony Lawson
THE LEVEL OF GENERALITY The first part of the preceding sentence is fundamental, of course. Certainly the interests of investigators and other concerned parties are significant in the determination of which causal factors, in the event, are pursued. Thus, for example, a feminist economics is likely to be distinctive in facilitating an orientation towards factors hitherto unappreciated in the (predominantly male) academy (see Chapter 19). But the interests of social scientists do not usually constitute or shape those factors. Causal mechanisms that are productive of actual phenomena exist at their own level of being, independently, for the most part, of any investigation. It follows that if the choice and conception of the phenomenon to be explained as well as that of the contrast are always dependent upon the individual investigator or a specific community of researchers, the level of generality of abstractions appropriate to determining an adequate explanation (given the contrast) will depend upon the nature of the structures or mechanisms that are really responsible. In other words, abstraction must be put to work in identifying and comprehending that aspect, or set of aspects, of reality that is essential to the phenomenon (including the contrast) that we want to explain, rather than, say, in determining that aspect which is merely the most general. If, for example, we want to account for the phenomenon that goods are now usually produced for exchange in the market place rather than for immediate or eventual use by their producers, our explanation is unlikely to turn on characteristics peculiar to a specific firm. Indeed, production for exchange, i.e. commodity production, appears to be a characteristic feature of the system of capitalist production. In consequence, any meaningful analysis of the mechanisms and structures which condition this phenomenon will require abstractions at a high level of generality, and specifically at the level of features common to all forms of capitalist production. If, instead, the phenomenon to be explained is of the character that a specific firm’s prices are, say, 50 per cent below those of its competitors, or that its productivity growth rate is twice as high, the explanatory focus will necessarily be at a significantly lower level of generality, upon distinctive features of the local employer-employee relations, and so on. Similarly, the level of abstraction required to understand the high concentration of women in secondary sectors of employment will inevitably be higher than that required to explain how, in 1979, Margaret Thatcher became the prime minister of Britain. Now it is important to recognise that when an abstraction at a relatively low level of generality is made, any mechanisms already identified, or the features of reality brought into focus at a higher level of abstraction, cannot
© 1997 Tony Lawson
thereby be disregarded as inconsequential or otherwise irrelevant. Thus, any insight into mechanisms that are fundamental to capitalist production and which are identified at a high level of abstraction will, if correct, be just as relevant to analysis and understanding at the lower level of explaining more regionally and historically specific factors and contrasts. Indeed, it is precisely because the basic process of accumulation, competition, and so on, operates throughout the capitalist world that, typically, geo-historical peculiarities, or unevenness in development, are phenomena that are considered particularly interesting and warranting of explanation. To repeat, to focus upon highly context-specific features of reality is not to treat that which is out of focus, which operates at higher (or at different) levels of generality, as something that can thereby be assumed away. Something, to repeat, is always abstracted from.
THE SCOPE OR EXTENSION We can note, finally, that the additional dimension to abstraction distinguished above, namely its scope or extension, also varies according to the vantage point determined by the investigator’s knowledge and interests, as well as the level of generality. Broadly speaking, the higher the level of generality of any feature of interest the greater the extension to abstraction required to understand it. Thus, if an abstraction is to identify a set of mechanisms essential to capitalist production, its extension will inevitably encompass the space-time region over which the capitalist system of production has existed. If, instead, the focus is upon peculiarities of a particular firm’s performance relative to those of others in a comparable situation, then its space-time extension is likely to be significantly circumscribed in comparison. In the light of the theory of ontology systematised under the heading of critical realism it is especially apparent, then, that the list of features of abstraction which are crucial to social analysis includes not only the vantage point, but also the level of generality and the scope or extension. For if considerations bearing upon the level of generality are fundamental to identifying a mechanism operating at a different level to some phenomenon for which it is essential, considerations of extension are especially significant in the light of the now clearly recognised intrinsic-dynamism and internalrelationality of the material of the social realm. When an abstraction sets spacial boundaries to the focus taken, limits are thereby set on the interdependencies which, at any point in time, can be brought under consideration. And when the process of abstraction sets temporal boundaries to the focus taken, limits are thereby set on the histories of any particular aspects that can © 1997 Tony Lawson
be comprehended, including the past development of any feature as well as what it may yet become. Significant skill is therefore required in choosing the appropriate generality and scope of an abstraction when addressing any specific question. In order to illuminate a structure responsible for the production of some phenomenon of interest it is necessary to identify connections and relations essential both to that structure’s efficacy and to its existence and mode of reproduction. A comprehension of any (set of) structure(s) will entail identifying the nature of its internal relatedness as well as its particular history.3 Notice that this conception does not undermine the possibility of features or aspects of an internally-related set of structures or dynamic process being considered explicitly and individually at a moment in time. The implication, rather, is that such features cannot be considered as isolated momentary phenomena, like punctiform events. Each must be seen as expressing the remaining features, or whole, to which it is related, as well as (or including) its own history. When, for example, we focus upon a particular footballer or hockey player setting off down one side of the playing field with the ball just in front of him or her, we do not suppose that other players, who may be momentarily out of view, cease to exist. Indeed, we interpret the objectives or tactics of the player in view conditional upon our understandings and expectations of others, the rules of the games, the player’s own ambitions, competencies, confidence and history, as well as upon our assessment of his or her understanding of the tactics and competencies of others, and so on.
ABSTRACTION AND GENERALISATION Let me re-emphasise that the just discussed insights on the central role for abstraction are facilitated by the understanding which critical realism provides. I do not want to imply, of course, that abstraction is other than essential, whatever the perspective on science that is provided. Different perspectives, though, do encourage competing conceptions of how abstraction is most centrally deployed. In particular, for those guided by the positivistic image of science as the elaboration of regularities between events or states of affairs, the goal of abstraction inevitably becomes that of seeking formal relations of similarity rather than uncovering the essential. As Dobb (1972 [1937]), distancing himself from this perspective, says of it, the aim is to base abstraction, not on any evidence of fact as to what features in a situation are essential and what are inessential, but simply on the formal procedure of combining the properties common to a © 1997 Tony Lawson
heterogeneous assortment of situations and building abstraction out of analogy. (Dobb, 1972 [1937]:40) Of course, generality of some kind is also an objective acknowledged here— to determine non-empirical features at the ‘deeper’ level of necessary relations and tendencies of things. These, however, are unlikely to be uncovered just by adopting the a priori aim of seeking broad generalisations. As Dobb concludes, what the positivistic form of ‘abstraction gains in breadth it more than loses, as it were, in depth—in relevance to the particular situations which are the focus of interest’ (ibid.). It is a characteristic of economic structures, for example, that much of what is essential to the explaining of a form of human activity is highly context-related. Thus, through a process of seeking merely wide generalisations, economic structures can easily be emptied of their contextrelated, but often essential, content. The employer-worker relationship may be identified as a widespread feature of human society. But this observation abstracts from the numerous variations in the nature of this relationship across time and space, and it is certainly insufficient for an understanding of the various work practices and activities that exist at a specific stage of human evolution in any particular region or place. Similarly, to note such generalities as people have preferences, or beliefs, or the ability to make choices, is to provide insufficient detail of these generative structures or causal powers for explaining any concrete form of human activity they govern.4 Perhaps the greatest danger of merely seeking broad generalisations, of pushing the abstractions so far that they are almost devoid of substantive content, arises from the fact that they then need to be supplemented by other propositions in order to have any analytic value; for this opens the door to the inclusion of ‘highly artificial’ or ‘bogus abstractions’, or, more accurately, convenient fictions. In other words, almost contentless abstractions can easily, if unwittingly, be manipulated or ‘strengthened’ in illegitimate ways to yield conceptions that really are no longer abstractions at all. In this, the strengthening additions may (correctly) be interpreted by those who formulate them as other than abstractions—typically as assumptions. But this does not render them, or the exercises conditioned by them, as thereby somehow harmless or neutral. In mainstream economics, for example, such ‘assumptions’, which may even creep in unnoticed, are usually designed to achieve mathematical tractability, system closure and completeness, or some such thing, rather than an understanding of the real causal mechanisms at work. Inevitably, if the original abstractions alone possess little explanatory content, then it is the additional ‘strengthening’ assumptions that do all the work and determine the upshot. And a failure to appreciate this may lead © 1997 Tony Lawson
either to an unthinking and erroneous belief that these assumptions can eventually be replaced with accounts of essential aspects of real generative mechanisms without the whole construction collapsing entirely, or to a misguided attempt to extract more meaning from the constructions than can possibly be legitimate. Once more this appears to be a point that Dobb has already emphasised: There is the danger of introducing, unnoticed, purely imaginary or even contradictory assumptions and in general of ignoring how limited a meaning the corollaries deducible from these abstract propositions must have and the qualifications which the presence of other concrete factors (which may be the major influences in this or that particular situation) may introduce. All too frequently the propositions which are products of this mode of abstraction have little more than formal meaning and at most tell one that an expression for such-and-such a relation must find a place in any of one’s equational systems. But those who use such propositions and build corollaries upon them are seldom mindful of this limitation and in applying them as ‘laws’ of the real world invariably extract from them more meaning than their emptiness of real content can possibly hold. (Dobb, 1972 [1937]:41)
ABSTRACTION AND ECONOMIC ‘MODELLING’ If those who accept the familiar positivist results inevitably set off in the wrong direction when making abstractions (seeking out merely broad generalisations rather than attempting to identify what is essential to some phenomenon of interest) there is a second greater error characteristic of contemporary economics which turns on a misunderstanding or misconstrual of what abstraction is. For in modern mainstream economics abstraction is interpreted not only, and not even mainly, as the legitimate activity of leaving momentarily out of focus something that is real. Rather, any explicit reference to the term is taken almost exclusively to denote the (typically illegitimate) activity, or result, of excluding something real, of assuming it away entirely. In place of abstraction as a one-sided focus upon an aspect of a concrete entity, an aspect brought momentarily into closer view, economic modelling thus interprets abstraction as a focus upon the aspect in question as though it existed in isolation—and typically as though it were free of internal instability as well. In other words, in mainstream economics the term abstraction stands in as rhetoric for the pretence that economic phenomena are, after all, generated © 1997 Tony Lawson
under conditions equivalent to those achieved though experimental control. Economies are ‘modelled’ as closed in the sense that the rest of the world does not exist, uncertainty is all but banished, as are becomings, and ‘begoings’, mortalities and (systematic) mistakes, conflicts and crises, internal relations and transformations. In the name of abstraction all features of social reality that prove inconvenient to deductivist modes of reasoning are ultimately assumed away. Those features which are not consistent with the two overriding closure conditions of the methods of contemporary modelling—that the material under discussion be viewed as atomistic-like and isolated—are, under the heading of abstraction, never allowed in the frame. It is as though when someone goes to a concert, say, and focuses momentarily upon one particular instrument, the rest of the orchestra does not exist; or that, when we focus upon the team player discussed above, the opposing players or supporting mid-fielders are no longer on the playing field. In short, the notion of abstraction is employed merely to gloss the already encountered fallacies of atomism and isolationism. Notice that such fictionalising cannot, in general, be legitimately defended as a heuristic device, as a useful first approximation or step. In Chapter 9 we came across the ‘method of successive approximation’, a procedure advocated by Musgrave (1981), supported by Nowak (1980) and examined by Mäki (1992a), amongst others. This is a step-wise procedure wherein one feature of reality is first looked at as though it existed in isolation, and then, when understood on those terms, combined with a second aspect, and then a third, and so on. Now it can readily be admitted that such a procedure may be valid if the material under analysis is of a nature such that it can really be decomposed into atomistic, isolated, components, whose effects can be mechanically added together or otherwise combined in some predictable fashion. The problem, however, is that, contra both Mill and recently Cartwright,5 but as Keynes clearly recognised, social material does not usually conform to this requirement at all. It is as though, to return to our football or hockey game, the team player with the ball can be treated as if in isolation at least as a first step or approximation. Or it is as if the coach of the team in question determines the tactics of the game by starting from the assumption that the noted player will be the only one on the pitch. Even this restriction is too lax, in fact, for the player may run with the ball at the opponents’ goal, or kick/hit it straight there from a distance. In other words, the player in our conception is not yet sufficiently atomistic. The intrinsic closure conditions must be shorn up, all but one of the player’s possible internal states must be assumed away or closed off. As with everything else intrinsic complexity can be added back in, it must be supposed, as part of the step-wise procedure that constitutes the method of successive approximation. © 1997 Tony Lawson
Such a procedure is patently absurd as a generalised tool for social science.6 Instead, continuing the sports example, the legitimate strategy can only be to cycle in and out of different frames of focus, sometimes ‘standing back’ and ‘seeing’ the movement on the pitch as a whole, sometimes focusing on individual aspects. On occasion, if to repeat, a player will be interpreted from the perspective of his or her relationship to the goal, sometimes from the perspective of his or her history (what are thought to be his or her skills, competencies and strengths, etc.), or from the perspective of the history of the game itself (e.g. in terms of a recognised trend for referees to penalise certain offences). And if this cycling in and out, this continual changing of focus and perspective (depending upon which phenomenon at any point we want to understand) is the only way to comprehend a football or hockey match (and, of course, in a televised match this cycling or zooming in and out on specific aspects of the play is literally what happens with the camera) it is also the only way of gaining an understanding of society and economy. When we focus upon varying productivity performances here, conditions of work there, rising or falling unemployment rates, and so on, we do not suppose that those features we choose to emphasise exist in isolation, even as a temporary, heuristic, measure. To do so is to assume a totally different world from the one in which we live, and one that has no bearing upon it. Indeed, the fiction of atomistic, quasi-omniscient, infallible (‘economically rational’) agents acting in the closed, isolated conditions described in contemporary modelling, does indeed constitute such a different world, one hopelessly irrelevant for providing insight into our own. In short, there is literally a world of difference between leaving something (temporarily) out of focus and treating it as though it does not exist. The achieving of an abstraction and treating something as though it existed in isolation are not the same thing at all. Abstraction, meaning looking at something in a ‘one-sided’ manner, is indispensable in science. Its object is to individuate some component or aspect of a concrete entity in order better to understand the latter. And it is essential to recognise that this entails understanding the aspect in question within the relationships in which it stands, relationships which may be essential to its existence and/or mode of activity. The purpose of abstraction is not to mask (or legitimate) a pretence that the aspect in question exists in isolation. Specifically, where—as in social science—meaningful, well controlled, experimentation appears impossible, abstraction is not usefully employed to denote (and usage of the term certainly does not render legitimate) ‘modelling’, which assumes that conditions analogous to those engineered through well controlled experiments have occurred after all. Indeed, the procedure of abstraction allows science to proceed in the absence © 1997 Tony Lawson
of such occurrences. In particular, abstraction, when skilfully executed, can, amongst other things, enable us to access and understand a structured, dynamic and holistic reality. In other words, abstraction, though not a licence for the largely irrelevant modelling activities of mainstream economists, is a procedure that can facilitate the illumination of the open social world in which we live.
© 1997 Tony Lawson
17 ON TRUTH IN ECONOMICS
Before considering an illustrative example of economic explanation consistent with the perspective I have been defending I ought to consider the status of any results that may be achieved. I have acknowledged throughout (and in Chapter 5 at length) that our theories are fallible and historically transient. I want now to dispel any impression this may give that the possibility of objective truth in knowledge is thereby precluded. This seems important. For a belief in the impossibility of true or even realistic theories could itself encourage the idea that acknowledged fictions of the sort which prevail in contemporary economics are sustainable after all.
THE NATURE OF TRUTH We can observe straightaway that many economists, including econometricians, do explicitly conclude that true theories or models are unobtainable in practice, and perhaps also in principle.1 It seems likely that in reaching such conclusions these economists are in effect holding, if implicitly, to conceptions of knowledge and truth as relations of identity or, more generally, correspondence.2 Certainly the widespread use of the terminology of economic models is consistent with this—as if the latter are somehow smaller, scaled down, or otherwise simplified, versions or mappings of the original. Economists frequently talk of ‘mappings’ in connection with their modelling, often remarking that any map drawn to a scale of one-to-one would be as complicated as whatever it is supposed to represent, and thereby useless. From this perspective it follows that because of the complex nature of social reality any feasible mapping is never isomorphic. In consequence, the argument seems to run, all models are inevitably distortions; true models or theories are most unlikely to be obtained. Whether or not any economist holds to this line of reasoning completely (or explicitly), or would express things quite so starkly, it is worth considering this position here anyway as an obvious point of reference. And it should be © 1997 Tony Lawson
apparent from everything that has gone before that any conception of knowledge and truth along the lines indicated, i.e. as relations of correspondence between discourse and extra-discursive reality, is profoundly misleading. In Part I it was seen how the thesis of the (existential) intransitivity of objects or ontological realism (that things in general exist and act independently of our knowledge of them) must be tied to the thesis of the historical transitivity of knowledge or epistemological relativism (that there is no way to know what exists except under particular, historically transient, descriptions). For example, the pervasive phenomenon of scientific continuity and change is intelligible only if we recognise this distinction between intransitive and transitive objects of knowledge. For only then can we make sense of scientific discovery and change in our knowledge of the world on the one hand, and changes in the world itself, on the other. That is, unless knowledge and its intransitive objects possess relatively distinct beings and histories, scientific change and changes in the nature of things could not be distinguished, with the consequence that the former would be unintelligible while the latter would never be noted. Put differently, the intelligibility of the phenomenon of scientific continuity and change entails not only that creative subjects do not constitute the world (as in idealism including its post-modernist or pragmatist variants) but also that thought cannot be regarded as a mechanical function of the world (as in empiricism). Rather, knowledge must be seen as consisting of stuff that is socially produced, typically symbolically mediated and formed, historically specific, and materially irreducible to its (intransitive) objects. Both inside and outside of science, when we talk or write of things, structures, events and states of affairs, we can speak or write of, and know, them only under particular descriptions, under descriptions that are neither reducible to, nor isomorphic with, the objects which (I am arguing) they nevertheless express.
OBJECTIVE TRUTH To accept the fallibility and transformability of knowledge, however, is not thereby to deny the possibility of objective truth. Rather, propositions are true, or contain truth, by virtue of the way that the world is. There is no need or possibility here of rejecting that implicit premise of the correspondence theory that truth can, under at least some of its numerous aspects,3 figure as an objective property of propositions. Here, the intuitions of many econometricians can be accepted. For it seems clear that when they reject either the possibility, or the actuality, of truth in formulating their models, especially when they declare that these models are not yet true ones, truth is being interpreted as something objective. It is being acknowledged as © 1997 Tony Lawson
something which holds or exists independently of any individual’s beliefs, and presumably, given the often found emphasis upon truth being unobtainable because of the complexity of reality, which turns upon the way that the world is.4 This implicit recognition of an objective aspect to truth is something I clearly retain. But in place of the conception of correspondence (or mapping) in knowledge and truth, the notion of expression is preferable. This term, as much as any, has connotations which capture that characteristic of able. It reminds us that there is no necessary relation of identity or able. It reminds us that there is no necessary relation of identity or correspondence involving knowledge. It indicates that what is expressed, the referent of any expression, can never be reduced to, or mapped onto, its expression—just as the expression can never be reduced to, or mapped onto, that which is expressed. But this latter situation, if to repeat, does not undermine the possibility of prepositional truth. Indeed, the statement that ‘contemporary western society is relatively complex’, which in effect appears to be a premise for the familiar (if erroneous) claim that economic theorising must be unrealistic or false, may even be absolutely true—by virtue of the way that the world is. Clearly, though, whether or not this is so, there is no question of a mapping or correspondence involved. To acknowledge this aspect of our knowledge is not to deny that economists, like everyone else, must take a one-sided view of their objects. But in so taking a one-sided view, in so abstracting from the concrete, there is no inevitable falsity about the result. In particular, and to repeat an assessment of the previous chapter, to focus upon particular features of something, for example a social system or a human being, is not per se to treat those features as existing in isolation or otherwise subject to any necessary distortion. If I focus on a person’s eyes in an attempt to gauge his or her reaction to what I am saying, or if I describe them in reporting my impression to others, I do not suppose that they exist in isolation; nor do I otherwise necessarily misrepresent the person’s reaction in any way. (Of course, as I emphasised in Chapter 9, there may not even be any distortion involved in expressing something as an idealisation. It depends on the referent of that expression. Idealisations can be real.) To take a partial approach is not per se to deform. It can involve a distortion; but the one is not equivalent to the other.
TRUTH AS AN EXPRESSIVE-REFERENTIAL DUAL A complication for the account I am defending is that in everyday discourse there is an inherent bipolarity in the way the concept of truth, as least in its objective aspect under consideration here, is employed. Specifically, in © 1997 Tony Lawson
everyday conversation the term truth functions both as an attribute of propositions and as a feature of the reality expressed. When we talk, for example, of seeking true statements, including on occasion ‘nothing but the truth’, it appears to be the former usage that is in question. When, in contrast, we talk of seeking (to know) the ‘whole truth’ it is the second usage or component that is at issue. While the former usage relates primarily to descriptions or expressions of events or states of affairs, etc., the latter usage (if it is not to be quite meaningless) relates to the referents of any such expressions, to a set of (aspects of) states of affairs that are, or can be, expressed and known. Now this ambiguity, if such it is, is not restricted to our conception of truth. It arises more generally with the non-identity of the transitive and intransitive objects of knowledge. Specifically, certain terms are widely used to denote, simultaneously, both transitive and intransitive objects—allowing the possible misapprehension that they are one and the same thing. For example, the term ‘scientific law’ in philosophical and scientific writing sometimes denotes a mechanism and its way of acting (or whatever), and sometimes an expression of it. The same duality is inherent in our usage of the term ‘explanation’ when there is some phenomenon to account for. Similarly, the term ‘history’ is applied both to the becoming and ‘begoing’ in time of some structure or individual, as well as to some account of its becoming and ‘begoing’. This ambiguous or double usage also applies even to terms such as ‘abstraction’ or ‘concrete’. Marx took pains to distinguish the ‘real concrete’ and the ‘thought concrete’ (see Ollman, 1993). Both are real, but this differentiation helpfully reminds us how the same term is often applied across the transitive/intransitive divide. The upshot is that an inherent bipolarity must be recognised in our usage of all such concepts, so that their meaning can usually only be determined from context. Put differently, any such term (and here we are following Bhaskar once more) can best be seen as an epistemic-ontic dual, as an essentially expressive-referential notion. Truth, in its objective aspect here under consideration takes the form of a duality. This duality of truth turns on the condition that, while the referents of discourse, i.e. the conditions and states of affairs which form the intransitive objects of knowledge, exist (relatively) independently of us, they can only be expressed by us via human mediation in cognitive discourse. In its epistemic aspect, truth functions to designate the truth-values, truth-judgements and truth-claims which are, or could be, made in scientific activity and discourse; in its ontic aspect truth functions to designate the states of affairs expressed, the referents (of the claims, judgements, propositions, etc.) which exist relatively independently of us and by virtue of which our claims are (or are not) true.
© 1997 Tony Lawson
CRITERIA OF TRUTH The preceding outline mainly elaborates what is meant by the claim that our theories can be true, or can contain truth. And according to the conception defended the real possibility of truth in economic statements or theories can be accepted. But, of course, a recognition of this possibility does not guarantee that even our currently most well grounded claims are without error. Certainly there is no external point, no ‘God’s eye view’ or ‘Archimedean point’, for determining the truth of any proposition. Propositions, or sentences, cannot be compared to states of affairs; as already noted there is no third thing or relationship called correspondence. The judgement of the truth of a proposition, the criteria for deciding which propositions are to be held as (possibly) true, are necessarily intrinsic to the science (or social activity) concerned. There is no way of comparing the world and a sentence to see whether they match. The fallibility of knowledge must be accepted. It can also readily be acknowledged that the criteria actually employed within (and outside of) science may not all be ontological/evidential, but will often turn on considerations of a pragmatist (persuasiveness, simplicity, standard of rhetoric, vested interests) or coherentist (consistency with existing general beliefs, or with those of some authority) sort. Of course, which pragmatic features are regarded as virtuous will depend significantly upon the prevailing social, historical, cultural and political context. This is something which feminist economists, for example, have recently emphasised. In the economics academy, as in many walks of life, the criteria of persuasiveness and other pragmatic virtues have been laid down by a body constituting a quite unrepresentative cross-section of the community when considered on the basis of gender, race, class, ethnicity and so forth. We can thus agree with Strassman (1994:156) that: ‘As long as methods of argument, training, and socialization remain differentially compelling to scholars with differently positioned lives, bodies, and experiences, then demographics matter in theory choice.’5 The sort of research which is carried out reflects something of the situated natures of the researches. Similarly, the standards of evaluation depend upon the situated natures of the evaluators.6 In short, not only is knowledge fallible and always potentially transformable, but the claims made and held in science are elaborated on the basis of various criteria, doubtless including many that have little to do with the determination of their ontological grounding.
© 1997 Tony Lawson
JUDGEMENTAL RATIONALITY It is a consequence of this realisation that a significant challenge facing any investigatory community lies in recognising its vulnerability (as well as indebtedness) to the interests, values and perspectives of its members. But it does not follow thereby that all claims in science are equally good, or, specifically, that knowledge held as possibly true cannot in fact be true. Strassman (1994:157) is correct to conclude that ‘accepting the impossibility of a sorting mechanism that is independent of time, history, gender, race, class, and culture is not the same thing as saying all accounts are equally valid’. The thesis of epistemic relativity of knowledge accepted here (that there is no way of knowing what exists except under particular, historically transient, interest-guided descriptions) does not necessitate the thesis of judgemental relativism, that any theory is as good as any other. Rather epistemic relativism when conjoined, as it is here, to ontological realism is a necessary condition for, and obvious complement to, a judgemental rationality. Although the criteria for accepting knowledge claims as potentially true may include pragmatist and coherentist ones, if amongst others, the commitment to realism entails that such criteria must also, and primarily, turn on the power of some theoretical claim to illuminate a (relatively) mind independent reality. Truth judgements, on a realist view, will incorporate descriptive and evidential components alongside any pragmatist and coherentist considerations. Such judgements will necessarily be ontologically grounded or justified, thus facilitating discovery and development in science as a rational, discriminating, and adequating process. According to the perspective defended here specifically, theoretical claims are scientifically-ontologically grounded according to the two basic models or schema of explanation that have above been found to receive philosophical support. Let me briefly restate them. The first mode of explanation, referred to as pure (or theoretical or abstract) explanation, proceeds in a few basic steps. First, a (rough) regularity in some phenomenon is described; second, some (explanatory) mechanism is retroduced, the theoretical construction follows of a hypothetical mechanism which, if it existed and acted in the manner supposed, would be capable of producing the phenomenon described; third, competing explanations are elaborated and some are eliminated on the grounds of their (inferior) empirical adequacy; so that, fourth, the identification of the causal mechanism at work is (hopefully) achieved, whereupon the latter now becomes the phenomenon to be explained. The second mode of explanation, referred to as applied (or practical or concrete) explanation, a form which is essential when conditions are fundamentally open, proceeds in a manner that is somewhat different. First, a complex event or situation of interest is resolved into its separate components, © 1997 Tony Lawson
is resolved into the effects of its separate determinants; second, these components are then redescribed in theoretically significant terms; third, a knowledge of independently validated tendency statements is utilised in the retrodiction of possible antecedent conditions; whereupon, fourth, competing accounts of possible causes are selected amongst on evidential grounds, i.e. empirically. In short, in both pure and applied explanation, competing theories are to be assessed empirically on the basis of their explanatory power (which may reduce to predictive power in those conditions, such as experimental control, where one (set of) mechanism(s) can be materially insulated). In other words, scientific theory can be ontologically-grounded in line with the perspective so far elaborated. In such fashion, theories can be held to be justified, as grounded, as (possibly) true, despite the inevitability of their possessing a relative aspect. It is precisely because scientific theories are both relative (historically transient and fallible) as well as (or at least where) suitably ontologically justified or grounded that I have argued, elsewhere, for the conception of a relative-absolute nature of knowledge and so of (judgements) of truth (see Lawson, 1987). Certainly, a recognition that all knowledge is relative does not entail the consequence that any degree of unrealisticness in economic modelling thereby goes (see also Sofianou, 1995).
FALSITY AND SHEER ERROR An implication of this conception is that if a theory once held as (possibly) true is eventually found to be inadequate in some sense, that it is outperformed in terms of explanatory power by some alternative conception, it is not inevitable that the initial theory be rejected as sheer error. I have emphasised above that usage of the term expression—or more comprehensively, the ascription of the label ‘expressive-referential’—to characterise the nature of the theory of truth here supported, is, in part, to remind us that neither our theoretical constructions nor their referents can be reduced to, or identified with, the other—that the two do not literally correspond. But the term expression also indicates that there are different, including better and worse, ways of capturing something. In consequence, there can be no a priori supposition that a given proposition will be absolutely true or absolutely false. It may be so. I have already mentioned above that a proposition of the sort ‘contemporary western society is relatively complex’ is conceivably absolutely true—depending on the nature of modern western society (and of course what we mean by ‘relatively complex’). But propositions such as ‘the sun sets in the evening’, while perhaps still held as true by some, but considered no longer to be the most adequate account by modern science, is © 1997 Tony Lawson
unlikely, nevertheless, to be sheer error. This proposition, which is certainly well grounded, can still contain truth by virtue of the way that the state of affairs expressed really is. Thus, whilst the truth of any given proposition is fixed or objective, truth in its (objective) aspect here under discussion nevertheless comes in degrees. Some expressions of a given aspect of reality can be truer than others, by virtue of the nature of the referent that is (varyingly) expressed. It does not follow, then, that the inevitable fallibility of knowledge is inconsistent with the possibility of specific theoretical claims (and in particular ones that have been systematically and extensively assessed empirically) being true, or possessing truth. Nor does it follow that when knowledge is transformed in the light of new evidence, original formulations must be rejected as sheer error and no better than any other falsity. Nor need we even lose interest in them. Indeed, when a theory once held as possibly true is superseded by some more explanatorily powerful alternative, it can be revealing to explain why or how the initial formulation performs as well as it does (or performed as well as it did)—to explain it from the perspective of recent, more adequate, conceptions. In so doing we may also want to learn its continuing domain of useful application, if any. This is precisely what happens when we interpret the continuing usefulness of the claim that the sun sets each evening. We also know the limitations of this conception from the standpoint of a more explanatorily powerful account. Now it is possible that the continuing usefulness of (aspects of) supplanted theories is a feature that is more characteristic of natural than social science. After all social structures depend upon human action and conceptions and, in consequence, can change with corrections to them. But it all depends. In any case, from the perspective here defended, any explanation of the pre-existing state of affairs, rendering it intelligible, investigating why certain views were held or proved as successful as they did, clearly has the potential to be illuminating. Previous successes, in short, can usefully be situated, their conditions of acceptance and successes comprehended.
TRUTH AND ECONOMIC ‘MODELLING’ This process of situating the conditions of successes of theories which were formerly, but are no longer, held as the most explanatorily powerful, is precisely what is achieved by transcendental realism with respect to the image of science that stems from the perspective of positivism. The latter’s conditions of success, outside astronomy, are mainly those of experimental control— where one or a set of intrinsically stable mechanisms is insulated and thereby empirically identified. In identifying the conditions of the successes of this © 1997 Tony Lawson
conception of science, its limitations are simultaneously revealed. And one most noteworthy feature of this achievement is that the positivistic conception of scientific analysis is found to be hardly appropriate to the social realm at all. Thus theories or conceptions contrived on the premise of widespread social closure are unlikely to bear much relevance to many aspects of the real social world. Certainly, the frequently postulated hypotheses of isolated economics, equilibrium states, omniscient agents, two commodity worlds, and so on, are not theories that were once found to be empirically powerful before being superseded by theories shown to be empirically more adequate still. They never were other than obvious fictions—contrivances, indeed, that, in the main, are immediately recognised as quite absurd. It is the image of science underpinning them, the deductivist structure of explanation which they presuppose, that has obtained a degree of success, but usually in conditions unavailable to social science as we have repeatedly seen. In short, the fallibility of knowledge neither rules out the possibility of truth in knowledge—and certainly does nothing to devalue its pursuit—nor does it license the absurd fictions of academic economics, sometimes construed as if more or less anything goes. Rather, the acceptance of ontological realism places the constraint of extra-discursive reality upon what can legitimately be maintained. A commitment to (ontological) realism combined with epistemological relativism sustains a judgemental rationality, and from this perspective the practices of contemporary economic modelling are found to be wanting. To pursue the latter in the name of abstraction, or on the basis of the intrinsic complexity of social reality or the recognition that all knowledge is fallible and historically transient, or some such, is merely to compound the confusion. The distance of the whole modelling project in economics from anything remotely realistic entails that, even accepting the foregoing comments upon the relativity of knowledge, this is one conception which may have to be rejected in its entirety after all, as having always been a manifestation of more or less sheer falsehood.
© 1997 Tony Lawson
18 ILLUSTRATION
It is time to be more concrete. For completeness I should at least indicate, and pass some comment upon, substantive economic theories, projects or analyses which seem broadly consistent with critical realism.1 Specifically, in the light of the emphasis taken in the preceding chapters it seems warranted that I include some reference to empirically grounded explanations of social phenomena which are couched in terms of structures and mechanisms that are reproduced over time, where that reproduction is itself explained. There are numerous examples I could focus on. At a general level it is easy enough to point to Hayek’s theory of spontaneous order; Smith on the division of labour; Keynes’ explanation of investment volatility in terms of the rise of stock markets, liquidity, and its facilitation of significant speculative activity; Marx’s analysis of capitalism, including the reproduction of structural conditions which give rise to various transfactual tendencies (such as for the rate of profit to fall), and so forth.2 I have argued elsewhere that much that is produced in contemporary institutionalist,3 post-Keynesian, Austrian and Marxian traditions is broadly consistent with this approach (see Lawson, 1994b). I also find significant overlap with much that is currently emerging under the head of feminist economics (see, for example, the contributions in the journal so titled; the papers collected in Gender And Economics edited by Humphries (1995); Folbre’s recent (1994) analysis of the sexual division of labour including the manner of reproduction of structures of societal caring; or Nelson’s (1996) study of the gender of the discipline of economics). Other recent output which adopts a broadly similar approach includes a body of work systematised under the heading of evolutionary theory (on this see Hodgson, 1993a), and also a ‘path-dependence’ literature (associated most closely with the writings of David, 1985, 1986, 1992, 1993, 1994), explicitly oriented to demonstrating generally that ‘history matters’ and specifically how certain technologies, conventions and other social forms come to be continually reproduced or ‘locked-in’. Any of these would provide a useful focus here. However, it is the latter path-dependence approach or project that I shall be most concerned with, at © 1997 Tony Lawson
least initially. Given that this ongoing, often explicitly methodological project appears to be the one that most straightforwardly overlaps with my own at the level of suggesting how social science might actually proceed, this seems the prima facie most sensible focus to take. In fact this adopted focus is doubly useful. Certainly it helps illustrate aspects of social reproduction that I take to be fundamental. But I do have some reservations with aspects of the pathdependence project, at least as it is sometimes interpreted. And by considering these in due course I am afforded the opportunity to re-emphasise and illustrate an aspect of the explanatory framework developed in the previous chapters that I take to be particularly significant. As a final remark at this stage, I should emphasise that my purpose in the remainder of this chapter is not to consider the empirical support for the theories considered. Any such consideration would be unlikely to provide insights of any generality. As I have stressed already (see Chapter 14), the high level of abstraction of the perspective systematised as critical realism facilitates inferences for explanatory procedure that are inevitably also fairly general in nature. Such references to empirical assessments by others will be made at relevant points. But my aim here is primarily to illustrate the sorts of steps that will be necessitated, and types of considerations that are likely to arise and be of significance, once a framework along the lines supported in the previous few chapters is accepted and acted upon.
PATH-DEPENDENCE The study of ‘path-dependence’, and in particular of the phenomenon of ‘lockin’, which seems to be best known to economists, has already been referred to in Chapter 15.4 It is provided by David’s (1985, 1986) account of an aspect of a familiar piece of technology,5 the letter arrangement of the computer keyboard that so many now regularly use.6 It is the case that newly acquired typing skills, typing manuals, classes in typing, and so forth, are oriented towards keyboards with a particular arrangement of letters, one in which the topmost row spells QWERTYUIOP. How is the pervasiveness of this particular letter arrangement to be explained? It is clear that personal computers available for purchase, and currently supplied in offices and other places, mostly contain keyboards with this arrangement of letters. Perhaps this is a sufficient explanation for a would-be typist looking to invest in some relevant training. But for the social theorist this explanation is hardly revealing. It is no more enlightening than, say, explaining the phenomenon that many people in the UK leave pubs just after 11.10 p.m. by the insight that this is the time pubs close. The interesting issue in such cases is to explain the explanation. In the case of the QWERTYUIOP keyboard, specifically, the obvious questions © 1997 Tony Lawson
to pursue are (1) why this particular keyboard letter arrangement came about in the first place (rather than something else), and (2) how it (rather than something else) came to dominate, or became apparently ‘locked-in’. It is such a two-part explanation that David basically pursues. Before briefly examining the nature of David’s explanation let me recall, in passing, how this example immediately reminds us that all aspects of explanatory endeavours are inherently interest and knowledge conditioned and motivated by conflict, surprise and/or doubt. I have already suggested in Chapter 15 that if the letter arrangement had been ABCDEFGHIJ then many of us would not regard the phenomenon as one that is at all surprising and in particular need of being explained. Of course, even if such an unpredictable letter arrangement as QWERTYUIOP is of prima facie explanatory interest to an ‘economic historian’ such as Paul David, to the unsuspecting mainstream economist a ready explanation is always at hand. For the latter’s acceptance of the deductivist framework usually (if not inevitably, as we saw in Chapter 8) encourages the a priori supposition that all social outcomes are the simple result of straightforward optimising decisions. From this perspective, it is to be anticipated that if the QWERTYUIOP arrangement has emerged and persists, this must be because it facilitates maximum typing speeds, or some such. Yet, although this sort of consideration is a factor in the story, there is evidence that things are not that simple. In fact, the (1991) Cambridge Encyclopedia of Language notes that the: [QWERTYUIOP] design has attracted generations of criticism, on ergonomic grounds. Although most typists are right-handed, this keyboard makes the left do 56 per cent of the work. Of all movements for successive letters, 48 per cent use only one hand instead of two— most noticeable when typing such words as addressed. Finger dexterity is not linked to letter frequency—for example, the two strongest fingers of the right hand are used for two of the least frequent letters, j and k. (1991:192) It is in recognition that this keyboard arrangement is widely regarded as suboptimal that David focuses on it in his attempt to persuade economists to rely less on a priori presumptions and to take more account of history. In pursuing the first part of the explanation, i.e. why this particular keyboard letter arrangement came about in the first place, David traces the origin of the noted arrangement of letters to a chance solution to a particular design problem that arose with one of the many early attempts to construct a reliable typewriter. Specifically, in the late 1860s a Christopher Sholes of Milwaukee in Wisconsin invented a version of the typewriter that was found to have a particular defect. There was a tendency for the typebars to clash and jam if struck in rapid succession. On ‘up-stroke’ machines such as Sholes’ © 1997 Tony Lawson
this problem was exacerbated by the fact that the printing point was located underneath the paper-carriage and so was not visible to the operator. For when a typebar would stick at or near the printing point, so that every succeeding stroke would merely serve to hammer the same impression onto the paper, the resulting string of repeated letters would be discovered only at the end of the paragraph or whenever the typist raised the carriage to examine what had been printed. Moreover, unsticking jammed typebars was an awkward and time-consuming manoeuvre. In consequence Sholes spent the next six years attempting to remedy this defect. David notes the outcome: It was during this painful interval that a four-row, upper case keyboard approaching the modern QWERTY standard emerged, from the inventor’s trial-and-error rearrangement of the original model’s alphabetical key ordering in an effort to reduce the frequency of typebar clashes. Vestiges of the primordial layout remained, as they do to this day in the ‘home row’ sequence: FGHJKL, with ‘I’ close by in the second row... (David, 1986:35–6) In pursuing the second part of the explanation, i.e. how the QWERTY arrangement became locked-in, David makes it clear that there was no inevitability about this. The US economic downturn of the 1880s was not a good time to be introducing a novel piece of office equipment. Moreover the beginning of the typewriter boom in the 1880s had...witnessed a rapid proliferation of competitive designs, manufacturing companies and keyboard arrangements rivalling the Sholes-Remington QWERTY. Yet, by the middle of the next decade, just when it had become evident that any micro-technological rationale for QWER-TY’s dominance was being removed by the progress of typewriter engineering, the US industry was rapidly moving toward the standard of an upright frontstroke machine with a four-row QWERTY keyboard that was referred to as ‘The Universal’. (David, 1986:38–9) Amongst the fortuitous events that David identifies as having facilitated the acceptance of the ‘QWERTY’ system were the timing and orientation of developments in typewriter instruction along with the advent of methods of ‘touch-typing’.7 David fills out the story in great detail, but the gist of his explanation is briefly stated. Basically, accidents of timing led to skills being taught and acquired that were fashioned to the QWERTY arrangement. Given © 1997 Tony Lawson
the various dispersed advantages of system consistency or parts-compatibility, along with the relatively high costs of rendering acquired individual skills obsolescent, set against the falling costs of typewriter hardware, the influence of these skills was such that the QWERTY system became locked-in.8 In short, David describes a ‘situation in which the precise details of timing had made it privately profitable in the short run to adapt machines to...[people] rather than the other way around. And the business has continued that way ever since’ (David, 1986:46). Examples such as this, and others gathered under the path-dependence heading, illustrate well some aspects of what an account of social reproduction/transformation will often involve. In particular, the emergence of structures can be heavily dependent on context, but once established, the very interactive, situated and continuity-preserving nature of human life is such that there are likely to be tendencies in place for the selected structures to get ‘locked-in’. David’s work admirably illustrates and provides comprehension of this process. It helps remind us of the inevitable heavy weight of the past in the present. However, there is a sense in which the path-dependence literature may also, on this very point, be misleading, and for my purposes it is equally illuminating to draw attention to this feature of it here. Specifically, I have reservations about the emphasis sometimes taken when generalisations are drawn from the case study evidence provided, including from the QWERTYUIOP example just considered. My worry is that, without due care to the way it is presented, the approach defended allows, and perhaps even encourages, the inference that once an account is provided of how a form of social organisation or a technology became established, this is more or less the end to serious enquiry, or anyway to the contribution of history. It facilitates the view that once a technology or social structure is in place then it can be treated as locked-in for good; that the past is not only ever present but also all determining.
FUNCTIONALISM AND MAINSTREAM THEORY This is not to deny that the generalised emphasis which the path-dependence literature places on the persistence of specific structures and technologies, serves as a valuable corrective to certain dominant features of the contemporary mainstream, including its ‘new institutionalist’ variant.9 In particular, it counters the mainstream treatment of resources and other relevant factors as completely malleable, and indeed perfectly adaptable, and adapted, to the pursuit of future ‘efficiency’, or some such. And it does seem to be against this functionalist aspect of much mainstream reasoning, along with a conception of © 1997 Tony Lawson
social structure and technology as plastic, that the path-dependence literature is basically orienting itself. David explicitly contrasts his own approach with ‘the “teleological” method...especially favoured by neoclassical economists’ (1994:206). Thus, he emphasises that while his own project is to link the present state of outcomes with some originating context and to interpolate...some sequence of connecting events that allow the hand of the past to exert a continuing influence upon the shape of the present...economists typically have followed [an alternative path]...in supposing that the present shape of things can best be explained by considering their function and particularly their function in some future state of the world. (David, 1994:206) But if it is clearly correct to criticise the excessive functionalism of aspects of mainstream thought along with the insistence that resources, including social structure, are perfectly malleable, there is a danger here of over-reaction. Specifically, there is a risk of reproducing the erroneous determinism of mainstream economics (including its stochastic variant) albeit in a displaced form, turning now on evolutionism and rigidity. I do not think David would wish to encourage this. However, it is a possibility that is apparently already being realised in some quarters. For it appears to be such a reading of the path-dependence literature which accounts for the fact that this approach is being cautiously embraced by various mainstream economists after all. We saw in Chapter 8 that a recurrent problem of mainstream ‘models’ is the presence of an overly large number of (equilibrium or other) ‘solutions’. The path-dependence literature appears to have encouraged the view in some that it is now possible to assert a unique equilibrium, at least in principle, without abandoning the familiar fictitious equations. And this is accomplished by assigning the task of achieving determinacy to the setting of ‘initial conditions’, a chore delegated in turn, in the name of path-dependence, to ‘history’—at which point history is effectively discarded. As David himself observes: Rather than shunning models with multiple equilibria on account of their indeterminacy, economists increasingly are content to leave open a door through which aleatory and seemingly transient influences, including minor perturbations specific to an historical context, can play an essential role in shaping the eventual outcomes. As the implications of mechanisms of self-reinforcing change in economics come to be more thoroughly and extensively investigated in the context of stochastic processes, using available methods from probability theory, economic © 1997 Tony Lawson
theorists and economic historians will be able to begin cooperating on useful programs of research. (David, 1992:1) It is possible to understand David’s emphasis here, given the desire to encourage economists to take history more seriously. But what real substance can there be to the claim that history matters if the sort of considerations that are intended to demonstrate this are also interpreted as being consistent with, or to have a counterpart in, a situation wherein history is all but denied? In truth, it is difficult to anticipate useful programs of cooperative research without mainstream economists first abandoning their basic ‘deductivist’ project. The fear is that in order to take part in cooperative research ‘historians’ end up trivialising or even abandoning history.
THE ECONOMY AS AN OPEN SYSTEM I stress, though, that the path-dependence literature is a useful contribution to economics. But its emphasis must be qualified by a recognition that every social phenomenon is a production of some sort, and to some degree at least could always have been otherwise. There is no social situation where explanation is inherently uncalled for. Even if decentralised decision-making leads to a system moving in one direction rather than another, there are likely to be intrinsic forces or tendencies which work, or could have worked, against this. In addition, if there is an absence of centralised direction this itself may warrant an explanation. Absences are no less significant as causal influences than anything else. This is acknowledged, for example, in world-wide debates about the consequences of NATO or UN non-intervention, or limited intervention, in the war in Bosnia, or economists’ debates about the consequences of non- or ‘minimised’ state intervention in ‘market economics’. In addition, conflicts of interests, incompatible goals, revolutionary ideas, technologies and other forms of novelty, are usually present. Where they are not this may once more constitute a situation that is itself in need of being explained. Let me emphasise that I am not merely acknowledging that items such as technologies, social organisations and procedures of action, can and do become obsolescent, or anyway displaced, albeit often long after they have outlived their usefulness. Rather I am insisting that even where a course of development appears to proceed unabated the potential for radical redirection may be present and liable to be triggered at any time. The extent to which specific absences, positive countervailing tendencies, and so forth, need to (or need not) be explicitly accounted for in order that some phenomenon and its © 1997 Tony Lawson
causes are satisfactorily explained, is a context-dependent issue, not an a priori one. The economy is an open system. Now to observe all this is not to deny that in some instances, depending on context, interests and understandings, little explicit elaboration may ex post be called for. We have seen in Chapter 12 that society can be viewed as a network of relationships and inter-related positioned-practices, with individuals slotting into particular positions and contributing to the reproduction of associated rules and relationships merely by drawing on them. It follows that once aspects of social structure are in place and acted upon they can come to endure merely through being an—often unintended and unacknowledged—outcome of the day to day activities of individuals ‘going on’ in life. Many conventions are like this, i.e. they express agreed ways of doing things that have become accepted and are regularly acted on (and thereby reproduced) even though alternative procedures could have arisen to serve the same purpose. Forms of greeting, boats passing each other on the port side, UK motorists keeping to the left side of the road, children raising hands to facilitate turn-taking in classroom speaking, are all of this sort.10 In numerous instances, clearly, individuals can have many very strong, and shared, vested interests in preserving specific procedures, forms of social organisation and technologies of which they are especially knowledgeable and/or upon which they perceive the usefulness of their skills (or those of their employees, etc.) to depend. At the same time it will often be necessary, or at least advantageous, for new organisational structures, functions, and technologies to be adapted to those already in existence, at least where the latter function in a satisfactory or desirable way, or anyway are not easily dislodged. For all such reasons particular aspects of some social structures may seem to evolve in a relatively smooth or apparently incremental fashion. Clearly, David presents the story of the QWERTYUIOP system as though, much of the time anyway, it conforms broadly to this scenario. But not all systems fit this pattern, even if some may appear to, at least for a while. Case studies such as David’s account of QWERTYUIOP capture aspects of social development; they cannot, certainly without more argument and evidence, be regarded as the general case. Often, and perhaps usually, social structures will be contested continuously. Sometimes there will be centralised attempts to ‘interfere’ with the outcomes of decentralised actions. In such cases it is transparent that the economist’s attention to history cannot be sidelined to the setting of initial conditions. However, I am maintaining that history cannot be sidelined even where structures have been found ex post to be relatively enduring. Rather, history, like geography, is an intrinsic aspect of economic science. Let me briefly illustrate some of these features with a well grounded © 1997 Tony Lawson
example. I choose a focus that is perhaps of more immediate relevance to economics than the QWERTYUIOP keyboard. It concerns a situation where once more a structure, or rather a set of structures, could be said to have become ‘locked-in’. But it is an example in which it is rather more clear that the structures in question could not, at any stage, have been taken as given, where their continuous reproduction requires explanation. Before proceeding with this example, however, I reiterate that my intention here is not to delve into the empirical grounds of explanatory conceptions examined. Although I do provide below some suggestive support for an explanatory theory focused upon, and do reference empirical assessments in footnotes, etc., to do more would occupy too much space and, if to repeat yet once more, be of little general relevance. Rather, my central purpose remains the illumination of the sorts of steps that a social explanatory exercise is likely to take if the framework defended above is acted upon.
BRITAIN’S RELATIVE PRODUCTIVITY PERFORMANCE AS A CONTRASTIVE DEMI-REG A partial regularity or demi-reg which has been referred to in previous chapters is Britain’s relatively slow rate of productivity growth over much of the last hundred years (at least prior to the 1980s) compared with that achieved by most other industrial countries.11 In particular, industrial economists regularly observe not only that staffing levels for any given type of machinery have regularly been higher in UK firms than in their overseas counterparts but also that, more often than not, UK firms have not responded as quickly and flexibly as others to technological and other changes in opportunities. British firms, for example, are frequently observed to have been slower at adopting new products and processes than their overseas counterparts even when they have been invented and often manufactured in the UK.12 This then is the phenomenon that I intend to focus upon here with the aim of pointing to the sorts of issues that arise in obtaining an adequate explanation. It can be noted immediately that the contrast which is captured by this demi-reg is fundamental to determining the particular explanatory factor which is sought. By its own standards Britain’s productivity record was more ‘impressive’ during the early post-war period than it had been for most of the preceding century. Clearly, a focus on this latter contrast seems likely to lead to the identification of a causal factor operative either during, or prior to, the post-Second World War period, but not in both periods— perhaps the extraordinary post-war expansion in world trade and demand. However, this is not the partial regularity which, for illustrative purposes, I am choosing to focus on here. Rather I want to consider an explanation of © 1997 Tony Lawson
Britain’s ‘poor’ performance relative to that of other countries throughout the period in question. And for this there is a prima facie case for supposing that the factor responsible is fairly unique to Britain’s experience and operative throughout the period. It thus seems plausible that the causal factor in question is a relatively localised set of structures that has been reproduced repeatedly. A first key question, then, is what (set of) structure(s) could, if it existed, give rise to mechanisms that would account for the contrast in question. Obviously, most economists posing this question do not do so in a knowledge vacuum; they inevitably draw on their prior knowledge of economic history as well as of the sorts of structures and mechanisms that can facilitate or inhibit ‘productivity-increasing’ changes in the place of work. Such knowledge then figures in the retroduction of hypotheses of the causal mechanism(s) responsible. Not surprisingly in the light of the phenomenon to be explained, the hypotheses most frequently retroduced range from theories of severe in-built obstacles to flexibility in the UK manufacturing context, to the persistent making of mistakes by those with the power to direct investment resources.That is, they range from the deterministic to the voluntarist extremes (see Northover, 1995). The most comprehensive attempt, that I am aware of, to compare the explanatory powers of the range of theories, has been carried out by Stafford (1983). This author deduces implications which would follow if the different theories considered were true, and checks each out empirically in the appropriate context. On the basis of a wealth of detailed evidence13 and an array of context specific strategies, which I cannot reasonably repeat here (though I provide some discussion of such issues towards the end of the chapter), this author concludes in favour of an explanation which attributes at least part of the responsibility for the experience of relatively slow growth in the UK to the latter’s far more highly decentralised system of collective bargaining compared to any which emerged elsewhere, along with the associated relatively localised nature of worker organisation in the UK.14
THE NATURE OF THE EXPLANATION Let me expand briefly on the substance of this particular explanation—which I have also addressed elsewhere (see Kilpatrick and Lawson, 1980).15 The system in question is one in which all aspects of work have traditionally been negotiated at local as well as national level; where workers are organised more on a craft- (perhaps in addition to an industry-wide) basis, certainly to an extent that is not replicated elsewhere. The argument usually advanced is © 1997 Tony Lawson
that this structure, more than others, facilitates mechanisms which inhibit coordinated decision-making and works against quick or smooth responses to changes in production possibilities. A crucial factor is that changes in productive potential are not neutral. As technology changes, so existing skills can become redundant. Britain’s history of local worker organisation, including relatively strict demarcation lines, has tended to produce a situation wherein any given group of workers has but a specific set of skills. This means that with any change in technology which reduces the need for certain specific skills, the incentive on the part of the workers involved (including any who in principle could be affected) to resist the change in question is that much greater. Of course, given the job-based nature of worker organisation, effective resistance can, and is likely to be, forthcoming. This explanation does not deny that the reorganisation of production continually takes place. It merely recognises that where resistance, or the threat of resistance, is strong, concessions can be, and often are, obtained concerning the conditions of work generally—including staffing levels of new machinery, or the maintenance of ‘skill status’ and associated rates of pay. Typically these concessions provide the basis for further resistance to future proposed changes, so that, one way or another, industrial change tends to proceed more slowly than would otherwise be the case. Nor does the hypothesis in question entail any necessary criticism of worker organisation, or of any other party. Rather, it constitutes a commentary upon the nature of uneven development. Workers and their unions need not be opposed to new technology per se. Rather, and understandably, they are concerned to adapt to new technologies under terms and conditions which avoid redundancies and at least preserve the skill status and the standards of the conditions of work of the people involved. In short, the explanation in question focuses on a set of structures that empowers workers (as well as others) to exercise a significant effect on all manner of outcomes. The exercise of this power can be, and usually is, situationally quite rational and comprehensible, and is essentially a defensive strength. Workers usually only react to management designs for change perceived as detrimental to their own situation; they rarely initiate, and certainly do not control, developments. Workers bargain to maintain living standards and working conditions, or at least to be partly compensated for detrimental changes to work conditions, skill status, and so on. However, negotiations can take time, and the likely opposition of workers may even deter management from taking up the potential for change. Of course, the defence of standards or ‘norms’ is carried out by all members of all societies. The important factor in the sphere of international competition is the relative ability to achieve these aims. If UK workers have achieved the ability to bargain over the conditions of introduction of a new process, where © 1997 Tony Lawson
in another country it is brought in without negotiation, then the productive system in the latter country will be inherently more flexible and, in a changing environment especially, tend to experience faster productivity growth. Needless to say, contrasts are rarely so sharp, but then conditions are never everywhere equal.
EXPLAINING THE EXPLANATION It is the case, to repeat, that there exists a good deal of evidence supportive of the view that the relatively decentralised nature of collective bargaining, and in particular the localised nature of worker organisation, have led to inflexibility in the workplace, constraining productivity growth rates compared to events elsewhere. The explanation as it stands, however, is nothing like sufficient for a comprehensive understanding of the noted phenomenon. Rather, in a world of perpetual change, where the competitive process includes tendencies to converge as well as incentives to differentiate, an adequate understanding of Britain’s persistently ‘poor’ relative productivity performance requires an explanation of the explanation. Specifically, an adequate understanding of the situation necessitate at least: 1 some insight into how a widespread structure of decentralised bargaining organised along craft lines came about in the first place; 2 an account of the process by which it has been reproduced over time; along with 3 some understanding as to why a similar system was not established elsewhere, or, if it was, then why it has persisted (longer) or perhaps operated differently in Britain. I have argued that the task of explaining the process whereby relatively enduring structures are reproduced is important in most contexts. But it seems particularly so in the one in question when a priori it might have been supposed that, given the outcomes sustained, all the pressures would have been for the identified system’s early demise. What explains the findings of the Donovan Report (HMSO, 1968) that it remains the case ‘that the craft system is deeply rooted in much of British Industry’ (1968:87), when the experience elsewhere has been so different? Why is it that, despite amalgamations that greatly reduced the number of unions in the twentieth century, from 1,269 in 1913 to 488 in 1975, by the late 1970s little had occurred by way of a simplification of the overall position, or a centralisation of control (Clegg, 1976:31–2).
© 1997 Tony Lawson
‘ORIGINS’ OF THE EXPLANATION What sorts of factors, then, explain the emergence, persistence and relative uniqueness of Britain’s industrial system? The obvious factor to highlight in regard to the emergence of this system is the timing and context of union ‘acceptance’ in the UK. The end of suppression of organised collective resistance by the military and the courts in Britain occurred relatively early on compared with other countries—in the mid-nineteenth century. Between 1867 and 1875, a series of acts were passed which made it easier for workers to strike without being imprisoned (Master and Servant Act, 1867; Employers and Workman Act, 1875); which helped secure the legal standard of trades unions (Trades Union Act, 1871); and which legalised peaceful picketing (Conspiracy and Protection of Property Act, 1875). And a significant feature of this early recognition is that it occurred well before the introduction of mass production techniques. Thus, the point at which workers achieved a significantly increased power to resist changes, to defend standards and procedures, coincided with a preponderance of customs and norms that reflected the craft-based form of much industrial organisation. And a consequence was that the system of industrial relations that in due course became institutionalised was highly decentralised, uncoordinated and certainly relatively unsuited to the later needs of mass production.16 If the timing of legal emancipation was a central contributory factor in the early impetus to the institutionalisation of the craft-oriented nature of worker organisation, a significant element in the explanation of the spread of collective bargaining to cover most aspects of industrial decision-making was the absence of any associated political party ready with a programme of positive rights for the unions. Absences, I have insisted, can be as causally efficacious as any other causes, and it is significant that the British unions in the nineteenth century, unlike all other European trades union movements at their inception or at comparable points in their development, did not have an associated political party ready and waiting with a programme of ‘rights’. While legal collective bargaining was achieved in other European countries on the basis of positive rights: the rights to associate, to bargain or to strike, in Britain trades unions obtained not the right so to act, only immunity from the law if they did. Thus, the Trades Union Act of 1871 provided unions with basic protection from criminal and civil consequences of the restraint of trade doctrine (although the Act stopped short of positive encouragement for trades unions); the Conspiracy and Protection of Property Act of 1875 allowed that where no criminal act is committed, an organisation or combination to strike in furtherance of a trade dispute is itself not a criminal conspiracy, and so on. As a result, trades unions came increasingly to argue for ‘freedom’ from © 1997 Tony Lawson
the law, meaning the removal of legal restraints on their activities so that they would be free to bargain as individuals. Indeed, modern research into the history of British labour law tends to draw upon the notions of voluntarism or legal-abstention in describing these developments and their consequences. These terms are not always used in particularly concise or consistent ways but, broadly speaking, are usually employed to signify that the degree to which the legal system in Britain comprehensively controls the conduct of industrial relations and actions, is markedly less than in other industrialised countries. The ‘voluntarist system’ which developed was not necessarily in line with the exact intentions of any of the participants. The unions, for example, pushed initially for legal protection—something they had campaigned for at least since the eighteenth century. It was because their campaigns for legal protection were largely unsuccessful that, bit by bit, the view developed that wages, hours and other basic terms and conditions of employment, should fall within the domain of collective bargaining rather than statutory regulation. There was never a stage at which this outcome was guaranteed. As Lewis (1976) notes: There was no historical inevitability about...[the development of collective bargaining to cover such a wide range of issues]. In the nineteenth and early twentieth centuries the unions campaigned for a legal minimum wage and the legal limitation of the maximum number of hours in the working day. Both objectives moreover still survive in rule 2, the objects clause, of the TUC’s constitution. But the campaigns met with little success. (Lewis, 1976:7) However, the development of collective bargaining was the outcome, with the consequence that unions looked increasingly to collective bargaining for dealing with all aspects of workers’ employment. As Lewis’ detailed history of labour law finds, it was because of their lack of success in achieving legal protection that the: unions themselves ultimately ceased to press for the wider statutory protections. Instead, they came to rely increasingly on collective bargaining as the single most important method for regulating the terms of their members’ employment. Over a period of time employers were prepared to concede what Parliament denied, and consequently through experience unions became deeply attached to the efficacy of industrial negotiations over the method of legal enactment. (ibid.: 8) © 1997 Tony Lawson
TENDENCIES TOWARDS ‘LOCK-IN’ The last observation is significant, for if factors such as the timing and context of legal emancipation appear to be crucial to understanding the source and the rootedness of the craft system in UK industry, as well as the extent of localised collective bargaining, they do not explain its persistence. And one often-cited factor which appears fundamental to understanding how the decentralised system of collective bargaining and workplace worker organisation became ‘locked-in’ is British management’s acceptance of the reality of workers’ power in many industries in the late nineteenth century and its preparedness to accommodate it. British management came to take a more conciliatory approach to workers’ organisation than its overseas counterparts, and attempted early on to institutionalise current practices, customs and the resolutions of disputes into workplace or firm ‘procedures’. In this way it sought to minimise the number of disruptions not by challenging the basis of worker resistance, but rather by attempting to limit its effect. Friedman (1977), in discussing the use of procedure and other forms of conciliation from 1870 onwards, even suggests that the most crucial development of the capitalist mode of production in Britain during the 1870s to 1914 was the acceptance by many employers of organised worker resistance as a permanent reality in their day-today dealings with workers (American acceptance came somewhat later). (Friedman, 1977:97) More recently, Brown and Rea (1995) find that ‘Collective bargaining was widely perceived by employers as unavoidable if not necessarily desirable’ (1995:367).17 They continue by observing that Throughout much of manufacturing and the public sector and over a wide range of issues, negotiation with trade unions was the normal practice. An internationally distinctive feature of joint regulation in Britain was the extent to which it was operated at the place of work, rather than at some higher corporate or industrial level (irrespective of whether there were industry-wide agreements), with the consequence that workplace union organisations often exercised detailed control over working practices. (ibid.) Unions were an active party to forming agreements with employers for they saw in them a means of defence against further losses of control over the organisation of work. Moreover, there were sectional gains to groups of workers, as procedures were established which related to promotion and © 1997 Tony Lawson
discharge. In this way, even relatively unskilled workers gained some protection from the competitive forces of the ‘external’ labour market. Thus, the structure of decentralised decision-making and internal ‘collective’ bargaining were further reinforced. No doubt many other factors are also likely to have contributed to the apparently ‘locked-in’ state, to the continuous reproduction, of a decentralised structure of bargaining. Likely candidates include the widespread existence of systems of internal labour contracting in the nineteenth and early twentieth centuries,18 and the contrasting experiences of two world wars.19 I shall not expand upon these matters here, though. My aim is mainly to indicate the sorts of explanatory factors which are usually identified by those who consider these issues seriously. All such considerations, however, indicate the existence of tendencies for certain structures to become increasingly integrated with a larger system as a whole.
SOCIAL REPRODUCTION/TRANSFORMATION: AN EX POSTERIORI CONTRASTIVE PHENOMENON So far the structure of the explanation provided is much the same as that found in the path-dependence literature.20 But there is still more to be determined for an adequate explanation. After all we are wanting to explain Britain’s relative productivity performance. If the system of industrial relations in Britain has exhibited a degree of continuity throughout the period in question there has also, and inevitably, been persistent change. For an adequate understanding of Britain’s relative productivity performance it is also necessary to explain the fact that the UK system of industrial relations has evolved more smoothly, or anyway in a manner that is significantly different from that in which other countries’ systems have developed. If British management came to take a more conciliatory approach to workers’ organisations than elsewhere this itself requires explanation. In other words, in an internationally competitive system, the very fact of a significantly different experience elsewhere indicates a prima facie case for supposing that Britain’s evolutionary path, its apparent locked-in state, is in need of being explained. Why then has the British industrial relations system evolved as smoothly as it has, and more so than systems elsewhere?
CONDITIONS FOR SOCIAL REPRODUCTION There are numerous causal factors that could be singled out here. Previous investigations of this phenomenon indicate convincingly enough that most other industrial countries not only underwent greater upheavals due to more © 1997 Tony Lawson
rapid industrial transformation, but many also experienced political traumas of a sort not experienced in the UK, sometimes associated with (defeats in) war and with strong state, or legal, influence over development, including the centralisation of industrial negotiations. In Germany, for example, the labour movement was suppressed by the Nazis, with unions being abolished in 1933 and unable to re-form until 1945. By this time organisational impediments to industrial unionism had disappeared. Moreover, the western forces supervising the German reconstruction favoured total centralisation. This latter objective was only dropped following strong opposition from the British TUC. German trade unionists aimed to create a unified structure to avoid the ideological, religious and craft divisions of the Weimar period, with the result that the present day structure is highly centralised (Günter and Leminsky, 1978:168, 169, 179). In France, the unions were effectively broken up with the suppression of the 1871 Paris Commune (Clegg, 1976:34). Although legal recognition was granted in 1884, it implied only the right of association; there was no obligation on the part of employers to negotiate. Unlike in Britain, no system of collective bargaining arose to give legitimacy to existing customs or procedures. The very notion of the obligation to bargain has always met with strong opposition from the courts, and it was only in 1971 that a law was enacted providing for penalties against the party that would not answer a demand for bargaining (Sellier, 1978:217–20.) In Italy, Mussolini replaced the Confederation of Labour by govern-mentcontrolled syndicates, which persisted until the overthrow of fascism in 1945. Moreover, the state-controlled unions which he set up were never properly dismantled in the post-war period, so that the centralised nature of organisation has continued. Indeed, for most of the post-war period there has been no direct union presence in factories in Italy or in France. Only since 1970 in Italy and since 1968 in France has this begun to change (Carew, 1976:31, 32, and 60–4). Of course, in any explanatory account the sorts of factors singled out will reflect the interests and understandings of the investigator. For many economists the most relevant question is why the competitive process itself did not lead to tendencies which undermined Britain’s relatively inflexible system of decentralised bargaining. The pursuit of this question has usually led to a focus upon the existence of protected markets which for long periods were available to UK producers, a factor which reduced the necessity for prolonged confrontation with workers over the restructuring of work,21 and so on. Social reproduction or continuity, then, is a phenomenon as warranting of explanation as transformation. It needs appropriate conditions (including absences of undermining mechanisms) just like any other phenomena. Now if © 1997 Tony Lawson
the preceding outline illustrates this, it is equally necessary to stress that even where the extrinsic factors seem supportive (or anyway not necessarily undermining) of definite social structures, the latter still need to be actively reproduced. I have stressed that there will almost always be an intrinsic opposition of interests and other tendencies that could have resulted in things turning out differently. Let me labour the point just a bit further through drawing once more on the history of the UK system of industrial relations.
RELATIVE CONTINUITY AS AN EX POSTERIORI PRODUCTION/ACHIEVEMENT I have emphasised throughout that social reproduction and transformation is achieved through the sum total of human everyday actions and interactions. And I have noted that the reproduction and transformation of certain structures is often an unintended aspect of this everyday interaction. However, this is not always the case. Indeed, it is a significant fact of much social life that social reproduction and transformation turn upon actions where the goals of changing, or of resisting changes, to aspects of social structures are explicitly formulated and actively pursued. Industrial relations in Britain (as of course elsewhere) have always been reproduced via a process in which overt conflict is an integral feature, where change as well as resistance to change, are continually in evidence. Where and when conflict of this sort is endemic it is evident that even where structures appear ex post to have endured there could never have been any inevitability about this outcome. I now illustrate by touching briefly upon a few significant moments in the development of the system of British labour law, a system that evolved in response to, and in turn shaped and became an internally related aspect of, the whole process of localised collective bargaining, a system of labour law which, I have noted above, is often referred to as voluntarist. I should emphasise immediately that usage of terms such as voluntarism or legal abstention does not signal anything like an absence of labour law in industrial relations. Rather, it indicates that much of the labour law that emerged after 1870 was intimately related to, and to a large degree supportive of, the institution of collective bargaining.22 Even so, when the industrial relations system that developed in the later part of the nineteenth century in Britain is compared with those of other countries it clearly has an appearance of relative legal abstention. Of course, developments were rarely smooth or straightforward. And abstention did not extend to war-time periods at all. Yet it is significant that as recently as 1985, when commenting upon the recent
© 1997 Tony Lawson
changes to labour law, Wedderburn was still able to conclude that the ‘laws of the 1960s and 1970s seem to have made little mark on the “voluntary” system in its overall operation’ (Wedderburn, 1985:33). The fundamental point I want to make, though, one which relates back to the discussion of the path-dependence literature (with its occasional encouragement of the inference that once a structure is locked-in, its persistence can be taken as given), is that the voluntarist system, no matter how broadly interpreted, and despite its ex posteriori longevity, was never something that, once in place, or recognised, could be regarded as fixed or given. Even in the period from 1870 up until the First World War, the period in which voluntarism or legal-absenteeism is usually considered to have become established, legal gains obtained by labour were frequently challenged or rolled back by the courts or by Parliament, only for labour to campaign once more for effective protection. In other words the legal structure was consistently under pressure for change and protective measures gained by labour had continually to be defended. Thus, although the Criminal Law Amendment Act, 1871 (which accompanied the Trades Union Act) repealed the 1825 Combination Act and supposedly dealt with criminal liabilities at large, this interpretation was overturned when organisers of an 1872 gas stokers’ strike were prosecuted for criminal conspiracy. Subsequently a union campaign ensued, and, in 1875, there followed the Conspiracy and Protection of Property Act in which immunity was obtained from the crime of ‘simple’ conspiracy when the combination was ‘in contemplation or furtherance of a trade dispute’. In their study of the period the Webbs note that in the last quarter of the nineteenth century ‘trade unionism had by its very success again become unpopular among propertied and professional classes as well as the business world’ (Webb and Webb, 1920:597). In the event, there was a wave of judgements against the trades unions. This time the legal threat was based upon civil liability for torts (civil wrongs) created by the judges. Criminal liability was replaced by tortious liability for ‘simple’ conspiracy as judges again condemned trades union objectives as unlawful (whilst regarding employers trade objectives as perfectly legitimate). Although numerous hazards were created for unions throughout the period around the turn of the century (also including the tort of inducing breach of contract), the main blow came with the Taff Vale case of 1901. Under this the House of Lords held that a registered union itself, as distinct from its officials, could be liable for tort damages. This effectively undermined the main advantage to the unions of their legal status as unincorporated associations and made them in many cases liable to heavy fines. Such developments probably helped boost the newly formed Labour Party, as union support and affiliated membership for it grew significantly in the early 1900s, with twenty-nine purely ‘Labour’ MPs being elected in 1906. In any case, as unions campaigned to regain their legal © 1997 Tony Lawson
protection, the Trades Disputes Act of 1906 was passed, following the election of the Liberals, which despite the protests of lawyers and employers alike, reversed Taff Vale and entrenched the tradition of ‘legal abstention’ as it has subsequently come to be called. But once more nothing was fixed. Following this demonstration of trades union political power and influence the judges became sharply hostile. In fact in the 1909 Osborne decision the Law Lords invalidated union financial support for the Labour Party and, in a ruling which according to Wedderburn (1985:31) ‘is now seen by no one as other than a technically incorrect and class-biased judgment’, limited the functions of trades unions to the matters explicitly mentioned in their statutory definition in acts of 1871 to 1876, thus excluding political activity from lawful purposes. Parliament validated political activity by the unions in the act of 1913 but only on condition that unions establish a separate ‘political fund’ and that each member be permitted to ‘contract out’ of the levy for that fund without disability to the union. In fact, the period from 1910 to 1914 may well have culminated in a general strike (Dangerfield, 1936) had it not been for the eventual outbreak of war. Once protective labour law was won, then, it was constantly under pressure for change; ‘voluntarism’ had to be actively and continuously reproduced. This sketch of events refers to the period prior to the First World War, but the general observation made applies throughout the last century. For example, not only the Taff Vale judgement of 1901 but also the 1927 Trades Union and Trades Disputes Act (which followed the defeat of the General Strike), the 1971 Industrial Relations Act, as well as the legislation of the 1980s, have been designed to make much effective trade union activity illegal. Of course, attitudes towards, and attempts to change, the law have been influenced by the relevant context prevailing. For example, the existence of formal and informal imperial (protected) markets, especially in the 1930s, offset the need for employers to embark upon protracted confrontation with workers over the needs of restructuring. And by the same token the disappearance of such markets, alongside the conditions of near full employment and therefore the strengthened position of workers within the process of collective bargaining which followed the Second World War, led to sustained attacks by the establishment upon the voluntarist system (Brown and Rea, 1995; Wedderburn, 1985). In other words, actions by individuals and groups of individuals, particularly in pursuing their demands concerning wages and employment conditions, have influenced the development, interpretation and use of (and challenges upon) law as much as vice versa. In fact, in the two decades before the 1980s the ‘development of labour law has been subject to bewildering changes of pace and direction’ (Lewis, 1976:1), mostly oriented around © 1997 Tony Lawson
attempts to undermine, or alternatively reproduce, existing structures. On itemising many of the changes that occurred Lewis notes: These events were a forceful reminder that major developments in labour law have occurred at times of economic and social conflict. The nature and extent of legal regulation has been determined not by some abstract rule-making force, but by the interplay of judicial innovations, public policy controversy, the relative power of management and labour and party politics with view to electoral advantage. Labour law past and present is only explicable within a firm historical framework which takes account of all these factors. Indeed the memory of previous conflicts over the law is itself a salient feature of political and industrial struggles concerning labour legislation. (ibid.) The central point here, though, is that at no stage during the period in question were participants and observers able to take developments for granted as an historical inevitability. Indeed, at each stage throughout the post-Second World War period observers questioned whether voluntarism was in decline (see e.g. Hawkins, 1971; Lewis, 1976; White, 1978); the continuance, or otherwise, of ‘legal abstention’ was something that had to be established rather than merely taken for granted. This questioning of whether voluntarism could continue, in fact, can be traced back to Dicey (1914) who, despite the 1906 Trades Disputes Act, saw fit to conclude, disapprovingly, that he was observing a trend of collectivist legal intervention, based upon ‘faith in the benefit to be derived...from the action or intervention of the State even in matters which might be, and often are, left to the uncontrolled management of the persons concerned’ (1914:259). The Webbs (1920) also predicted that legal enactment’ would soon supplant collective bargaining as the main method of dealing with industrial relations. In brief, the endurability of the British system of (relative) ‘legal abstentionism’ must be interpreted as an ex posteriori development. Continuity of structures presupposes human action as much as vice versa. If the system in question has proven to be relatively enduring (although of course always subject to some change) there was never any inevitability about this outcome.
THE EXPLANATION IN BROADER PERSPECTIVE In the last few sections of this chapter I have focused upon one possible explanation of the ‘inferior productivity performance’ of much of UK industry over the last hundred years or so, a set of structures which clearly comprises © 1997 Tony Lawson
a geo-historically grounded, intrinsically dynamic, internally-related, totality. And I have considered, amongst other things, aspects of the process of accounting for the continuous reproduction of the set of structures in question. Primarily, I have attempted to indicate that even where recognisable aspects of sets of structures appear to be reproduced relatively ‘intact’, there was never any inevitability about this, there was no inherent fixity involved. The potential for fundamental change may often be present and, where this is so, its non-realisation a possible matter for enquiry. It is also perhaps worth re-emphasising that in accounting for the phenomenon in question, many of the causal factors identified as explanatorily significant constitute a set of relative absences—of defeats in wars and occupation, of severe competition in certain imperial and other markets, of out and out hostility by British management to workers’ organisation in many industries, and so on. These are important considerations; absences, I repeat once more, are just as real and consequential as any other causal forces. Yet any actually enduring structure or system must also be actively reproduced. Social structures or systems, once in existence, are never thereby just fixed or given. Even in the relative absence of countervailing forces, the structure(s) in question must necessarily be reproduced.23
THE PARTIALITY OF ALL EXPLANATION If it is a clear implication of the foregoing illustrative overview that social explanation which ignores the social reproduction of causally efficacious structures will often be too partial to constitute an adequate or sufficiently comprehensive understanding of some phenomenon, it is perhaps warranted that I explicitly acknowledge once more that partiality of sorts is nevertheless inevitable. In fact, explanations of Britain’s relatively poor productivity performance which have followed the lines sketched above (for example, Kilpatrick and Lawson, 1980) have received most criticism for what is perceived as their partiality. Various types of omission (features abstracted from) can be distinguished in regard to the sorts of factors discussed above. These include especially (1) forces that, during the period or over the spatial region in question, can be treated as having been ex posteriori broadly constant, or homogeneous (across countries, industries, etc.) in their mode of operation; and (2) causal conditions that either complement or counteract the explanatory mechanism identified. I briefly consider each in turn. Explanations of the sort considered above have recently been criticised by Kennedy (1993) for failing to allow for the capitalist drive to accumulate, the ceaseless attempt to increase the rate of surplus value. But it cannot be © 1997 Tony Lawson
supposed that any such mechanisms not explicitly emphasised in such studies are thereby treated (implicitly or explicitly) as non-existent. The sort of mechanisms that Kennedy is pointing to will have been operative throughout the region of time-space (throughout the last hundred years and within all countries and areas mentioned) in question. The preceding outline, for example, refers explicitly to workers’ resistance, new technological possibilities, inflexible work practices and protected imperial markets. But, of course, such conceptions presuppose tendencies that are being resisted, mechanisms that bring about new technological possibilities, interests of capitalist enterprises that require flexibility, that protectionism counteracts competition, and so on. Such factors as identified by Kennedy, then, are not neglected. Indeed they are essential to comprehending the whole process. But the point, rather, has been to render explicit certain structures and tendencies that vary across regions and appear crucial to explaining certain aspects of the productivity process in question: that growth rates have been uneven across regions over a significant period of time, with the UK, in particular, often experiencing slower growth rates than other industrial countries. The second form of criticism is that numerous UK specific mechanisms bearing directly upon productivity performance are neglected. In particular, certain actions or inactions of the UK government, or of UK managers, are usually singled out in this regard. In competing explanations either the state is interpreted as not intervening effectively enough in industrial relations (Elbaum and Lazonick, 1986; Fine and Harris, 1985; Hall, 1986), or UK management is variously seen as uneducated, insufficiently capable in some respect, as possessing the wrong agenda, having invested too little in new technology, and so forth. Fine and Harris (1985) even emphasise ‘management complacency’. To the extent that such claims can be supported empirically and are complementary to the sort of explanation here identified there is no problem in all this for the latter. The aim of any explanatory project will usually be to identify and draw attention to one efficacious mechanism, in the current example one which is crucial to Britain’s relative unique productivity growth performance; certainly not to identify all that has been going on. In an open social system the provision of a complete explanation of some state of affairs is not a realisable goal. But this does not undermine the explanation provided, any more than the theory of gravity as a causal mechanism is undermined by the realisation that the path of any given leaf warrants, for its complete explanation, an attention to aerodynamic, thermal, and very many other tendencies as well. To the extent, however, that the other explanatory claims compete with the one considered here, it is clearly desirable to determine which is the most empirically adequate. I have referred to Stafford’s work above as an example © 1997 Tony Lawson
where competing hypotheses of Britain’s relative productivity performance have already been assessed according to their relative empirical grounding. It is the set of structures and mechanisms discussed throughout this chapter that is found to be the most empirically adequate. Of course there is always the possibility of further hypotheses being proposed whose comparable empirical adequacy will then require examination. And the possibility of success in any ensuing attempts to gauge a theory’s empirical support, of course, will depend upon context and availability of relevant evidence. If, for example, British management is really somehow less capable or more complacent than others, then presumably those brought in from overseas, and perhaps operating UK-based production units of overseas multinational companies, should be found to be associated with productivity performances that are frequently notably better. If insufficient investment is considered to be the problem then analyses which examine the contrasting rates of investment across time or space should reveal some evidence of a corresponding relation in productivity performances, and so forth. The explanation here being focused upon can itself be further assessed, moreover, by examining whether the different productivity tendencies that are manifest across regions outside, or even within, the UK are associated noticeably with variations in the degree of decentralised collective bargaining and job-based worker organisation, and so on. This is not the place to explore such matters; the aim here is merely to indicate how explanatory endeavours of a certain sort are possible and warranted. But, whatever the stage of any analysis, there will always be more to do. Certainly the suggested explanation of Britain’s relative productivity performance, if correct, leaves many relevant questions unanswered. Why, for example, were British trades unions legalised when they were? Why did fascism not take off in Britain in the way it did in some other countries? Why was there no political party with a programme of positive rights available to British trades unions at the time of their legal emancipation; and so on? Each explanation is in terms of a set of structures and mechanisms which eventually is itself in need of explanation. Economic analysis as conceived here, then, will usually be a complicated and messy affair. Unlike the simplistic positivistic conception of science as elaborating event regularities, the process of uncovering and explaining significant causal structures and mechanisms, including geo-historically rooted and dynamic totalities, will usually be a painstaking, laborious, and time-consuming, transformative activity, one that gives rise to results that will always be partial and contingent (and usually contested). The explanatory process will inevitably involve looking at certain features of some structure or mechanism or system to the (momentary) neglect of others,
© 1997 Tony Lawson
understanding some structure, etc., from a particular angle, leaving certain questions at any stage unanswered (so far) and warranting of further attention. But if, in economics, such complications are unavoidable, in this respect as in many others, I repeat yet once more, the situation is just as in any other science.
© 1997 Tony Lawson
Part V
ECONOMIC POLICY AND FORECASTING
© 1997 Tony Lawson
19 ECONOMIC POLICY AND INTENDED CHANGE
I turn to consider the implications of the argument for the process of policy formulation, or, more generally, any programme of rational, intentional, social change. The current dominant practice of policy analysts is to formulate large scale econometric models which are used for extrapolating past trends in producing ‘best forecasts’. Such models comprise an integrated system of hypothesised (typically probabilistic) relationships of the generic form ‘whenever event (or state of affairs) x then event (or state of affairs) y’. The most familiar representation is as a functional relation broadly along the lines of the following: y = f(x1,x2, ...xn,z1, ...;zm,)
(A)
where (1) y, the x’s and z’s are variables (or vectors of variables) denoting measurable events or states of affairs, (2) the x’s and y are considered to be endogenous, that is determined within the system, and (3) the z’s, in contrast, are treated as exogenous. The e component, which is often referred to as the disturbance or error term, is usually interpreted as a surrogate for those variables omitted from the equation (including supposed explanatory variables, any intrinsic randomness in y, as well as errors of measurement) which collectively affect, or are a part of, y.1 Notoriously, such econometric models, whether considered on an equation by equation basis or when combined into a larger system, perform rather badly on the forecasters’ own criterion of predictive accuracy. And, as we saw in Parts I and II, there are other problems or tensions as well. Let me briefly review.
ORTHODOXY, ECONOMIC POLICY AND CHANGE Most noticeably, such models, not withstanding their probabilistic gloss, are essentially deterministic in the sense that human choice is all but denied. In a standard ‘consumption function’, for example, and in line with the exposition © 1997 Tony Lawson
in Chapter 7, consumers’ expenditure (under some transformation) is represented as a fixed function of disposable income (if amongst other things), so that a unit change in the income (or in its rate of change or whatever) induces an adjustment in consumption by an amount that is fixed and given even before actual consumers come to know about any change in income that eventually transpires, whatever its source and context. In short, people do not make history in this scenario but are merely propelled along by external events and conditions. In consequence, there arises the question of whether it is even meaningful to talk of economic change in this context (see Chapter 12 above) and so whether there is any point to the activity of policy prescription. It is true that if we look back to formulation (A) there appears to be some scope for change of sorts if (at least some of) the exogenous variables (the z’s) are interpreted as choice variables, as instruments of government control or whatever. That is, if such exogenous variables are really instruments in the control of policy makers it would appear that an element of choice is afforded in affecting their values, allowing the possibility of some influence over y. But any such separation between, on the one hand, the inability of lay agents to choose and, on the other, the real possibility of government policy formulators to make a difference, is hardly credible, as we have seen. Of course, from the perspective of contemporary orthodox economics, with its emphasis on ever-rational, near omniscient, individuals, the obvious anomaly here is the supposition that the variables treated as exogenous, the apparent instruments of government policy, are in fact unpredictable. Put differently, the noted dualism is inconsistent with the mainstream belief that everything can ultimately be endogenised, that everything can be deduced from some ‘whenever this then that’ formulation. And as we saw in Chapters 1 and 7 certain advocates of the influential rational expectations hypothesis have recently asserted explicitly that changes in government policy instruments are indeed predictable, so that any real choice in that department is assumed away after all. But then we are returned to the question of whether there is any point to policy analysis—or indeed to any other activity which is premised on the belief that things could have been otherwise? We have seen that this is a question or puzzle which mainstream economists themselves frequently raise, albeit without achieving any satisfactory answer. To repeat Reder’s (1982) conundrum addressed in Part I: ‘To the extent that variables are endogenized—choice is explained—society’s freedom of choice is seen as illusory’ (1982:35). We can also recall that even if we were to overlook these tensions and allow a role for the government control of policy instruments in this rather forced set-up, the notion of change which could then be supported is extremely impoverished. There is no scope for transforming modes of determination; social life reduces to little more than the flux of events. In fact, the idea of © 1997 Tony Lawson
change reduces to nothing more than adjustments to, or the amelioration of, events and states of affairs. Our assessment of the traditional mainstream framework on these issues is thus easily summarised: while ‘economics’ is made too easy, choice and change are rendered little more than illusory, and the practice of policy formulation rendered pointless.
CRITICAL REALISM, ECONOMIC POLICY AND CHANGE What then is the alternative view on policy and intentional change that I intend to defend? The central claim is simply stated. Whereas the most that supporters of the traditional, positivistic, constant-conjunction view of science can sustain with any consistency is the goal of control along with the amelioration of events and states of affairs, the critical realist perspective instead offers the real possibility of human emancipation through structural transformation. For if according to the former view the (only conceivable) point is to (attempt to) fix a set of events x in order to determine and thereby control a dependent set of events y, on the critical realist understanding a feasible aim is to transform real social structures in order to facilitate alternative opportunities. There thus arises the possibility of enhancing the scope for human potentials to be realised, of broadening human opportunities; it is feasible to think in terms of replacing structures that are unwanted, unneeded and restrictive by those that are wanted, needed and empowering. Choice is no longer denied. On the contrary, it lies within the realms of policy objectives to aim to widen the scope of choice, with respect, in particular, to options that are both needed and desired. Two relatively distinct scenarios for bringing about change of this sort can be distinguished. The first is a situation in which there is no necessary misunderstanding of the nature of the structures or mechanisms involved, even if the social transformation in question may require considerable planning, organisation and/or agreement. Examples that are often mentioned in contemporary UK society include a switch from driving on the left to driving on the right side of the road, with all accompanying upheavals which that would entail; altering clock-times corresponding to sunrise, etc; or allowing Sunday trading. The second scenario, which in many ways is the more interesting one, is that of explanatory critique. This is a situation in which beliefs are shown to be inadequate to their situation. This scenario focuses on a critical possibility, available only to social (i.e. not to natural) science, wherein it may be possible to transform a set of structures through facilitating a change in the manner in which each is understood. Specifically, it lies within the potential of social science both to identify discrepancies between social objects and general © 1997 Tony Lawson
beliefs about and expectations of, o r relevant to, those objects and also to provide an explanation of such discrepancies, i.e. to identify the social causes responsible. When this is achieved the basis is clearly laid for the possibility (although it will never be sufficient for the realisation) of rational, intentional, social transformation, in particular through a modification of the practices on which the structures of relevance depend. It is this scenario, this objective of changing features or aspects of society through transforming the conceptions on which they depend, that, in part at least, underpins all radical contributions. It is the basis, for example, of Marxian analyses of capitalism, of environmentalist analyses of environmental destruction or degradation, or of feminist analyses of gender relations and hierarchies.
CONDITIONS OF SOCIAL EMANCIPATION Social emancipation, then, is an objective which critical realism, unlike positivistic orthodox economics, is able to embrace. Even so, an acceptance of critical realism is obviously nothing like sufficient for any progressive change actually to take place. Various conditions for such change are essential, the more fundamental of which can now, if briefly, be considered. The first is the existence of shared human objectives, i.e. real interests, needs and motives. Clearly if everyone’s needs are, for example, merely subjective, with the possibility of being irreconcilably opposed, then projecting the goal of social emancipation is likely to be question begging from the outset. But this requirement of shared real interests is not merely a condition of social emancipatory action as conceived of here, it is a presupposition of all policy analysis, whatever perspective is accepted. Every national government, for example, professes to act on behalf of ‘its people’ supposing that certain interests of the people of the relevant country are the same. To be sure, such claims and conceptions warrant unpacking. But all programmatic stances with respect to (changing or sustaining) any aspect of the economic and social order presuppose some implicit moral realism, some presuppositions of shared real interests of all, or a group of, individuals which exist by virtue of a common human genetic make-up, or as a result of shared occupancy of specific societal, including national, positions. As we saw in Chapter 13, the basis for a moral realism has indeed been accepted here. That is, the possibility of moral theorising is grounded in a recognised common human nature. A conception has been sustained which acknowledges the existence of rights of all human beings as human beings, by virtue of a common nature grounded in our biological unity as a species. Of course it has also been argued that this common nature is always historically © 1997 Tony Lawson
and socially mediated, so that needs will be manifest empirically (as wants) in a potential multitude of ways. It follows that the pursuit of change will always take place in a context of conflicting position-related interests. Certainly, conflicts centring on the interests of class positions, age, gender, nation states, and so forth, are as real and determining as anything else. Even so, different groups may cooperate and persistently reproduce relations of trust, and so on. But opposed, position-related, interests exist. And it may be upon our unity as a species and the more generalised features of our social and historical experience and make-up, that the greater possibility of unambiguous and more enduring progress rests. Second, the possibility of emancipatory social change requires not merely shared real needs or interests but, equally, a self-awareness of these needs and interests. Needs and rights can be formulated as goals or wants or demands, and treated as legitimate or illegitimate, only under definite historical conditions. As such they may be poorly, and even misleadingly, formulated. Specifically, real needs can be manifest in a variety of historically contingent wants, which may then be met by any of perhaps a multitude of potential satisfiers (Doyal and Gough, 1991). To assume that either actual satisfiers, for example specific commodities purchased, or expressed objectives, such as owning more than others, are defining of human needs is to commit the epistemic fallacy all over again, albeit this time in an ethical form—to reduce needs to wants and wants to the conditions of their being satisfied or expressed. I am not suggesting that wants as expressed in actions bear no relation to underlying needs, of course. Indeed, although certain activities sometimes appear quite undesirable from the point of view of facilitating human development and potential, it is often easy enough to see how they are nevertheless motivated by various real needs on the part of the perpetrators— for example, to obtain respect from others, inner security or simply a release of frustration. But the two, real needs and expressed wants, should not be conflated. Yet in contemporary orthodox economics the individual’s needs and rights are precisely collapsed onto preferences, tastes and wants, and measured by the purchases made or whatever. Where, for example, people abuse either themselves (excessive drinking, harmful drugs) or others, economic orthodoxy can only treat such actions as the direct expression of preferences, rather than the often complicated and masked manifestation of deeper needs. Little basis exists or can even be allowed for the drawing of such a distinction. We can note in consequence that the numerous mainstream attempts to measure and thereby assess human well-being in different contexts according to simple indices of poverty, levels of income, or whatever, are seriously open to question and reformulation. Now it should be clear, although I emphasise it anyway, that I am not here suggesting that we replace the orthodox insistence on the sovereignty of © 1997 Tony Lawson
individual preferences with any presumption of the sort that scientists or policy-makers know (or could know) the choices that people should make. Even less am I suggesting that the latter group, or anyone else, be empowered to take more decisions on behalf of others. Rather, the implication of the foregoing is that science aims to provide a more adequate understanding of real needs in order to inform the choices that individuals themselves make (possibly in some informed, democratic forum). For, to repeat, a condition of emancipatory social change is a self-awareness of real human needs and interests.2 A third condition for emancipatory, or indeed of any intended change is that the goal in question constitutes a real possibility. Counter-factuals are legitimate in policy discussions if they express real goals or options. Then they can serve as guiding conceptions to action. This seems obvious enough, but is an issue rarely addressed in contemporary economics. In particular, it is rarely questioned whether the usual economic goals—zero unemployment and/or price inflation, a balance of payments ‘equilibrium’, economic growth, and so on—are attainable under the current (or any other realistically achievable) institutional system. Rather, competing policy-oriented groups usually accept without question that definite goals are achievable and spend much of their time arguing the advantages of their own strategies compared to those of others, usually in terms of ‘whenever this then that’ conjectures. There are exceptions to this; the more visionary economists certainly question the limitations of what can realistically be pursued within existing institutional and relational conditions. Such a stance does not reflect any specific political outlook. It is as true, for example, of (the later) Hayek as it is of Marx and, at times, of Keynes. For Hayek and Marx I think this is clear. But it applies much of the time to Keynes as well. Consider, again his analysis and explanation of stock market volatility/stability as presented in Chapter 12 of the General Theory (and briefly referred to in Chapter 13 above; see also Runde 1997). Essentially three conditioning or explanatory factors are identified: the prevalence of psychological motivations, the nature or structure of uncertainty regarding future yields and the organisation of modern investment markets with a view to facilitating ‘liquidity’. In order to reduce or eradicate such inherent volatility as there is, little can be achieved with regard to psychological factors—although government spokespeople do attempt to talk up (or down) relevant market prices when the situation so demands. Even less can be done to counteract the fact of fundamental uncertainty. Not surprisingly, Keynes’ policy-orientated comments are largely directed not at the amelioration of events, but at the necessity of a quite revolutionary reorganisation of the basic institutional structure governing the markets for existing investments.3 Two further conditions of emancipatory change can briefly be mentioned here. I suggested above that emancipation may be facilitated by explanatory © 1997 Tony Lawson
critique. This itself presupposes not only that discourse is ontologically efficacious, i.e. that others can be persuaded (Bhaskar, 1989:113), but also that values such as truth, freedom and justice are immanent and widely accepted. In their absence normative discourse is quite futile. A final condition is an inclination to act in accord with real interests, if indeed they can be identified correctly and their pursuit rendered practically feasible. In any actual situation, such action may involve a large cost, a degree of risk, or just an effort of will. Not every person grasps every opportunity that is recognised as coming his or her way. Rational, intentional, emancipatory change, then, is unlikely very often to be a simple matter, even if we can now acknowledge it as a real possibility. Notice too that emancipation or freedom is not reducible to human choice. The latter is certainly a condition of freedom but it is by no means the same thing. Even condemned people can affect the manner in which they meet their fate. Nor are choice and freedom undermined by a recognition of necessity. Thus Reder’s (1982) assertion, that ‘Freedom...consists not in power of choice, but (pace Hegel) in recognition of necessity’ (p. 35), is incorrect if Reder’s intended implication is that power of choice is something at odds with a recognition of necessity. It is further erroneous if it presumes that freedom is merely the recognition of necessity. At the very least freedom requires knowing real interests, being in a position to act to realise them, and being disposed to do so. In this the power of choice and recognition of necessity are clearly necessary but hardly sufficient conditions. In short, from the perspective here argued for, a transformational, rather than (or rather than merely) an event-ameliatory, conception of intentional change is sustained and the goal of emancipation, in place of control, emerges as a real possibility. Of course, any project that aims to be emancipatory is likely to face difficulties. But once the blinkers of positivism are thrown off it must become easier to adapt the policy process to exploring the real possibilities for human betterment. In short, if the cost of accepting the framework here elaborated is an abandonment of much of the output of the contemporary discipline of economics, the gain includes not only the possibility of an emergent science of economics, but a firmer basis, a more appropriate and coherent framework than hitherto possessed, for exploring how to make the world a more secure, facilitating and empowering place, more at one with our liabilities and potentialities as needy, creative and purposive social human beings.
© 1997 Tony Lawson
20 ECONOMIC SCIENCE AND PREDICTION
THE BASIC THESIS The argument of this book has been formulated with the intention of being critical but also constructive. The objective has been not only to reveal the inherent limitations of contemporary mainstream economics, and to explain the discipline’s state of disarray, but also, and especially, to sketch the structure of a more sustainable, potentially fruitful, alternative approach or perspective. My central thesis is briefly stated. The essence of contemporary mainstream economics does not lie at the level of substantive theory as most of its critics suggest, but at the level of methodology. Specifically, the most fundamental feature is a generalised insistence on the deductivist mode of explanation, including an unsustainable commitment to the ‘whenever this then that’ structure of ‘laws’. And it is in this very essence that the perpetual disarray of the subject is rooted. For, while the generalised usefulness of deductivism is dependent upon a ubiquity of closed systems, the social world, the object of social study, is fundamentally open and seemingly insusceptible to scientifically interesting local closures, or at least to closures of the degree of strictness that contemporary methods of economics require. The ultimate source of all the problems is the epistemic fallacy, the belief that questions of ontology can be reduced to questions of epistemology. In the writings of Hume this leads to reality being reduced to the course of events given in experience. And with reality so contained the ‘whenever this then that’ conception emerges as the only form of scientific generality or ‘law’ that can be sustained. In this way, the real is collapsed onto the actual which is anthropocentrically identified with a human attribute.1 I do not wish to suggest that in accepting deductivism and/or the associated conception of science, mainstream economists are in agreement with the © 1997 Tony Lawson
manner of its derivation from Humean philosophy. Indeed, many have probably never given any thought to the question of how this conception of science is grounded. At the same time it is difficult to believe that many do not at some level accept the richer ontology of structures, change, emergence, and internal relationality, etc., that has been systematised under the heading of critical realism. But if such is the case, the charge against mainstream economists must be that of complicity, of agreement to accept the results of the Humean project without questioning their justification. Most obviously this is manifest in a widespread disdain for methodology and in particular in a total neglect of explicit attention to ontology in formal contributions. This serves to deflect attention from the unbridgeable schism between the world view sustained by critical realism and that of atomism and closure, presupposed by mainstream deductivist methodology. By persevering with a conception of science without explicitly questioning the nature and relevance of its ontological presuppositions, mainstream economists are in effect committing their own scientistic fallacy. Something though has to give way. If our best theory of ontology is inconsistent with the officially sanctioned practices of economists, at least one of them, ontological theory or sanctioned practice, has to be transformed. If it were the case that the deductivist project in economics had proven ex posteriori to be successful, on its own terms or more widely, there would be some grounds for persevering with it. But it has not. What is more, critical realism can situate and explain those successes which deductivism has achieved in the natural sciences as well as account for the failure of the deductivist programme in economics. The choice then is ultimately between contemporary mainstream economics and reality. It is largely the dogmatic refusal to focus explicitly on questions of methodology2 that hinders a generalised recognition that this is the real choice. And this refusal serves, in turn, to sustain the deductivist faith that it is only a matter of time before success is achieved, that it is but a question of (repeatedly) trying harder. It must by now be clear that I am arguing for nothing less than orthodox economics’ effective demise (as a general approach and mainstream position). This orientation follows inescapably once it is recognised that the source of all the project’s problems and difficulties (in dealing with an open system) stem from that project’s very essence. The optimistic message, however, is that if the choice is made in favour of reality rather than deductivism, economics as science properly conceived emerges as a real possibility. For the social world has been found not only to be open but also structured, consisting in definite and identifiable social structures, powers, mechanisms and tendencies that are responsible for the actual course of social events and states of affairs. In accepting the challenge of seeking to illuminate them, economics can yet be a science in the sense of any other science. © 1997 Tony Lawson
OMISSIONS At this stage I should acknowledge briefly the ways in which the account I have provided here is incomplete. Most obviously, I have not attempted to trace the precise path whereby the mainstream economics project emerged; nor have I explored all the mechanisms whereby it is reproduced. No doubt there are numerous contributory factors to both aspects. Ideology must come into it: amongst other things the emphasis of orthodoxy upon surface phenomena rather than the deeper structures of society, including social relations, sustains a view of society as rather more harmonious than it is (although it is important not to adopt too functionalist or conspiratorial a view prior to a fuller historical investigation). A further likely contributory factor of significance is the inspirational role of mathematics in western culture, certainly since the Enlightenment, and especially amongst economists (there were of course numerous attempts to mathematise the discipline long before the apparent successes of Walras and others—especially in France). Currently, new students to economics are not so much taught the supposed virtues of deductivism directly, as immersed in mathematical techniques to an extent which presupposes the widespread relevance of deductivist method. Any justification for this training appears to rest on little more than an uncritical enthusiasm for formalism on the part of those who direct the discipline (individual academics who, by and large, appear quite blind to the fact that specific formalistic systems are necessarily limited in their scope of application).3 But these are not matters I consider further here, significant though they obviously are.
CONSEQUENCES Nor do I think it necessary that I summarise further the main steps of the argument, some of which have been repeated already. Rather, it is likely to be more useful if I concentrate upon drawing out various central implications of the argument. And I think that this is most usefully achieved by my going through those aspects or consequences as they bear upon that most favourite of economists’ pastimes: the activity of event prediction or forecasting. For it is this activity, more than any other, that distinguishes the applied and policy orientations of economists from those of other social scientists, and it is with respect to this activity, as much as any, that the discussion that has gone before bears fundamental implications.
© 1997 Tony Lawson
THE PREOCCUPATION WITH PREDICTION The situation to which I am referring is the near universal practice in economics of tying conceptions of economic science to the possibility of successful prediction. This supposed association of science with successful prediction is insisted upon as much by economists who reject the possibility of economics as science as by those who think that economic science, so conceived, has already been achieved. Examples abound of explicit assertions to this effect. But let me merely sample from winners of the Nobel Memorial Prize in economic science. The list of prize-winners who suppose that a predictively successful science has actually been achieved includes not only such well known cases as Friedman but also perceived critics within orthodoxy such as Allais. The group which concludes that economic science is impossible precisely because successful prediction is not realisable includes Hicks, a former ‘theorist’ who seems at some stage to have abandoned the orthodox project entirely. Consider, for example, the following assessments: The essential condition of any science is the existence of regularities which can be analyzed and forecast. This is the case in celestial mechanics. But it is also true of many economic phenomena. Indeed, their thorough analysis displays the existence of regularities which are just as striking as those found in the physical sciences. This is why economics is a science, and why this science rests on the same general principles and methods of physics. (Allais, 1992:25) A science, I would say, consists of a body of propositions, which have three distinguishing characters. (1) They are about real things, things we observe, about phenomena. (2) They are general propositions, about classes of phenomena. (3) They are propositions on which it is possible to base predictions, predictions which command some degree of belief. All of these three appear to be necessary, if the body of propositions is to be a science... (Hicks, 1986:91) [Economics] is no more than on the edge [of science], because the experiences that it analyses have so much that is non-repetitive about them. (ibid.: 100) It is instructive to focus briefly upon Hicks’ assessment somewhat further. In truth Hicks has little to say on his premises (1) and (2) (except in relation to the question of whether mathematics is a science), considering them to be relatively uncontroversial. In consequence, he concentrates almost wholly © 1997 Tony Lawson
upon his premise (3), on the supposed necessity, in science, of generating successful predictions. In doing so Hicks distinguishes: (a) conditional from unconditional prediction—where the former, unlike the latter, is qualified by the stipulation that some known condition must be satisfied; (b) strong from weak (conditional) prediction—where the former entails that, given the stated conditions, a specified event will follow; the latter that, given the stated conditions, the specified event will follow if there are no disturbances. In short, Hicks asserts the following prediction typology where x and y are observable phenomena: 1 y will happen—unconditional prediction 2 If x then y—strong conditional prediction 3 If x then y if no disturbances—weak conditional prediction. Hicks associates the lay person’s view of science with (1), Popper’s (falsificationist) view of science with (2), and his own view of science with (3). Economics according to Hicks fits most closely with (3). But because it deals with phenomena of a highly non-repetitive nature economics is considered to be no more than on the ‘edge of science’ (and equally on the ‘edge of history’). Hicks’ account apparently resonates with many, and this is doubtless because it does capture insight. Certainly the possibility of weak prediction appears more plausible than that of strong prediction. The problem with accepting Hicks’ position is not only that the possibility of economics as science (as Hicks conceives of it) is ruled out, but also that it is not obvious that economics remains of much use at all. For if the objects of interest and relevance in economics are observable phenomena, and if regularities of the form ‘whenever event x then event y’ are subject to ceteris paribus clauses which rarely hold in any interesting form, what is the subject’s practical worth? It is easy to see that the distinction Hicks recognises between the notions of strong and weak (conditional) prediction is relevant to certain strands of debate in economics. If orthodox economists presuppose the possibility of strong conditional prediction in their substantive contributions, members of the (typically more philosophically reflective) non-orthodox schools tend to emphasise the possibility only of weak conditional prediction. Of course, if challenged to explain themselves, or to defend current failures, orthodox economists can also be found sliding into accepting only the possibility of weak conditional prediction. But the issue is then whether disturbances are likely to be pervasive. To admit that they are is tantamount (in this © 1997 Tony Lawson
predictionist framework) to accepting the case for abandoning substantive work more or less altogether. In consequence, the compromise usually effected is that the rhetoric of weak conditional prediction remains while the presupposition of strong conditional prediction creeps surreptitiously in via the back door. Of course, the few who consistently follow the logic of the argument such as put forward by Hicks, invariably do end up, along with various subjectivists or hermeneuticists, including various Austrians, and Shackle, abandoning any significant role for economics at all. In short, whatever the reasoning or rhetoric, the perceived choice in modern economics is quite often between an acceptance that strong conditional prediction is possible or a resignation that economics can achieve very little—between the poles typified by the assessments of Allais and Hicks respectively. The position argued for in this book, in contrast, is that this ‘choice’ reflects a false dichotomy, which arises only because the basic conception of science held by economists is flawed. However, to return to Hicks’ assessment, the error creeps in not just with his discussion of prediction, but also in his formulation of the two associated premises: that science is restricted in its concern with observable phenomena, and that the object of science is to obtain generalisations about classes of such phenomena. Rather, as we have seen throughout, science is primarily concerned to identify and illuminate the structures, powers, mechanisms, processes and tendencies that produce or facilitate such actualities as the events, including human actions, that we experience. What, then, follows with respect to the perpetual endeavours of economists to forecast successfully? Three conclusions of significance can be stated:
Success at economic forecasting is unlikely If the predictive goal is the successful forecasting of scientifically significant economic events or states of affairs then an implication of the preceding discussion is that such a goal is likely to be only rarely realised, at least in an unqualified form. Prediction of non-experimental events rests upon spontaneous occurrences of constant event conjunctions which, as Hicks observes, are not widely in evidence in the social realm and seem unlikely to be. This insight does not obviate the possibility of predicting tendencies. I refer only to the (unconditional) prediction of events. There may well be predictable tendencies for, say, leaves and profits to fall even when (in specific instances) leaves and profits actually rise. It may also be possible, on occasion, to anticipate the limits to, or bounds on, the range of realisable outcomes. If, for example, income taxes are reduced, individuals within the economy can either spend (some or all of) their increased income or save it. It is thus © 1997 Tony Lawson
conceivable that the broad consequences of it being entirely spent, or all saved, can be reflected upon and so the range of possibilities determined. Of course, other context-specific mechanisms may be known. Perhaps the tax reductions are aimed at the very poor who have little option but to spend any additions to their incomes. It all depends on a host of contingencies that cannot be foreseen. The one clear implication is that without attention to context-specific structures and mechanisms, there can be little rational basis for supposing that x will follow y on this occasion merely because that is what happened before.
Success at economic forecasting is not essential If event prediction is usually infeasible it is in any case not required for a successful science of economics. For it can now be accepted that the primary aim of science is not the illumination or prediction of events at all but the identification and comprehension of the structures, powers, mechanisms and tendencies which produce or facilitate them. And this understanding is all that is required for policy analysis and (where feasible) effective action. Hutchison (1994:30) is therefore quite wrong to suppose that to abandon the aim of event prediction necessitates a ‘form of outright intellectual nihilism’. It is not, for example, the prediction of the pattern of spots on a patient’s skin that is the ultimate goal of medical research but the identification of the virus or agent that is causing it and the production of, or at least search for, an effective antidote. Although the prices of financial assets and interest rates, influenced as they are by such factors as uncertainty, whim, sentiment and chance, may be more or less anything, the conditions which encourage or facilitate the (unwanted) instability of financial markets are issues that, if Keynes (1973a) is correct (see Runde, 1997), can nevertheless be identified, and the question of an efficacious policy response considered. Similarly, economics can aim both at identifying mechanisms which contribute to the emergence and continuance of worldwide poverty, unemployment, discrimination, and indeed all situations in which the production and exercise of human potential is denied or severely limited, and also be put to work in the formulation of effective responses, such as alternative institutional structures, including relationships. In short, with science now more adequately conceived, it follows that, contra Hicks, economics need not be merely at the ‘edge of science’ at all. Indeed, economics can be a science in precisely the sense of natural science, albeit depending upon a wholesale transformation in the practices which economists employ.
© 1997 Tony Lawson
Success at forecasting economic events is not wanted If event prediction is neither possible nor necessary, it is equally undesirable. For the possibility of successful prediction, turning as it does on the existence of constant conjunctions of events, would mean either that the future is already determined, or, if exogenous variables could be fixed by us, open to social control. Either way the situation would be inconsistent with the possibility of generalised human choice and freedom. An attraction of the theory of reality and science defended here is that human choice, and indeed emancipation are sustained as real possibilities. From this perspective it follows that policies and strategies can be formulated with the objective not merely of ameliorating events and states of affairs but also of replacing structures that are unwanted by others that are needed and empowering, of facilitating a greater or more desirable or equitably distributed range of human opportunities. Rational, intentional, emancipatory, real change is no longer found to be, as in positivism, in contradiction with the explanatory function of science including economics—indeed, it is recognised as being a very condition of science, properly conceived. Rather, critical realism provides a perspective on science, nature, society and economy that is not only explanatorily powerful but also able to preserve the intuition that human social history is explicable and yet actively made.
© 1997 Tony Lawson
NOTES
1 ENDURING TENSIONS AS POINTS OF DEPARTURE 1 2
So dominant is this project, in fact, that its proponents and many others often just refer to it as ‘economics’ (see Gee, 1991). Nor, incidently, is Leamer the only top-rate econometrician who is prepared publicly to air these sorts of concerns and attempt to respond to observed theory/ practice inconsistencies (although there are not many others). An important additional voice is that of David Hendry. Indeed, Hendry, like Leamer, is unusual in being prepared to discuss all aspects of methodology. Interestingly, in fact, a methodological discussion between these two and Dale Poirier has been published recently in Econometric Theory (Hendry et al., 1990). And, amongst the numerous topics and themes elaborated upon and debated, the continuing prevalence of econometric theory/practice inconsistencies is reaffirmed by Leamer and also acknowledged by Hendry. Leamer observes: There is a great discrepancy between the popular theories of statistical inference and actual practice. What I’ve tried to do is to think about that discrepancy, to think about the issues raised by actual practice as compared to theory, and either to alter the theory to conform with those aspects of practice that I thought of as desirable or sensible, or alternatively, to make recommendations as to how practice ought to be altered, ought to be policed, and made more effective. (Hendry et al., 1990:178)
On the following page Hendry concurs: At present there are peculiar gaps between theory and what people actually do: I think the sinners and preachers analogy in Leamer [1978] is the correct one here. The theoretical econometrician says one thing but as a practitioner does something different. I am trying to understand why economists do that, given that they know the theory, and they are obviously trying to solve practical problems. (ibid.: 179) 3 4
In fuller form Duesenberry’s well-known aphorism is that ‘Economics is all about how people make choices. Sociology is all about why they don’t have any choices to make.’ This attitude is not new of course. The Royal Economic Society Newsletter, for example, recently (October 1992) reproduced the opening paragraph of living Fisher’s Presidential Address to the American Statistical Association in December
© 1997 Tony Lawson
5 6
1932, with its assertion to this effect (see Chapter 18 below, note 15). For a recent critical discussion of this attitude see Hoover (1995). This is the main grant-providing body for UK post-graduates in economics. To repeat Hoover’s (1995b:733) apposite paraphrasing of Keynes: ‘Practical economists, who believe themselves to be quite exempt from any methodological influences, are usually slaves of some defunct methodologist.’
2 REALISM, EXPLANATION AND SCIENCE 1
2 3 4
The basis of their adoption is often merely that they are perceived as having been successful in some other realm or context, or some such. It is as though, for example, the pneumatic drill which has facilitated digging up the road can thereby be immediately accepted as an appropriate tool for cleaning the glass window without regard to (our knowledge of) the nature of glass. As, for brevity’s sake, I shall frequently refer to the formulations in question. This ‘symmetry thesis’ is insisted upon by ‘Popperians’ such as Blaug (1994:123). This is basically a rather severe adaptation and simplification of the following well-known example provided by Hempel (1965 [1942]): Let the event to be explained consist in the cracking of an automobile radiator during a cold night. The sentences of group (1) [i.e. the ‘set of statements asserting the occurrence of certain events...at certain times’] may state the following initial and boundary conditions: The car was left in the street all night. Its radiator, which consists of iron, was completely filled with water, and the lid was screwed on tightly. The temperature during the night dropped from 39°F, in the evening, to 25°F in the morning; the air pressure was normal. The bursting pressure of the radiator material is so and so much. Group (2) [i.e. the ‘set of universal hypotheses’] would contain empirical laws such as the following: Below 32°F, under normal atmospheric pressure, water freezes. Below 39.2°F, the pressure of a mass of water increases with decreasing temperature, if the volume remains constant or decreases; when the water freezes, the pressure again increases. Finally, this group would have to include a quantitative law concerning the change of pressure of water as a function of its temperature and volume. From statements of these two kinds, the conclusion that the radiator cracked during the night can be deduced by logical reasoning; an explanation of the considered event has been established. (Hempel, 1965 [1942]:232)
5
Notice that although many economic theorists do not in fact confront their ‘models’ with the data of experience this does not detract from the observation that they adhere to the deductivist explanatory form in question. Indeed, as we shall see in Chapters 8 and 9, their economic ‘theorising’ is usually sustained on the promise of obtaining models that can be confronted with data eventually—as a result of ‘trying that little bit harder’ to overcome current problems. Nor is this situation peculiar to economics, even within the ‘social sciences’. Nor, furthermore, is the acceptance of deductivism in economics particularly contemporary (even if the extent of this acceptance is). As Miller (1987) recently observed:
© 1997 Tony Lawson
The pioneers of modern academic sociology, anthropology and economics, Weber, Durkheim, Radcliffe-Brown, Malinowski, Menger, Jevons and Walras, all regarded subsumption under general laws as essential to scientific explanation, and took the discovery of such laws to be the means for making the social sciences truly scientific. Commit-ment to the coveringlaw model has kept such goals alive in the face of continual disappointment. For example, when social anthropologists discovered that their fieldwork yielded few interesting general laws involving relatively concrete phenomena, such as motherhood or farming, they did not abandon the pursuit of general laws. Many responded by seeking such general relationships among more abstract structural characteristics, such as ‘binary opposition’. Many economists elaborate the internal logic of some general model, serenely accepting that their work makes no appreciable contribution to explaining specific episodes of inflation or unemployment, or specific international economic relations. The intellectual justification is, basically, that the elaboration of general models is the most promising route to the discovery of general laws, an essential aspect, in turn, of explanations. (Miller, 1987:28) 6 7
For a similar statement also see Allais (1992:25). For interpretations similar to, or inclusive of this one, see entries on positivism in Bullock, Stallybrass and Trombley, 1988; Gould and Kolb, 1964; or Outhwaite and Bottomore, 1993. Also see the discussions, especially in Bhaskar, 1978; and in Miller, 1987. I should acknowledge, though, that while a reasonably consistent collection of themes tends to be systematised in social theory as positivism there is no unanimity over which are, or have been, of greatest historical moment or cognitive importance. In the social sciences, perhaps three (related) principles are most frequently singled out, vis phenomenalism (an ontological theory according to which knowledge is founded on experience alone); the methodological unity of science; and the axiological principle that facts are rigidly separate from values, and normative statements cannot achieve the status of knowledge. Although many themes included under ‘positivism’ have been around at least since Aristotle the term was first systematically employed by Saint-Simon. It was adopted by Auguste Comte when expressing the claim that the world consists of phenomena which are real, certain, precise and useful, and that knowledge consists only in the coexistence and succession of such phenomena. Comte’s aim was to instigate a naturalistic science of society capable of explaining humanity’s history and of predicting its future employing those methods of enquiry (comparison, experimentation and observation) that were regarded as successful in the study of nature. In his Course of Positivist Philosophy Comte (1830–42) argues that the human spirit has evolved though three necessary stages. In the ‘theological or fictitious’ period, phenomena are explained in terms of supernatural entities, in the ‘metaphysical’ stage by ‘abstractions’, and in the ‘scientific or positive’ stage by ‘laws’ of facts interpreted as the ‘invariable relations of succession and resemblance’ which connect them. Comte coined the term ‘sociology’ for the science intended to synthesise all positive knowledge, reveal the workings of society and guide the formation of the positive polity. From the mid-nineteenth century onwards, positivism became an extremely influential intellectual and cultural trend forming the basis of the most widely accepted view of science. In France, Durkeim abandoned Comte’s philosophy but retained his method, insisting on the affinity of social and natural science; in
© 1997 Tony Lawson
Britain a more radically empiricist version was developed by J.S.Mill; while in Austria the Vienna Circle (including Schlick, Mach, Carnap, Hemple and Neurath) attempted a synthesis of Humean empiricism and logical analysis that would forever rid philosophy of ‘metaphysical speculation’ by grounding all knowledge in experience. For a discussion of many of these and related developments see Caldwell (1982). Somewhat ironically, although the term ‘positivism’ was originally coined as a term of affirmative self designation and heralded by the Vienna Circle as the ‘philosophy to end all philosophy’, it appears today to be used (as here) only negatively, i.e. by those concerned to assess critically, or situate some chosen theme. Indeed, it is a standing joke in some quarters that there are nearly as many versions of positivism as there are critics. The plethora of interpretations no doubt arises precisely because the term is no longer used by anyone as an affirmative description or prescription. In consequence, the varying usages tend to reflect that set of aspects or components of positivist themes which each critic regards as most essential. The advantage of the conception sustained here is that, as Bhaskar notes, it ‘describes a pole not merely in relation to which other philosophical positions may be defined, but a pole in relation to which they have had typically to defend and justify themselves’ (Bhaskar, 1986:229). In any case, whatever the manner in which positivism is characterised, the historical significance of Humean empiricism on its development is undeniable. 8 Of course, in proceeding in this fashion, albeit only for illustrative purposes, I am taking it for granted, for the time being, that there exists a separate sphere which can be designated the social realm. This is a preumption that will be justified explicitly in due course. 9 Of course, aspects of the theory in question, and in particular the conception of reality as consisting in part of structures, powers and tendencies, etc. are hardly original to Bhaskar’s project. However, to my knowledge it is within the project stylised by Bhaskar as transcendental realism that the philosophical ontology in question has been most explicitly and effectively systematised, furthest developed and rendered most coherent. It is thus fair to accept Bhaskar’s heading here. (The reason that the heading employed by Bhaskar is transcendental realism will become clear in Chapter 5.) 10 And, in the natural sphere at least, it is in the notion of a generative mechanism at work that natural necessity (i.e. necessity independent of people and their actions) lies. 11 The term ‘governing’ is intended to be generic here. Its interpretations include ‘influence’, ‘facilitate’, ‘produce’, etc. 12 At the time of writing (1995) the best explanation of the phenomenon, i.e. of ‘mad cow disease’ or bovine spongiform encephalopathy (BSE), and indeed of the apparently related fatal human brain disease Creutzfeldt-Jakob Disease (CJD) as well as scrapie in sheep, is not something like a virus or bacteria (as analogical reasoning might lead us to expect). Rather it is a prion, a ‘rogue’ form of ‘normal’ protein that apparently spontaneously changes shape, and in so doing (1) becomes immune to break-down by enzymes in the fashion experienced by other proteins, and (2) carries the potential to induce other proteins similarly to change their shape. In consequence there can be a build-up of diseased proteins in the body resulting in the body’s self-regulation system being unable to function adequately. If this theory is correct there is much about the working of the agent and the transmission mechanism of the disease that is not understood. For example it is not fully understood why proteins should change their shape, how © 1997 Tony Lawson
many diseased proteins are required for a chain reaction to be initiated or how long the whole process takes. And there remains uncertainty whether diseased proteins that have passed from cows to the human food chain can cause the normal human protein to change shape. The whole episode, though, does provide a very clear example, one very much in public view, of scientists not resting content with manifest symptoms or correlations but seeking to uncover agents and mechanisms which are responsible for these actualities. 13 Notice that with this conception an event is no longer to be interpreted, as with Hume’s objects of sense experience, as necessarily atomistic. Rather, I think they are best thought of as inherently transitory, unitary occurrences or actualities, as possibilities rendered actual, which are essentially related to their multiple determinants. Whitehead (1926) appears to hold to a similar conception: The general aspect of nature is that of evolutionary expansiveness. These unities, which I call events, are the emergence into actuality of something. How are we to characterise that something which emerges? The name ‘event’ given to such a unity draws attention to the inherent transitoriness, combined with the actual unity. (Whitehead, 1926:131) 14 This (Aristotelian) mode of inference is interpreted as retroduction by, for example, Peirce (1867), Hanson (1958), Bhaskar (1978), McMullin (1984), and as abduction by Peirce (1867) and various Institutionalist economists. Peirce is probably the most important contributor here. Adopting a scientific realist perspective which in many ways is consistent with the position to be developed below (for an overview see Hoover, 1994) Peirce divides inference into ‘explicative inference’, including deductive logic and mathematics, and ‘ampliative inference’, including induction and abduction. Concerning the latter (which he also terms hypothesis and presumption) Peirce observes ‘abduction is Originary in respect of being the only kind of argument which starts an idea’ (1867:2:96). Elsewhere he expands: [Induction] never can originate any idea whatever. No more can deduction. All the ideas of science come to it by the way of Abduction. Abduction consists in studying the facts and devising a theory to explain them. Its only justification is that if we are ever to understand things at all, it must be in this way. (Peirce, 1867:5:145) [Note that references to Peirce of the form (x: y) indicate his Collected Papers, volume x, paragraph y.] 15 Thus, in this framework usage of a ceteris paribus clause, the qualification that other things are (or must be) equal, does not apply to a law’s application, properly interpreted. If a mechanism is triggered, a tendency is in play, and a law can be cited, unconditionally. The satisfaction of the ceteris paribus clause, rather, is a condition for the actualisation of the tendency designated in the statement. It acts as a reminder that the system in question may not be closed, that the tendency postulated in a law-statement may not act in isolation and so be fulfilled. Once we use the category of tendencies to denote the activity of transfactual mechanisms, and not events that would come about if things were different, it is clear that, in principle, the ceteris paribus clause can be dispensed with entirely (see also Pratten, 1994). 16 It is important to be clear that intransitive objects need be no more fixed or enduring than transitive ones, including thought itself. Indeed, if I am attempting © 1997 Tony Lawson
to study the thought of a second person, that person’s thought constitutes an intransitive object of my own study (see Parts III and IV below). 17 As Peirce (1867) observed more than a century ago ‘there is but one state of mind from which you can “set out”—a state in which you are laden with an immense mass of cognition already formed...’ (1867:5:416). 18 As Bhaskar formulates it: ‘On the transcendental realist view of science, then, its essence lies in the movement at any one level from knowledge of manifest phenomena to knowledge, produced by means of antecedent knowledge, of the structures that generate them’ (1989:20). 19 Actually, it is presumably the case that what we treat as immediate sense experience must also be understood in this sort of way. The point here is that all perception is ‘theory laden’; there are no pure uninterpreted facts. When we see, hear, touch, taste or smell something in a definite context I think we retroduce instantly (and without choice) a phenomenon that would account for our sensation, with the result that we perceive the sensation holistically just as a recognisable phenomenon. A similar view appears to be maintained by Peirce (1867:5:181, 183, 291, 584–6)—see also Hoover (1994:301).
3 THE CASE FOR TRANSCENDENTAL REALISM 1 2
Of course, at least the first of these ‘observations’ has been repeatedly noted in the philosophy of science literature. Still, I think their full implications as discussed below have rarely been appreciated. The nature of the argument employed here is explored in detail in Chapter 5.
4 THE LEGACY OF POSITIVISM 1
Put differently, choice is not consistent with the orthodox emphasis upon prediction. We have seen earlier that in the deductivist framework explanation is symmetric with prediction. In this framework, if to deduce an event before it has been observed is to predict it, to do so after it is observed is to explain it. Recognising this, Buchanan also acknowledges that it is the conception of science in question that defeats the objective of allowing for the human power of choosing: ‘Choice, by its nature, cannot be predetermined and remain choice. If we then define science in the modern sense of embodying conceptually refutable predictions, a science of choice becomes self-contradictory’ (Buchanan, 1979:40).
5 THE NATURE OF THE ARGUMENT 1
2
McCloskey (1986) for example writes of ‘methodological authoritarians’ (1986:36) and asserts that ‘Methodology is Bourgeois’ (ibid.: 24), suggesting that ‘If it were not so damaging to the sense, Methodology strutting around issuing orders to working scientists, would only be funny’ (ibid.: 24). On all this see Sofianou, 1995. According to Weintraub (1989) ‘any normative role for Methodology rests upon a profound misconception, and thus Methodology cannot possibly have
© 1997 Tony Lawson
3 4 5
consequences for the way economics is done. Methodology...cannot have any impact on the manner of practice’ (1989:478). This point has been well made in Lawson, C., Peacock and Pratten (1996) where a position similar to the one developed below is taken in the context of ‘underlabouring for an economics with institutions’. Any reader impatient to see how the argument develops, however, can easily skip this chapter and perhaps return to it later. Thus Weintraub (1989) drawing upon Fish’s (external?) injunctions from Literary Criticism concludes that ‘there is no position totally apart from the doing of economics which can inform the consideration of the doing of economics’ (1989:486). In supporting this result Weintraub reproduces the following passage by Fish, inviting us to insert the term ‘Methodology’ where Fish uses ‘Theory’: Theory can be seen as an effort to guide practice in two senses: (1) it is an attempt to guide practice from a position above or outside it..., and (2) it is an attempt to reform practice by neutralizing interest, by substituting for the parochial perspective of some local or partisan point of view the perspective of a general rationality to which the individual subordinates his contextually conditioned opinions and beliefs.... The argument against theory is simply that this substitution of the general for the local has never been and will never be achieved. Theory is an impossible project which will never succeed. It will never succeed simply because the primary data and formal laws necessary to its success will always be spied or picked out from within the contextual circumstances of which they are supposedly independent. The objective facts and rules of calculation that are to ground interpretation and render it principled are themselves interpretative products: they are, therefore, always and already contaminated by the interested judgements they claim to transcend. (Fish, 1985:110)
6
Weintraub (1989:486) seems to take this position too. Relying on Fish’s conclusions once more, and again allowing us to insert ‘Methodology’ where Fish uses ‘Theory’, Weintraub reproduces the following passage with approval: Theory cannot guide practice because its rules and procedures are no more than generalizations from practice’s history (and from only a small piece of that history), and theory cannot reform practice because, rather than neutralizing interest, it begins and ends in interest and raises the imperatives of interest—of some local, particular, partisan project—to the status of universals. (Fish, 1985:111–12) For Weintraub, on accepting Fish’s conclusions that local interests are always involved, it seems to follow unproblematically that such interests are necessarily theoretical so that methodology must be inconsequential: The differences [between the various competing groups in economics—Post Keynesian, neoclassical, Marxian, Austrian] are theoretical, not epistemological, and certainly cannot be resolved through Methodological argumentation’ (Weintraub, 1989:487). He adds, just a few lines on: ‘And if the Methodological agenda, of providing a sure basis of judging whether one theory is better than another from
© 1997 Tony Lawson
considerations that stand apart from the two theories, cannot be begun, cannot succeed, then Methodology has no consequences for practice’ (ibid.). 7 That is, it may result from reflections upon such matters as natural scientific practices. Certainly it is erroneous to suppose that philosophy has access (and misleading to represent all philosophy as supposing itself to have access) to some autonomous realm of generalisations, or being, etc. 8 Of course, it should be clear that while a rejection of normative methodologism does not actually render the questioning of the possibility of naturalism illegitimate, necessarily inconsequential, or debilitating, a foreclosing of any questioning of the possibility of naturalism sits uneasily with a declared attitude of pluralism. 9 To understand Popper’s falsificationist criterion and how it emerged it is necessary to go back to Popper’s student days in Vienna of the 1920s. Here philosophy was dominated by the Vienna Circle, whose members included Carnap, Feigl, Frank, Gödel, Hahn, Neurath, Schlick and Waismann. Their system, known as logical positivism, was in essence a restatement of Machian empiricism (his theory of elements proposes that all phenomena can be reduced to complexes of sensations, and he dismisses the idea of a ‘thing in itself) in a form made possible by Russell’s developments in logic (especially as contained in Principia Mathematica, produced jointly with Whitehead), under the influence of Wittgenstein (especially Tractatus logico-philosophicus). According to logical positivism scientific propositions are about the world, known in sense experience. If a sentence does not refer directly or indirectly to experience, if no possible observation is relevant to the determination of its truth-value, then it was regarded by logical positivists as unscientific and thereby ‘meaningless’. In the attempt so to demarcate science from non-science, logical positivists formulated the criterion of verifiability, or susceptibility to positive test. Now while Popper certainly accepted the need to distinguish science from non-science, he proposed instead his own ‘better criterion of demarcation: testability or falsifiability’ Popper, 1992 [1974]:79). On this criterion any theory must be at least potentially falsifiable by empirical observation to qualify as ‘scientific’. Popper found empirical, logical and moral reasons for putting it forward. He was impressed that Einstein’s refutation of Newtonian mechanics empirically undermined the empiricist emphasis on certainty. At the same time he realised that universal statements of the form ‘all metal conducts electricity’, which he took to be amongst the greatest achievements of science, could never be verified by a finite number of observations. And he was also taken by a perception that while physics was vulnerable to refutation, the same appeared not to be true of Marxism, Freud’s psychoanalysis and Adler’s ‘individual psychology’ (all prominent topics in contemporary background Viennese culture). This perception convinced Popper that true scientists are distinct from non-scientists in accepting, in effect, a scientific morality—in being prepared to specify in advance the conditions under which they would be prepared to abandon their theories. It is precisely this, according to Popper, that Marxists, psychoanalysts, etc., unlike physicists, refuse to do. And all such lines of reasoning led Popper to prefer the falsifiability criterion for demarcating a science. 10 Of course, Popper soon realised that his approach to the demarcation problem facilitated a new general methodology of science. Science does not proceed inductively, according to Popper, through accumulating positive instances until a generalisation is confirmed. Instead, scientific generalisations are first proposed
© 1997 Tony Lawson
as bold conjectures and then subjected to rigorous tests by drawing out their implications; when a theory is refuted it is replaced by another bold conjecture, and so on. In short, Popper espoused his version of ‘deductivism—the view that theories are hypothetical-deductive systems, and that the method of science is not inductive’ (1992 [1974]:81; for good discussions of Popper’s method in the context of economics see especially Caldwell, 1991 and Hands, 1993b). 11 A major problem with this methodology is that if the deductivist criterion of predictive accuracy is employed almost every hypothesis is all too easily ‘falsified’. As Chalmers (1978) puts it: An embarrassing historical fact for falsificationists is that if their methodology had been strictly adhered to by scientists then those theories generally regarded as being among the best examples of scientific theories would never have been developed because they would have been rejected in their infancy. Given any example of a classic scientific theory, whether at the time of its first proposal or at a later date, it is possible to find observational claims that were generally accepted at the time and were considered to be inconsistent with the theory. Nevertheless, those theories were not rejected, and it is fortunate for science that they were not. (Chalmers, 1978:66)
12
13 14
15 16
Chalmers provides numerous examples, and details how other criteria developed in reaction to Popper also fail. In economics, the main consequence of the deductivist criterion is that, because all theories can be so easily ‘falsified’, the status quo prevails. Certainly the existence of science is not denied. Rather those who adopt positions similar to Caldwell’s ‘pluralism’ can be interpreted as wanting to steer a path between, on the one hand, the absurdity of either denying that science exists or of abandoning methodology altogether as irrelevant to science, and, on the other, the apparent foolhardiness of claiming some insight into what scientific practice and/or method entail. Caldwell, indeed, is explicit in identifying these ‘opponents’ of critical pluralism: ‘On the one hand there are those who claim to know what the scientific method is...’ while the ‘other group...either denies that science exists, or denies that methodology has any important connection to science’ (1989:46). And such references are found just as much in the writings of McCloskey, Mirowski and Weintraub along with most others who engage in metamethodology. The word ‘must’ is used here because it figures in the transcendental question as originally posed. But of course, and as emphasised below, the question being asked is really what is the most explanatorily adequate, albeit fallible, account that we can come up with which renders intelligible the phenomenon of interest— here experimental practices as conceptualised in experience. For discussions of Shackle’s views on these issues see Rotheim (1993) and Runde (1997); of Hayek’s, see Chapter 13 of this volume. Thus, on this issue, I am once more in sympathy with the Popperians, despite coming from a different perspective. As Blaug (1994) concludes, if methodologists refuse ‘to prescribe they end up with economics just as it is. “Economics is what economists do”, Jacob Viner once said. This ironic definition of the science of economics could well serve as the rallying cry of the antiPopperians. “Recovering practice” is what they call it but it is not much more than accepting economics as it is, for better or for worse’ (Blaug, 1994:129).
© 1997 Tony Lawson
17 Nadeau (1995) observes, for example, that ‘neoclassical economics is a theoretical framework thoroughly based on intentional concepts’ (1995:3); that ‘the phenomenon of intentionality...[is a] reality [which] is constitutive of the domain of economic theory...’ (ibid.: 7); ‘that all economic concepts are thoroughly intentional’ (ibid.: 9); that ‘intentionality should be acknowledged as the trademark of all important economic concepts since the marginalist revolution’ (ibid.: 13); suggesting that it was the ‘subjectivist approach to economic value that seems to have made it necessary to recognise the intentionality of economic phenomena’ (ibid.: 9); and so on. While our contributions are in many ways mutually supportive, Nadeau, however, does seem to suggest that, in assessing the possibilities for economics, any approach must be either external or internal: Now, there are at least two radically different ways of looking at this situation. One can start with some preconception concerning what a ‘legitimate empirical science’ must look like and try to see whether neoclassical theory passes the test, or one can choose to deal somewhat more directly with the epistemological specificity of economic reality as captured by neoclassical theory. (Nadeau, 1995:3) My own view is that these ‘two radically different ways’ of proceeding (appropriately interpreted) are actually but two mutually supportive aspects or strands of a single endeavour. 18 Of course, and as briefly alluded to above, subjectivists or hermeneuticists (such as the Austrians, including Hayek—see Chapter 13) have recognised this all along. However, their acceptance of the positivist ‘whenever event x then event y’ conception as nevertheless constitutive of, or necessary for, science has tended to encourage, in reaction, an essentially voluntarist conception of human agency and an almost non-existent role for social science. That is, by erroneously relinquishing science to positivism, and recognising the inapplicability of that conception to the social realm, the tendency has been to abandon the task of explaining social phenomena at all, or rather to abandon it, in effect, to some more basic science such as physics, biology or psychology. Hayek, (1942–4), for example, suggests that ‘What is relevant in the study of society is not whether...laws of nature are true in any objective sense, but solely whether they are believed and acted upon by the people’ (1942–4:51); and concludes ‘It is a mistake to which careless expressions by social scientists often give countenance, to believe that their aim is to explain conscious action. This, if it can be done at all, is a different task, the task of psychology’ (ibid.: 68). See also Hayek, 1943. (For a further elaboration of this reaction see Lawson, 1994a, 1996.) 19 In Locke’s formulation: The commonwealth of learning is not at this time without master-builders, whose mighty designs, in advancing the sciences, will leave lasting monuments to the admiration of posterity; but everyone must not hope to be a Boyle or a Sydenham; and in an age that produces such masters as the great Huygenius and the incomparable Mr Newton, with some others of that strain, it is ambition enough to be employed as an under labourer in clearing ground a little, and removing some of the rubbish that lies in the way to knowledge...
© 1997 Tony Lawson
(Locke, 1985 [1690]:xiii) 20 And, indeed, in recent years especially, a significant amount of extremely valuable servicing is being done. Thus, the project reported upon here must be seen as a contribution to a wider one, to a body of work which, I think, is best interpreted as under-labouring for an emergent science of economics. Although much of this wider work takes a very different tack and orientation on many issues compared to the position argued for here (for example, Blaug, 1980, 1994; Boland, 1982; Boylan and O’Gorman, 1995; Colander, 1994; Hands, 1993a; Hausman, 1992; Mäki, Gustafsson and Knudsen, 1993; Vanberg, 1994) there is also a good deal that is coherent with it. Indeed, I should emphasise that an increasing number of contributions are explicitly embracing a realist framework which is the same, or very similar, to that which I am defending here. This covers not only contributions in the field of economics as social theory, including economic methodology (Caldwell, 1990; Fleetwood, 1995, 1996a, 1996b; Lawson, C. 1994; Lawson, C., Peacock, and Pratten, 1996; Mäki, 1988, 1990a, 1990b, 1994; Peacock, 1993; Pratten, 1993, 1994, 1995; Runde, 1993, 1995c, 1995d, 1997; Sofianou, 1995), but also much that is interpreted (in addition) as contributions to specific programmes. Prominent amongst the latter are the Austrian (O’Driscoll and Rizzo, 1996), Institutionalist (Dopfer, 1989; Foss, 1994; Hodgson, 1985, 1986, 1987, 1989, 1992, 1993a, 1993b, 1993c; Hodgson and Screpanti, 1991; Jennings and Waller, 1994, 1995), post-Keynesian (Arestis, 1990, 1992; Arestis and Chick, 1992; Cottrell, 1994; Davis, 1987, 1989, 1992; Dow, 1990a, 1990b, 1990c, 1991, 1995, 1997; Hamouda and Harcourt, 1989; Harcourt, 1988, 1992; Rotheim, 1988, 1989, 1993; Runde, 1990, 1995a, 1997) and Marxian (Kanth, 1991, 1992; Meikle, 1985; Reuten and Williams, 1988) traditions. Furthermore, because the position defended here supports a conception of social science which basically encompasses the concerns of economists, sociologists, geographers and philosophers among others, it is not surprising to find people nominally located in other fields undertaking the same sort of ground work. Thus in addition to philosophers (e.g. Bhaskar, 1978, 1979, 1989; Collier, 1994) comparable work is to be found in human geography (Pratt, 1995; Sayer, 1984, 1985), sociology (Ekström, 1992) as well as history (Lloyd, 1993). Furthermore, there are numerous groups involved with projects in economics where the term realism itself is rarely if ever mentioned, and yet the goals and general perspective are very similar. An increasingly significant one is the developing programme of ‘feminist economics’ (see, for example, Folbre, 1993; Harding, 1995; Nelson, 1996; Strassman, 1994 and Strober, 1994).
6 TOWARDS A RICHER ONTOLOGY 1 2
More recently, with the ‘post-modernist’ or ‘deconstructionist’ turn, reality has often been identified with language, the text, or some such (for an excellent analysis, see Sofianou, 1995). Which Bhaskar (1986) collects under the head of synchronic emergent powers materialism. See also Bhaskar (1993, 1994).
© 1997 Tony Lawson
7 ECONOMETRICS 1
Rarely is the basic objective clearly stated or framework explicitly elaborated, however, and even less frequently is any connection with Hume’s conception of science and causality made. Yet there are noteworthy exceptions. Hendry, for example, has recently expressed his own thinking as follows: I think ‘causality’ is only definable within a theory. I am a Humean in that I believe we cannot perceive necessary connections in reality. All we can do is to set up a theoretical model in which we define the word ‘causality’ precisely, as economists do with y=f(x). What they mean by that in their theory is that if we change x (and it is possible to change x), y will change. And the way y will change is mapped by f, so we have a causal theory. They could give a precise or formal definition of the mapping f(.). Empirically, concepts such as causality are extraordinarily hard to pin down. In my methodology, at the empirical level, causality plays a small role. Nevertheless, one is looking for models which mimic causal properties so that we can implement in the empirical world what the theorist analyzes: namely, if you change the inputs, the outputs behave exactly as expected over a range of interesting interventions on the inputs. (Hendry et al., 1990:184)
2 3
4 5
It is though only a modification. Given the nature of the set-up, where parameters are specified to change or ‘drift’ this can be handled analytically only if constancy is attributed to other parameters elsewhere in the system. See, for example, Hutchison, 1994; Leamer, 1983; Rosenberg, 1992. Although econometric failure is manifest at many levels an outwardly familiar sign is the poor forecasting record of econometric ‘models’ designed to track developments in the economy. For a recent assessment of the (unsuccessful) performance of UK forecasting groups, see for example Kay (1995:19). (In fact, in examining the record of thirty-four UK groups, including the most quoted ones, Kay summarises the findings as follows: ‘Economic forecasters do not speak with discordant voices; [keeping an eye on each other] they all say more or less the same thing at the same time. And what they say is almost always wrong. The differences between forecasts are trivial relative to the differences between all forecasts and what happens.’) That is, observations that were not at hand when any given specification was ‘estimated’. Even prior to the obtaining of new observations, accepted results are usually be found to be rather ‘fragile’ in the specific sense of being extremely sensitive to any change in the specification. Consider Leamer’s assessment: I have a sense that most economists feel that conclusions from data sets are fragile. Somebody will add another variable, or they will control for some aspect of the time series phenomena in some other way, which will yield a substantially different conclusion. One of the reasons that we don’t treat empirical work seriously is that there have been so many cases of fragile conclusions. Somebody claims to have found something, and then six months later a new equation is estimated, and the same finding seems to be reversed. It creates the feeling among economists that conclusions from data are very fragile. (Hendry et al., 1990:187)
© 1997 Tony Lawson
6
7
Because Lucas is primarily concerned with what he terms the ‘theory of economic policy’ (1976:20) he mainly focuses upon the implications for policy simulations using already derived econometric models. However, it is clear enough that the identified problem is more generalised. Specifically, it extends to the endeavour of constructing and estimating such models in the first place if the data used in this relate to periods in which policy rules have been frequently changing. There is a point in his discussion where Lucas appears to hold out some possibility of uncovering stable econometric relationships which may be useful for policy analysis. But he quickly expresses a view that this is an unlikely eventuality. Specifically, Lucas writes: Perhaps the adaptive character of this early stage of macro-economic forecasting is merely the initial grouping for the true structure which, however ignored in statistical theory, all practitioners knew to be necessary. If so, the arguments of this paper are transitory debating points, obsolete soon after they are written down. Personally, I would not be sorry if this were the case, but I do not believe it is. (Lucas, 1976:24)
8
9
In consequence he recommends that policy changes be made in a pre-announced, well-understood fashion in the expectation that any move to a new structure will thereby be as stable and predictable as possible. Of course, Lucas himself focuses on the consequences for policy-makers. But what if the government (or policy-making body) is unwilling or unable to comply with his suggestions? Remember that Lucas does not offer anything along the lines of the transcendental realist perspective as an alternative approach to economics. In any case it seems most unlikely that econometricians will abandon easily the methods and orientations in which they have invested so much. Lucas presents his contribution in the fashion of someone outside the econometrics tradition. My interest, though, is in identifying the most compelling or attractive strategic options available to those professional econometricians who take seriously the Lucas, or any related, critique, but nevertheless intend to persevere with econometric modelling. Thus Leamer indicates that he distinguishes: three steps in a data analysis: planning, criticism, and revision. Planning refers to preparing responses to hypothetical data sets. Criticism occurs when you notice that your plans are inappropriate given the actual data you have in front of you. Revision occurs to the extent that the actual response to the data is different from the one you had originally planned. My sense is that most of econometric theory has been emphasising the process of planning responses to hypothetical data sets, and has not concerned itself very well with those issues that arise when plans cannot be carried out. (Hendry et al., 1990:200–1)
10 Thus Leamer acknowledges that the phenomena of criticism and revisions open up deep philosophical questions concerning the inferences you can legitimately make from a data set. Statistical theory, whether it’s Bayesian or classical, is confined
© 1997 Tony Lawson
completely to the planning phase. We know how to draw inferences only to the extent that there are no revisions. (Hendry et al., 1990:203; also see Leamer’s comments on pages 207, 211 and 228 as well as some responses from Hendry.) 11 There are many reasons to suppose that parsimonious specifications are likely ceteris paribus to be preferred or at least sought out. Maxims to this effect stylised as ‘Occam’s Razor’ or the ‘law of parsimony’, and others associated with such positivists as Mach, Poincaré and Pearson, are often quoted by economists as if authoritative (for an interesting discussion of this in the context of econometric modelling see Keuzenkamp and McAleer, 1995). My point here is simply that strategies to achieve simplicity in this sense are likely to be advocated primarily within Leamer’s planning stage. Such is clear for example in the ‘reductionist strategy’ associated most notably with the work of Hendry. Here the aim is to obtain a highly parameterised or ‘general model’ that captures the features of the data-set at hand in a desired manner, and then to simplify, to impose zero or other constraints on the model’s parameters. The tendencies which are likely to be in play outside of the planning stage are the focus of the discussion below. 12 The better econometricians do criticise this neglect, of course. Hendry, for example, suggests that one of the reasons that empirical work isn’t highly regarded...is that many researchers view a piece of empirical work as the final product. They set up their model, and analyze it, once and for all: Here is the theory, the data, and hence the result. Modelling is not seen in economics as an incremental progressive accumulation of knowledge in the way I see it in other subjects where you get credit for taking an experiment somebody else did and improving upon it in one direction, throwing some light back on what previously had been found, high-lighting some artifact, showing that if you change one aspect, another moves in a certain direction. (Hendry et al., 1990:180–1) Hendry’s own work is admirable in its concern to explain previous results and to embed or encompass previous apparent successes. 13 Of course, in certain modelling strategies it is suggested at the planning stage that all conceivable influences should be incorporated in the model at the outset. However, this is rarely feasible. First the data sets employed are rarely large enough to allow this. And no matter how many potential influences the researcher can think of, others will always be able to come up with additional ones, especially after the event. 14 This set of conditions is often ensured through a performance known as single exit modelling, or situational analysis. This usually involves contriving a set-up in which it is assumed not only that agents always act rationally (in some stronger or weaker but determinate sense) but also, and crucially, that the circumstances of action is such that, relative to the rationality conception in question, there is only one particular course of action that can ensue (see for example, Caldwell, 1994; Koertge, 1975; Langlois, 1986; Langlois and Csontos, 1993; Latsis, 1972, 1976a, 1976b; Popper, 1957, 1966). 15 Often, in fact, the identical relationship is merely assumed to hold at the aggregate level as is formulated at the level of the individual. Although this assumption seems prima facie implausible, there is nothing particularly novel in it. Even © 1997 Tony Lawson
before modern econometrics became established Hicks was suggesting that the move from the analysis of individuals or single firms to aggregate behaviour be made ‘by using the same principle...that the behaviour of a group of individuals, or a group of firms, obeys the same laws as the behaviour of a single unit’ (Hicks, 1939:245). And nearly seventy years earlier still Jevons can be found asserting that ‘the general forms of the laws of economics are the same in the case of individuals and the nations’ (Jevons, 1965 [1871]:16). See Klein (1946) for a criticism of this view following the emergence of econometrics more along the lines as we now know it. 16 Even if each individual is regarded as atomistic there is as yet no built-in assumption that all such crypto-atoms are identical. 17 For useful detailed discussions of the various strategies adopted see Stoker (1993) and Janssen (1993). 18 Such a conception of human behaviour is explicitly acknowledged to be ‘superficial’ by Lucas (in later years) but interpreted as a positive attribute of economic theorising in that it facilitates model predictions. He writes: To observe that economics is based on a superficial view of individual and social behaviour does not seem to me to be much of an insight. I think it is exactly this superficiality that gives economics much of the power that it has: its ability to predict human behaviour without knowing very much about the make up and lives of the people whose behaviour we are trying to understand. (Lucas, 1986:5425) 19 This fallacy of atomism clearly underpins the more familiar fallacy of composition. 20 At this point we can understand Keynes’ comments noted in the preface, namely: If we are dealing with the action of numerically measurable, independent forces, adequately analyzed so that we knew we were dealing with independent atomic factors and between them completely comprehensive, acting with fluctuating relative strength on material constant and homogeneous through time, we might be able to use the method of multiple correlation with some confidence for disentangling the laws of their action... (Keynes, 1973b:286) As is well known Keynes concludes that ‘every one of these conditions is far from being satisfied by the economic material under investigation...’.
8 ECONOMIC ‘THEORY’ 1 2
I enclose the term theory in scare quotes throughout this chapter in order repeatedly to remind the reader that there is actually much more to economics as theory than this particular project. As in the previous chapter I prefer, where possible, not to focus my criticisms on any specific pieces of substantive work—which would be invidious and too easily dismissed as biased selection. Rather I seek out relatively clear and explicit summary statements by reflective spokespeople of the project, statements which others in the project appear to treat as authoritative.
© 1997 Tony Lawson
3 4
5
6
7
8
9
Of these features Hahn (1984) notes that his commitment to (1) is ‘pretty strong’; that his commitment to (2) is less so but that he still accepts it for its ‘theoretical fruitfulness’ (1984:2); while his feelings for (3) are ‘rather secure’ (ibid.: 3). Some of the features which remain unelaborated or explained are, of course, rather fundamental. Why, for example, has the study of equilibrium states been consistently included amongst the project’s essential features? Note that the commitment to the study of equilibrium states is not of the form of an axiom. And Hahn (1985) explicitly adds ‘there is no axiom to the effect that an economy is in equilibrium at all times’ (1985:14). Why then the significant concern with it? Furthermore, why, on what justification, have the individualist perspective and the rationality axiom been adopted anyway? On what basis or what evidence can such stances or axioms be regarded as being beyond question? Questions of this sort appear never to be addressed seriously. Clearly, though, any explanation which I provide here of the nature and problems of orthodox theorising should deal with such questions. And this indeed is the case below. For the time being, however, the point to emphasise is that Hahn’s assessment is insightful and, amongst theorists, widely accepted. Indeed, for some time now, at least until quite recently, it appears to have remained largely unchallenged from ‘within’. In the general equilibrium context, for example, with many agents each endowed with given preferences, beliefs and assets, the question of interest is whether potential trades exist that would make each participant better off, or at least which would make none worse off while the situation of one or more improves. Of course, desired sales or acquisitions depend upon the prevailing structure of prices. The traditional notion of an equilibrium is a particular structure of prices in which everyone’s desired actions are mutually compatible. The specific ‘internal problems’ of the orthodox, general equilibrium theory, project then are: (1) if one such price vector is conceivable many others may also be; and (2) there is nothing in the typical set-up to suggest why or how any such ‘equilibrating’ price structure could ever come about. For a detailed and illuminating history of the reliance upon the equilibrium notion in economics see Ingrao and Israel (1990). It is possible to argue that the reason game theory took off in the 1980s (despite few recent results at that time to add to those established significantly earlier by von Neumann and Morgenstern (1944) and Nash (1950, 1951)) is that it was ready and waiting when the consensus grew that the arbitrariness of general equilibrium theory necessitated some alternative. It would be a distraction to explore the historical and technical details of this development here. Rizvi (1994a) provides an excellent discussion, detailing just how, in addition to various conceptual problems, game theory necessarily gives rise to numerous problems of arbitrariness of its own. Note that greater determinacy is to be expected only in those analyses which make definite specifications of expectations part of the system solution concept. If, in contrast, models of learning are introduced explicitly the system may easily become more indeterminate. I am not denying, of course, that deducibility and covering-law requirements can each be non-trivially satisfied without the other. Both, though, are part of what has become known as the Popper-Hempel theory of explanation, which I am here referring to as, or subsuming under, deductivism, the theory of explanation I am ascribing to orthodox economic ‘theorising’. Examples where this is not the case, i.e. where the law-like statements are not taken to be of necessity empirical, are examined in Chapter 9.
© 1997 Tony Lawson
10 And sometimes it is asserted. In a paper interpreting game theory, for example, Rubinstein concludes: ‘Overall, game theory accomplishes only two tasks: It builds models based on intuition and uses deductive arguments based on mathematical knowledge’ (1991:923). 11 Even Solow, who has argued explicitly for incorporating context-specific features of an agent’s situation into economic ‘theory’, remains seemingly unprepared to question the orthodox reliance upon the axiomatic deductivist form of reasoning: I am not prepared to abandon the exhaustive study of the implications of particular axiom systems, though I admit that I do not expect a lot from that sort of theory. What I am arguing against is the foolish belief that when it comes to studying the real world there is only one useful system of axioms and we already know what it is. (Solow, 1986:22) 12 Doubtless most economists who accept the universality of the basic deductivist mode of explanation suppose that event regularities of the noted sort are, if necessary, nevertheless insufficient for science and/or explanation. The point I have been making, however, is that such regularities are not only insufficient, but also unnecessary, for it. 13 Of course most theorists never construe indeterminacy in their models as a virtue. Consider the following assessment by Kehoe, Levine and Romer: Models with indeterminacy are undesirable from a scientific point of view. Starting from fundamentals of the economy, a theory based on this kind of model offers little guidance about what should be observed. Nor is it possible to condition on the equilibrium values that are observed and to perform comparative statics analysis because small changes in the underlying parameters can lead to large and discontinuous changes in the observed outcomes. Indeterminacy is especially troublesome in dynamic models because it undermines the interpretation of these models in terms of an equilibrium where agents trade in spot markets and form expectations about the future. If there is indeterminacy concerning the equilibrium that obtain[s] in the present, there is also indeterminacy in the equilibria starting from future dates, and this makes the formation of expectations problematic. (Kehoe, Levine and Romer, 1990:2) 14 Thus Robbins observes: At the same time it must be admitted that the propositions which have hitherto been established are very general in character. If a certain good is scarce, then we know that its disposal must conform to certain laws. If its demand schedule is of a certain order, then we know that with alterations of supply its price must move in a certain way. But, as we have discovered already, there is nothing in this connection of scarcity which warrants us in attaching it to any particular commodity. Our deductions do not provide any justification for saying that caviare is an economic good and carrion a disutility. Still less do they inform us concerning the intensity of the demand for caviare or the demand to be rid of carrion. (Robbins, 1932:106) © 1997 Tony Lawson
15 In the light of repeated failure, it is to be expected that some ‘theorists’ will deny any ambition or desire to illuminate the real world. Certain heroic attempts to rationalise this position are considered in Chapter 9 below. Mainly though any such denials amount to a ‘free-riding’ reliance upon others to ‘justify’ the existing significant share of available (teaching and research) resources being absorbed by this largely irrelevant project. It is equally clear however that many serious and reflective theorists do want the project explained. Not surprisingly they conclude, as does Rubinstein in his tribute to Nash’s contribution to game theory, expressing puzzlement as to the project’s ultimate purpose and direction: The issue of interpreting economic theory is, in my opinion, the most serious problem now facing economic theorists. The feeling among many of us can be summarized as follows. Economic theory should deal with the real world. It is not a branch of abstract mathematics even though it utilises mathematical tools. Since it is about the real world, people expect the theory to prove useful in achieving practical goals. But economic theory has not delivered the goods. Predictions from economic theory are not nearly as accurate as those by the natural sciences, and the link between economic theory and practical problems, such as how to bargain, is tenuous at best. Although I have never heard an economist seriously claim that the Nash bargaining solution is a good predictor of bargaining in real markets, this solution is a standard tool in modeling interactions among negotiators. Economic theory lacks a consensus as to its purpose and interpretation. Again and again, we find ourselves asking the question ‘where does it lead?’ (Rubinstein, 1995:12) 16 We can now observe how it is that neither new institutionalist economics or behaviouralism are alternatives to mainstream economics. Both rest upon regularity determinism. What ultimately distinguishes these two, moreover, is merely the way in which determinateness is achieved. Langlois and Csontos (1993) make the latter point when, in essentially embedding ‘new institutional analysis’ within situational analysis (or the ‘situational logic’ or SL approach), they write: If we compare the ‘programmes’ of situational logic and behaviouralism, we find close family resemblances as well as important differences. What unites the two programmes is a similar explanatory objective: to generate determinate conclusions from models that employ assumptions about human behaviour. The principal difference between the two approaches lies in the way in which they achieve determinateness. (Langlois and Csontos, 1993:118) Two pages further on they observe: There is nothing inherent in the SL approach that requires perfect information or a ‘thin’ conception of rationality. We can place the agent in any kind of situation we choose, with one proviso: the logic of the situation must be such that it leaves the agent with only a single exit. (ibid.: 120)
© 1997 Tony Lawson
They add: ‘It is true, however, that the SL approach focuses on the situation, whereas the behaviouralist approach focuses on the agent. And herein lies the real difference between the two...’ (ibid.: 121). 17 As Arrow emphasises in referencing competitive general equilibrium theory: ‘“atomistic” assumptions concerning individual households and firms are not sufficient to establish the existence of equilibrium: “global” assump-tions...are also needed’. Thus a limit is set to the tendency implicit in price theory, particularly in its mathematical versions, to deduce all properties of aggregate behaviour from assumptions about individual economic agents’ (Arrow, 1968:382). Of course, despite this recognition the desire amongst ‘theorists’ to derive macrolevel phenomena from ‘strict microfoundations’ has, with some recent exceptions (e.g. Dierker, 1974, 1986; Grandmont, 1987; Hildenbrand, 1983b; Kirman and Koch, 1986; Kirman, 1989), remained strong. Essentially the object has been to establish macro-level results on the basis of ‘microfoundations’ employing only a minimum of specifications regarding any macroeconomic structure, basically those which allow the ‘existence’ of an equilibrium to be demonstrated. As Drazen (1980) conceives the enterprise: ‘Explanations of macroeconomic phenomena will be complete only when such explanations are consistent with microeconomic choice theoretic behaviour and can be phrased in the language of general equilibrium theory’ (1980:293). The upshot, though, is that the project’s own goal of determining macro-level regularities from micro-specifications has failed, and the project’s own results indicate this failure was inevitable. This is hardly the place (it is not essential to my argument) to survey this episode, but, for a good account of this—drawing out the negative implications of results of Debreu (1974); Mantel (1974, 1976, 1977) and Sonnenschein (1972, 1973a, 1973b) in particular, and demonstrating that there are no strict microfoundational approaches to macroeconomics in general equilibrium theory, and that such existence proofs as there are depend upon unwarranted or (relative to the project) ad hoc assumptions—see Rizvi (1994b). For an equally insightful and extended study which reaches the same conclusions though exploring all aspects of the question of ‘microfoundations’, see Janssen (1993). 18 Hausman usually writes only of ‘economics’, but makes it clear that by this term he wishes to refer to something he understands as ‘neoclassical economics’. He writes: ‘Indeed, to avoid unnecessary repetition, I shall usually omit the adjective “neoclassical” and just speak of “economics” when I am discussing neoclassical economics. This is merely a convenience, not a covert attempt to denigrate other schools of economics or to define them out of existence’ (1992:3; emphasis in the original). 19 Thus in an article that summarises the main thesis of the book, Hausman (1994) writes: To simplify the discussion I shall focus exclusively on orthodox theoretical microeconomics and general equilibrium theory. In speaking of ‘economics’, I shall be speaking only of this part of the discipline. This is only a terminological simplification, not of course a denial that there are other important parts of economics. (1994:196) 20 Thus Hausman (1994) writes: ‘Economists (of course) recognize that causal factors left out of economic theory sometimes influence market phenomena. Economic theory is thus inexact’ (ibid.: 207–8). © 1997 Tony Lawson
21 There is also a problem for Hausman’s account in understanding why or how the attention to rational greed as a causal factor leads to an emphasis upon equilibrium notions anyway. Certainly, the existence of equilibrium is not treated in ‘economic theory’ as a given, as an empirical law, as ‘theorists’ such as Hahn (1984, 1985) often explicitly point out. It is noticeable, moreover, that while ‘abstract general equilibrium theorizing’ is hardly mentioned in Hausman’s (1994) summary article, in the (1992) book he has a difficult time trying to make it fit in (see especially his Chapter 6). Problems of reconciliation subside, however, once we recognise equilibrium just as a solution concept to be employed in determining the consistency of some set of deductivist formulations. 22 Although assumptions such as perfect foresight, rational expectations, two commodity worlds, closed economics—which have nothing to do with real capacities—are of course fundamental problems; indeed they are irrelevancies.
9 ALTERNATIVES AND/OR PRELIMINARIES TO SUCCESSFUL EXPLANATION 1
In fact, Hahn reserves his most severe criticisms not for those who call for greater realisticness in economics (whom he mostly dismisses without argument as not understanding his project) but for those who attempt to employ the models produced by ‘theorists’ to explain/predict actual social phenomena. That is, Hahn is particularly critical of attempts to estimate or test the formalistic models of economic ‘theory’ using measured data on actual phenomena. Consider, for example, the following not untypical assessment: The economists I have been discussing might be taken to be engaged in the following programme: to enquire how far observed events are consistent with an economy which is in continuous Walrasian equilibrium.... [Even if this programme was successful]...it would not be true that we understood the events. For we would not understand how continuous equilibrium is possible in a decentralised economy and we do not understand why a world with Trade Unions and monopolies behaves like a perfectly competitive one. Theorising in economics I have argued is an attempt at understanding and I now add that bad theorising is a premature claim to understand. What has happened in this instance is that ‘as if’ prediction methodology has taken over. Recall Friedman’s example: leaves on a plant orientate themselves as if the plant were maximising the surface area of leaves exposed to the sun. This may well predict the orientation of leaves but it is not an account which we understand precisely because we know that plants are not, as Friedman notes, capable of any calculations. So we do not leave matters there. We investigate chemical feedback mechanisms which account for the observation and which we understand. They fit to what we know generally about chemical substances in quite different contexts and into what we quite generally stipulate about causal processes. (Hahn, 1985:15; see also Hahn (1994:240)
2
Hahn takes the view that if real world phenomena are to be addressed we need to understand the mechanisms that actually produced them. Thus any approach
© 1997 Tony Lawson
3
along the lines of that advocated by Friedman (1953) is rejected because it disavows any need to account for the observed actual phenomena it addresses in terms of the processes which produced them. The instrumentalist claim, which Friedman’s 1953 methodological essay is usually taken to illustrate, is that in accommodating or predicting actual phenomena, mechanisms responsible for them either cannot be, or do not have to be (or perhaps do not exist to be), identified. Alternatively put, it is held that the criteria of theory adequacy can be satisfied independently of any consideration of (the reality of) any mechanism by which the phenomenon in question was produced. In Friedman’s case such criteria boil down essentially to one: predictability. Indeed, not only is predictive accuracy reckoned to be the only essential criterion of theory choice or selection, but the prediction of events is thought to be the sole, or anyway the primary, use to which any successful theory is eventually to be put. Notice, parenthetically, that Friedman’s instrumentalist posture does not, as many (including Hahn on occasion) suppose, turn on the ‘as if’ clause per se but rather on the more highly qualified ‘merely as if’ expression (see Lawson, 1992). Retroductive arguments, including transcendental ones (see Chapter 5), are forms of ‘as if’ reasoning. The realist aim in employing retroductive arguments is to formulate hypotheses of the sorts of conditions that could have given rise to some phenomenon of experience. In other words, the object is to infer conditions which, if they had been operative, could have given rise to the observed phenomena in question. Certainly, in such analyses the theories formulated are such that the noted phenomena are as if generated in accordance with them. But there is no ‘merely’ about such exercises. In other words, once various mechanisms are theorised their reality must be assessed. If competing theories are retroduced they must be selected from on the basis of their (relative) explanatory adequacies. In addition, the facilitating conditions of any mechanism must in turn be identified and empirically assessed—the mechanism thought to be producing some identified phenomenon must in its turn be explained; and so on. In short, there is nothing wrong with the ‘as if’ mode of reasoning per se, so long as it is utilised to initiate an investigation that attempts to identify the real mechanism(s) via which the (equally real) phenomena in question were actually brought about. Of course, the instrumentalist reliance upon prediction in the manner which Friedman advocates presupposes, for its validity, that constant event conjunctions are ‘out there’ waiting to be found. The necessary premise is that event regularities of the sort noted throughout this book are spontaneous and ubiquitous, and that the scientific task at hand is basically to identify them. It is the implicit belief that the world is so constituted that allows the conclusion that (excess) theoretical content can in practice be treated as an optional extra. But it is precisely the presupposition of a general availability of spontaneous event regularities, certainly in the social sphere, that is here being contested and, on empirical grounds amongst others, rejected. Friedman, in his 1953 essay at least, merely asserts that the conditions essential to predictive successes are available to economists. And it is because most of the contributions orientated as explicit responses to Friedman’s account have likewise taken the widespread availability of event regularities as given, that his conclusions have not been dislodged (see Lawson, 1992). It is not only mainstream economists who suppose that such constructions necessarily have use. Even some very competent methodologists appear to concur. Thus, Mäki (1992a) seems dismissive of
© 1997 Tony Lawson
those who are suspicious of any formulation of an economic theory or model employing assumptions which are not believed to be precisely true or which serve to exclude factors that might be thought of as relevant. In such cases the issue is to be specified as that of unqualified narrowness versus comprehensiveness or the unqualified truth versus the falsehood of the representations of a theory. Assumptions involving idealizations, such as the assumption that goods are perfectly divisible and homogeneous, or that the economy is closed and in full employment equilibrium, may be objected to simply because they are, in most cases, false, and contribute to eliminating from consideration factors that may appear to belong to the total situation. Paying little or no attention to the goals of such isolations and idealizations, the critics become annoyed simply because of the obvious perceived discrepancy between a theory and the way the world appears to them. Such critical doubts are sometimes held by practically minded people, beginning students of economics, and some scholars in other social sciences, or, more generally, by people who are more accustomed to making uncontrolled isolations by using omissions... (Mäki, 1992a:345, 346; see also Hausman, 1992:131, 132)
4
But what are the ‘goals of such isolations and idealisations’ that would render them worthwhile? It is this and related questions that are pursued in the current chapter. Perhaps this conclusion has been recognised most clearly by Rosenberg (1983, 1994). Apparently holding to the empirical law precondition for science; believing that the character of intentional variables—belief, desire, expectation—is such that ‘we cannot expect to improve our intentional explanations of action beyond their present levels of predictive power’ (1994:224); and (erroneously) concluding thereby that this prevents economics from becoming a successful science, Rosenberg correctly observes where such a line of reasoning logically leads: to a purely formal or logical apparatus, a branch of applied mathematics, with no obvious pretensions whatsoever to providing understanding, explanation, or any meaningful input into policy discussion. Echoing Hayek’s (1937) assessment that mainstream economics is ultimately little more than a branch of logic, Rosenberg concludes: Much of the mystery surrounding the actual development of economic theory—its shifts in formalism, its insulation from empirical assessment, its interest in proving purely formal, abstract possibilities, its unchanged character over a period of centuries, the controversies about its cognitive status—can be comprehended and properly appreciated if we give up the notion that economics any longer has the aims or makes the claims of an empirical science of human behaviour. Rather we should view it as a branch of mathematics, one devoted to examining the formal properties of a set of assumptions about the transitivity of abstract relations: axioms that implicitly define a technical notion of ‘rationality’, just as geometry examines the formal properties of abstract points and lines. This abstract term ‘rationality’ may have far more potential interpretations than economists themselves realize, but rather less bearing on human behaviour and its consequences than we have unreasonably demanded economists reveal. There are some important practical consequences of this answer to the question of what economics really is. If it is best viewed as more akin to a
© 1997 Tony Lawson
branch of mathematics on the intersection between pure axiomatization and applied geometry, then not only are several cognitive mysteries about economics solved, but more importantly our perspective on the bearing of economic theory must be fundamentally altered. For if this view is correct we cannot demand that it provide the reliable guide to the behaviour of economic agents and the performance of economics as a whole for which the formulation of public policy looks to economics. We should neither attach much confidence to predictions made on its basis nor condemn it severely when these predictions fail. For it can no more be relied on or faulted than Euclidean geometry should be in the context of astrophysics. Admittedly this attitude leaves a vacuum in the foundations of public policy. For without economics we lose even the illusion that we understand the probable, or potential, long term or merely possible consequences of choices that policy makers are forced to make. Of course, the caution that loss of illusions may foster is certain to be salubrious. On the other hand the vacuum may attract really useful foundations for decisions about the economy and its improvement. (Rosenberg, 1983:391–2) 5 6
The debate of methods with the German Historical School in which Menger’s ‘Investigations’ figured prominently in undermining the supposed antitheoretical methods and attitudes of that School. There is also, according to Menger, a third group of sciences which he refers to as the practical sciences or technologies. These, however, do not reveal aspects of reality but concentrate on how such knowledge as is available can be instrumentally applied: The sciences of this type do not make us aware of phenomena, either from the historical point of view or from the theoretical; they do not teach us at all what is. Their problem is rather to determine the basic principles by which, according to the diversity of conditions, efforts of a definite kind can be most suitably pursued. They teach us what the conditions are supposed to be for definite human aims to be achieved. Technologies of this kind in the field of economy are economic policy and the science of finance. (Menger, 1985:38)
7
8
The latter regularities, or empirical laws, connecting as they do observable phenomena never appear in a strict or universal form: ‘The conclusion that the phenomenon C follows the phenomena A and B in general (that is, in all cases, even those not observed!), or that the phenomena under discussion here are in general coexistent, transcends experience, the point of view of strict empiricism’ (ibid.: 57). Menger writes: All these factors are ultimately given by the particular situation, independent of human choice. The starting point and the goal of all economy (need and available quantity of goods on the one hand and the possible completeness of satisfaction of the material needs on the other) are ultimately given to the economic human, strictly determined in respect to their nature and their measure. (ibid.: 63)
© 1997 Tony Lawson
9
See Menger (1985): The purpose of the theoretical sciences is understanding of the real world, knowledge of it extending beyond immediate experience, and control of it. We understand phenomena by means of theories as we become aware of them in each concrete case merely as exemplifica-tions of a general regularity. We attain a knowledge of phenomena extending beyond immediate experience by drawing conclusions, in the concrete case, from certain observed facts about other facts not immediately perceived. We do this on the basis of the laws of coexistence and of the succession of phenomena. We control the real world in that, on the basis of our theoretical knowledge, we set the conditions of a phenomenon which are within our control, and are able in such a way to produce the phenomenon itself. (1985:55–6)
10 Menger writes: For the most important and most basic of the theoretical natural sciences also suffer from the same weakness for which our historical economists reproach previous theories of social science. Chemistry and physics, too— no less, however, a series of other exact sciences like mechanics, mathematics, etc.—appear, when measured by the critical standard of our historians, contrary to reality, unempirical, and therefore in need of the same reform as theoretical economics. Chemistry does not teach us ‘real concepts’ of definite groups of concrete phenomena. Its elements and compounds are, rather, unempirical in their complete purity, they are not to be observed in nature uninfluenced by human art; indeed, to some extent, they cannot be prepared synthetically. Pure gold, pure hydrogen and oxygen, and their pure compounds are not given empirically, either per se, or in that ideally strict measure which the laws of chemistry presuppose. Chemistry operates with factors which are unempirical qualitatively and in certain respects quantitatively also. It furthermore does not deal with substances in the totality of their phenomenal form. It apprises us of their nature and laws, not in respect to all sides of their being, but only to a specific one. Chemistry, to use the language of our historical economists, starts with the dogma that the basic chemical substances and their compounds are empirically present in their complete purity, that ideally they are exactly measurable, that gold and oxygen in their real phenomenal form are exactly identical in all places and at all times. Moreover, it is concerned with only one single side of the real world, and its laws are accordingly, in contrast to the totality of the world of phenomena, based upon arbitrary assumptions and unempirical. As we probably do not need to amplify any further, the same is true of physics, and in particular of mechanics and mathematics. Pure mechanics starts in the case of its most important laws with the arbitrary and nonempirical assumption that bodies move in a vacuum, that their weights and their paths are measured exactly, that their centres of gravity are determined exactly, that the forces by which bodies are moved are known exactly and are constant, that no disturbing factors develop their activity; and thus—to use the language of our historians—it © 1997 Tony Lawson
starts with a thousand other arbitrary, unempirical dogmas. And just like mathematics, whose unempirical presuppositions probably need not be pointed out particularly (think of the mathematical point, the mathematical line, the mathematical surface, etc.!), they, too, do not deal with the world of real phenomena in its totality, but only with a single side of it. Also in this respect they are therefore, in contrast to ‘full empirical reality,’ arbitrary and unempirical... (ibid.: 85–6) 11 Notice, parenthetically, that the ‘thing’ which possesses the tendency may not always be the same ‘thing’ recorded in any statement of its tendency. Science is replete with successful elaborations of novel kinds—electronic structures, genetic codes, magnetic fields—as the bearers of tendencies that are manifest in the behaviour of observed (other) ‘things’. The point here, though, is that, once triggered, any tendency set in train will go on acting no matter what the context or milieu, irrespective of whether countervailing mechanisms are active. If triggered, a tendency can be in play whether or not fulfilled, and whatever the actual result that eventually ensues. 12 It is not enough, in other words, to attempt to explain how more complex or ‘organic’ phenomena may have come about—see Menger’s Book III. 13 Certainly it is consistent with assessments of the following sort: Man’s powers are limited: almost everyone of nature’s riddles is complex. He breaks it up, studies one bit at a time, and at last combines his partial solutions with a supreme effort of his whole small strength into some sort of an attempt at a solution of the whole riddle. For breaking it up, he uses some adaptation of a primitive but effective prison, or pound, for segregating those disturbing causes whose wanderings happen to be inconvenient for the time: the pound is called Ceteris Paribus. The study of some group of tendencies is isolated by the assumption other things being equal: the existence of other tendencies is not denied, but their disturbing effect is neglected for a time. The more the issue is thus narrowed, the more exactly can it be handled; but the less closely does it correspond to real life. (Marshall, 1898:40) 14 He writes: The fact is, of course, that the assumption of perfect rationality in the sense of complete consistency is simply one of a number of assumptions of a psychological nature which are introduced into economic analysis at various stages of approximation to reality. The perfect foresight, which it is sometimes convenient to postulate, is an assumption of a similar nature. The purpose of these assumptions is not to foster the belief that the world of reality corresponds to the constructions in which they figure, but rather to enable us to study, in isolation, tendencies which, in the world of reality, operate only in conjunction with many others, and then, by contrast as much as by comparison, to turn back to apply the knowledge thus gained to the explanations of more complicated situations. (Robbins, 1932:93–4—see also 97) 15 Of course, even in cases where the object which is treated as acting in isolation is a real tendency, it will generally be misleading or distracting to focus upon an event regularity imputed to such a situation. This is so because the focus of © 1997 Tony Lawson
attention is then unduly restricted; it is concentrated upon a situation that may never come about or have any relevance. As I have already noted in the discussion of Menger’s position, a statement about a tendency which is transfactually in play may license a subjunctive conditional about what would have happened at the level of the course of events if the system had been insulated from the activities of other mechanisms. But a comprehensive understanding of the situation cannot be captured or conveyed in this manner. Rather, to obtain a full and indeed useful or practical understanding of the situation, tendency statements must be interpreted as categorical and indicative, to the effect that, if triggered, a mechanism is really in play whether or not its effects are fully manifest.
10 SUBJECTIVISM 1
Hayek writes: The confusion about the concept of a datum is at the bottom of so many of our difficulties in this field that it is necessary to consider it in somewhat more detail. Datum means of course something given, but the question which is left open, and which in the social sciences is capable of two different answers, is to whom the facts are supposed to be given. Economists appear subconsciously always to have been somewhat uneasy about this point, and to have reassured themselves against the feeling that they did not quite know to whom the facts were given by underlining the fact that they were given—even by using such pleonastic expressions as ‘given’ data. But this does not solve the question whether the facts referred to are supposed to be given to the observing economist, or to the person whose actions he wants to explain, and if to the latter, whether it is assumed that the same facts are known to all the different persons in the system or whether the ‘data’ for the different persons may be different. There seems to be no possible doubt that these two concepts of ‘data’, on the one hand in the sense of the objective real facts, as the observing economist is supposed to know them, and on the other, in the subjective sense, as things, known to the persons whose behaviour we try to explain, are really fundamentally different and ought to be kept carefully apart. And, as we shall see, the question why the data in the subjective sense of the term should ever come to correspond to the objective data is one of the main problems we have to answer. (Hayek, 1937:39)
2 3
Although some commentators appear to conclude otherwise—see e.g. Caldwell (1988b) or Hutchison (1981). In his later work, produced in the twenty years or so following the publication of his scientism essay, things do at least appear to change (on this see Böhm, 1989; Hodgson, 1993a; and especially Fleetwood 1995). Thus, by the 1960s, we find Hayek (1967) writing explicitly of ‘the knowledge that rules exist in the objective world and a disinclination [and so implicitly a real possibility] to deviate from the rules commonly followed’ (1967:79). It is also clear that action is no longer considered always to be discursively conscious but conditioned by tacit knowledge (amongst other things). Thus we can find Hayek concluding that we learn as children to use language ‘according to rules that we do not explicitly
© 1997 Tony Lawson
4 5 6 7
know’ (ibid.: 87); and that in interpreting the world and acting appropriately we draw upon ‘rules that will guide us though we have never explicitly formulated them’ (ibid.). It follows, of course, ‘that our conscious reason can therefore always take account only of some of the circumstances that determine our actions’ (ibid.); so that the various conditions of action are matters that need to be identified, that action is something that we can and should attempt to explain. (See also Lawson, 1994a, and, for a related discussion, Fleetwood, 1995, 1996a, 1996b). The point though is that if transcendental realism or something approximating it is eventually embraced by Hayek it does not underpin his influential scientism essay. And as we shall see it is from the latter essay that his subjectivism and various influential methodological injunctions, stem. For example Peacock (1993). And see also 1942–4:97 where Hayek writes of social phenomena constituting ‘an order which is not a spatial or temporal order but can be defined only in terms of relations which are intelligible human attitudes’ (emphasis added). Note that psychology is not included here within social science. It seems reasonable to suppose that at some point Hayek worked out these views partly in (over) reaction to Neurath’s (erroneous) eliminativist physicalism. According to the latter all terms that are not reducible to physical descriptions must be eliminated from science. Hayek is correct to reject this form of reductionism and to defend a hermeneutic component to social science. However, this in no way justifies the radical subjectivism of the scientism essay. On all this I am quite in agreement with O’Neill (1996). However, I think that O’Neill is wrong to suppose that Neurath is ‘Hayek’s main positivist opponent in the scientism essay’ (1996:2), and that eliminative physicalism forms the main positivist target (ibid.: 2, 3). While in any case the influences upon Hayek, as well as his targets of criticism within the group of Viennese positivists, included others in addition to Neurath, it seems clear from the text and context of the scientism essay that at the time of writing it, Hayek is engrossed with what he perceives economists actually to be doing. It does, however, seem likely that the extreme form of physicalism of Neurath originally alerted Hayek to the limitations of the whole positivist programme. Indeed Hayek appears to acknowledge as much in his posthumously published ‘autobiographical dialogue’ (Hayek, 1994). Thus when asked his interpretation of the meaning of (Viennese) ‘positivism’ in the first part of the twentieth century, Hayek replies: It was almost entirely the influence of Ernst Mach, the physicist, and his disciples. He was the most influential figure philosophically. At the time, apart from what I had been reading before I joined the army, I think my introduction to what I now almost hesitate to call philosophy—scientific method, I think, is a better description—was through Machian philosophy. It was very good on the history of science generally, and it dominated discussion in Vienna. Joseph Schumpeter had fully fallen for Mach, and while I was still at the university, this very interesting figure, Moritz Schlick, became one of the professors of philosophy. It was the beginning of the Vienna circle, of which I was, of course, never a member, but whose members were in close contact with us. There was one man [Felix Kaufmann] who was supposedly a member of our particular circle, the Geistkreis, and also the Schlick circle, the Vienna circle proper, and so we were currently informed of what was happening there.
© 1997 Tony Lawson
What dissuaded me is that the social scientists, the science specialists in the tradition of Otto Neurath, just were so extreme and so naive on economics; it was actually through them that I became aware that positivism was just misleading in the social sciences. I owe it to Neurath’s extreme position that I recognised it wouldn’t do. And it took me a long time, really, to emancipate myself from it. It was only after I had left Vienna, in London, that I began to think systematically on problems of methodology in the social sciences, and I began to recognise that positivism in that field was definitely misleading. (1994:49, 50)
12 SOCIETY AND ECONOMY AS REPRODUCED INTERDEPENDENCIES 1
2
Even observable driving acts, of course, may still be conditioned by the rule in question. It is conceivable, and likely, that motorists neither wish to exceed notably the average speed of others, so risking being singled out for a penalty if a police car appears, nor wish to exceed the legal limit to such an extent that if a police car appears on the horizon, it is impossible to slow down to the legal maximum within a given time or space span. Of course the category of institution is central to the school of institutionalist thought in economics. I thus hesitate to insist on the above interpretation in the face of historical tradition or precedent. My perception, however, is that the institutionalist tradition is surprisingly undecided on the meaning of this term. Veblen, for example, views institutions as ‘settled habits of thought common to the generality of men’ (1909 [1919]:239); Hamilton defines an institution as ‘a way of thought or action of some prevalence and permanence, which is embedded in the habits of a group or the customs of a people’ (1932:84); and Neale’s entry in a recent Companion to Institutional and Evolutionary Economics reads as follows: ‘Institutions’ is the word that evolutionary (institutional) economists use for the regular, patterned behaviour of people in society and for the ideas and values associated with these regularities. Various phrasings have been used to define institutions or an institution: a usage that has become axiomatic by habitation; a collective action in control of individual action; widely prevalent, highly social habits; a way of thought or action embedded in the habits of a group or the customs of a people; prescribed patterns of correlated behaviour. (1994:402) While most of these interpretations focus on recurrent activities or habits, Neale himself links them to social rules, albeit with the latter understood as ‘regularities or patterns’ (ibid: 403). It seems to me that a basic problem with this literature is a failure to elaborate a notion of social structure, i.e., of a level of social reality which, though dependent upon, is irreducible to human thought and practice. Of course, all categories are open to continuous revision in line with transformed understandings. And it seems that the category of institution is ripe for reconceptualisation in the light of the set of understandings systematised as critical
© 1997 Tony Lawson
3 4
5 6
realism. In consequence, if not insisting on it, I nevertheless propose the above interpretation as one sustainable from the perspective now set out. Specifically, I suggest that the term ‘institution’ be used to designate those systems, or structured processes of interaction (collecting together rules, relations and positions as well as habits and other practices), that are relatively enduring and identified as such. For a discussion from the ‘new institutionalist’ perspective that is alive to the inherent danger of the latter strand of this possible two-way reduction, albeit whilst erroneously advocating the first, see Langlois (1986, 1989). In recent years the question of social continuity has received some over-due attention in economics under the heading of path-dependence. This literature expresses the explicit aim of bringing more history into economics. See especially David (1985, 1994). However, despite its numerous insights this literature is often formulated in such a manner that it is easily misunderstood and even used to justify essentially non-historical approaches—see the discussion in Chapter 18. It will be clear that this discussion ties in with some of that recently covered under the heading of evolutionary theory in economics. For discussions of this see Foss (1994), Hodgson (1993a), Jennings and Waller (1994) and Vromen (1995). This discussion has necessarily been pitched at a rather general level. In addition, the emphasis, albeit implicitly, has been primarily on reproductive or metamorphic change whereby a structure, organisation, system or thing does, or at least can, remain essentially the same sort of thing despite continuously undergoing some alteration or transformation in its form. The intention has been to emphasise the pervasiveness of (endogenous) change in social life (as everywhere else). In a more sustained analysis, of course, specific types of change warrant elaboration, including, explicitly, changes in the nature or types of things, and their causes and conditions. An obvious example is the quantity/ quality type of transformation. This is the process whereby, as some aspect of something gets bigger (or smaller), more numerous (or less numerous), hotter (or colder), a point is reached at which a qualitative transformation takes place (e.g. money, through accumulation, into capital; a pressure group, through increased membership, into a movement). Similarly, the manner of emergence of new, or different forms of, structure, due to crises resolution of contradictory (i.e. internally related and simultaneously supportive and antagonistic) processes, also warrants attention. Although forms of change such as these cannot be further elaborated here they are clearly consistent with the general transformational conception outlined in the text.
13 A SKETCH OF THE ACTING SUBJECT 1
Feminist economists seem to be one group who recognise the need for social scientists to seek a fuller understanding of the nature of the connection of the social to material reality. See for example Nelson (1996). As Nelson formulates matters: Without such an understanding of material connection, we have the scandal of professional economists working out endless theoretical yarns about preferences while a majority of people in the world live in a state of neediness apparent to any observer who has not lost his or her humanity.
© 1997 Tony Lawson
With an understanding that incorporates both choice and material connection, comes the possibility of abundance and a hospitable nature, if we choose wisely. (Nelson, 1996:35) 2
3
Or as Giddens puts it: ‘since the earliest experiences of the infant, shaping the basic security system whereby anxiety is canalized or controlled, predate differentiated linguistic competence, they are likely to remain thereafter “outside the bounds” of discursive consciousness’ (Giddens, 1984:49). As Keynes formulates it: The outstanding fact is the extreme precariousness of the basis of knowledge on which our estimates of prospective yield have to be made. Our knowledge of the factors which will govern the yield of an investment some years hence is usually very slight and often negligible. If we speak frankly, we have to admit that our basis of knowledge for estimating the yield ten years hence of a railway, a copper mine, a textile factory, the goodwill of a patent medicine, an Atlantic liner, a building in the City of London amounts to little and sometimes to nothing; or even five years hence. In fact, those who seriously attempt to make any such estimate are often so much in the minority that their behaviour does not govern the market. (Keynes, 1973a:149–50)
4
Keynes writes: Nevertheless the above conventional method of calculation will be compatible with a considerable measure of continuity and stability in our affairs, so long as we can rely on the maintenance of the convention. For if there exist organised investment markets and if we can rely on the maintenance of the convention, an investor can legitimately encourage himself with the idea that the only risk he runs is that of a genuine change in the news over the near future, as to the likelihood of which he can attempt to form his own judgment, and which is unlikely to be very large. For, assuming that the convention holds good, it is only these changes which can affect the value of his investment, and he need not lose his sleep merely because he has not any notion what his investment will be worth ten years hence. Thus investment becomes reasonably ‘safe’ for the individual investor over short periods, and hence over a succession of short periods however many, if he can fairly rely on there being no breakdown in the convention and on his therefore having an opportunity to revise his judgment and change his investment, before there has been time for much to happen. Investments which are ‘fixed’ for the community are thus made ‘liquid’ for the individual. (ibid.: 152, 153)
5 6
See Winslow (1986) for a wider analysis of Keynes’ thinking on psychological motives. This is a study which also explores Keynes’ indebtedness to Freud. Thus, for example, (good) ‘feminist research does not introduce political assumptions, values and interests into research fields that are otherwise valueneutral; it identifies the ones that are already there’ (Harding, 1995:7).
© 1997 Tony Lawson
15 ECONOMIC SCIENCE WITHOUT EXPERIMENTATION 1
I think this characterisation is appropriate given both of the common interpretations of the term ‘demi’ (as either half-way or as false). First any regularity observed is partial or incomplete. Second although any such partial regularities may be about real phenomena and capturing associations, they are not real laws at all. As Cartwright (1989) expresses the matter: Nature selects the capacities that different factors shall have and sets bounds on how they can interplay. Whatever associations occur in nature arise as a consequence of the actions of these more fundamental capacities. In a sense, there are no laws of association at all. They are epiphenomena. (Cartwright, 1989:181)
2 3 4
5
Though some, such as Lipton (1991) express reservations about whether this conclusion can be drawn. As is its meaning in other contexts such as sculpture or pictorial art. Notice, however, that there may be an inconsistency between the phenomenon reported and certain normative views of how things ought to be. Under such conditions we might well see a further enquiry initiated, either into ways of ameliorating the situation, or perhaps into a fuller understanding of the conditions sustaining it with a view to encouraging their transformation (see Chapter 19). I take it that this orientation is essentially Peircean. According to Peirce the point of enquiry is to assuage doubt and attain a state of (‘fixed’) belief. For Peirce, you cannot criticise what you do not doubt. (Such doubt, though, is not that of Cartesian rationalism—indeed, just as you cannot choose whether or not to doubt something so doubt presupposes previous belief.) Such doubt then is essential to enquiry; without it, and the contradictions that give rise to it, enquiry does not ensue. He writes for example: The irritation of doubt causes a struggle to attain a state of belief. I shall term this struggle enquiry, though it must be admitted that it is sometimes not a very apt designation. The irritation of doubt is the only immediate motive for the struggle to attain belief. It is certainly best for us that our beliefs should be such as may truly guide our actions so as to satisfy our desires; and this reflection will make us reject any belief which does not seem to have been so formed as to insure this result. But it will only do so by creating a doubt in the place of that belief. With the doubt, therefore, the struggle begins, and with the cessation of doubt it ends. (Peirce, 1966:99, 100)
Elsewhere Peirce adds: Every enquiry whatsoever takes its rise in the observation, in one or another of the three Universes, of some surprising phenomenon, some experience which either disappoints an expectation, or breaks in upon some habit of expectation of the inquisiturus; and each apparent exception to this rule only confirms it. There are obvious distinctions between the objects of surprise in different cases.... The enquiry begins with pondering these phenomena in all their aspects, in search of some point of view whence the wonder shall be resolved. (ibid.: 367) © 1997 Tony Lawson
6
7
8 9
There can be no context-independent account of what is meant by adequacy. The specificity of any criteria by which adequacy is determined is characteristic of all scientific explanation. It is once more the lingering influence of positivism, with its insistence on the generality of rules across disciplines, that misleads economists from recognising the necessarily contingent, pragmatic and field-specific nature of this aspect of explanation (see Miller, 1987:6). A recent consequence of this mistake in the UK at least, is the tendency to direct economic graduates away from fieldwork, and case study, where skills and methods can be determined on the job, to longer ‘taught masters course’, where largely irrelevant a priori techniques provide but a superficial semblance of an education in economics. Leamer’s own (largely methodological) contribution is admirable in such respects. But it aims to make life too straightforward for others. Indeed, a major problem with contemporary economics generally is that it attempts to render science far too easy. As it currently stands, with its almost exclusive reliance on computer packages, official data, and sanctioned statistical, or other mathematical, procedures, economics can be pursued, or so it is clearly supposed, without any knowledge of, or thought to, the economy at all. No wonder, then, that so many senior posts in economics faculties are being taken by engineers and applied mathematicians without previous experience of studying features of actual economies. No wonder, too, the current disarray within contemporary economics discipline and the fast growing suspicion of its affairs by those who remain on the outside. Of course there may be many pragmatic reasons for preferring one theory. In a contribution which I cannot fully do justice to here (it appeared as the present one was being finished) Boylan and O’Gorman, in putting forward their ‘causal holism’, write: Our objection is to the Lawsonian realist thesis that the structured entities which possess these powers are claimed to be non-empirical. If one postulates non-empirical entities, how can economists go on to furnish theoretical descriptions of these and know that their descriptions are correct or approximately true? Quine has taught us that it is not beyond the ingenuity of economists to construct a range of incompatible economic theories such that each one is compatible with the empirical evidence. Given this pluralism, causal holists do not see any rational way of deciding which non-empirical referents they should choose nor which theoretical descriptions are true of these non-empirical entities. (1995:212)
On the following page they add: Furthermore, causal holists acknowledge that the process of inference to the best explanation (Lipton, 1991) is a valuable inductive process. For instance, if we see shoe prints in the snow the best explanation is that a human being, rather than a monkey wearing shoes, passed by. However, if the scientific realist wishes to explicate the notion of abductive inference in terms of an inference to the best explanation and thereby argue that the non-empirical entities postulated by our mature, best confirmed economic theory is the best explanation of the observable economic facts and observable tendencies, the causal holist refuses to take this final step. According to causal holism, abductive inferences understood as inferences to the best explanation are limited to the domain of the observable. The © 1997 Tony Lawson
principal reason for this limitation is the same as that used against the transcendental realist’s use of non-empirical mechanisms. Quine teaches us that there is a multiplicity of such explanations, each compatible with the empirical evidence and hence we have no rational way of choosing between these. In other words, when the process of inference to the best explanation is extended to postulated non-empirical entities, there is no best explanation. There are many best explanations with no rational way of deciding between them. In causal holism economists can discover the hidden causal webs operating through economic systems and their transformations without recourse to the additional realist strategy of non-empirical mechanism and specific abductive inferences to these mechanisms. (ibid.: 213) 10 Under the heading of strict regularities I include ‘well-behaved’ i.e. tractable, probability ‘laws’. 11 As Haavelmo (1944) is forced to admit in setting out the ‘probability approach to econometrics’ ‘there is, in general, no limit to the number of...factors that might have a potential influence’ on any variable of interest (1944:24). 12 This, to repeat, is no less true of natural phenomena. As Cartwright (1989) argues: Nature, as it usually occurs, is a changing mix of different causes, coming and going; a stable pattern of association can emerge only when the mix is pinned down over some period or in some place. Indeed, where is it that we really do see associations that have the kind of permanence that could entitle them to be called law-like? The ancient examples are in the heavens, where the perturbing causes are rare or so small in their influence; and the modern examples are in the physics laboratory, where...our control is so precise that we ourselves can regulate the mix of causes at work. Otherwise, it seems to me, these vaunted laws of association are still very longoutstanding promissory notes: laws of association are in fact quite uncommon in nature, and should not be seen as fundamental to how it operates. They are only fundamental to us, for they are one of the principal tools that we can use to learn about nature’s capacities; and, in fact, most of the regularities that do obtain are ones constructed by us for just that purpose. (1989:182) 13 If I understand Boylan and O’Gorman (1995) correctly this distinction mirrors their contrast between pure and applied economics. The difference is that they construe only the latter to be an explanatory exercise (see e.g. pp. 135–41). However I not convinced that anything very significant hangs on this contrasting use of terminology. 14 For an example of this sort see Lawson (1981). In this enquiry the phenomenon singled out as warranting an explanation is that a particular Cambridge firm over a longish period—throughout the twentieth century up until the 1960s at least (when it was taken over by a large overseas-based multi-national)—had successfully combined a primary product market structure (high and stable demand for [advance technology] products) with secondary employment conditions (low wages, limited possibilities for personal advancement), while the
© 1997 Tony Lawson
general pattern in the economy at large had been for product market and employment conditions to positively correspond (see e.g. Wilkinson, 1981).
16 ABSTRACTION 1
As Marx (1973) formulates it, the concrete is concrete because it is the concentration of many determinations, hence unity of the diverse. It appears in the process of thinking, therefore, as a process of concentration, as a result, not a point of departure, even though it is a point of departure in reality and hence also the point of departure for observation [Anschauung] and conception. (Marx, 1973:101)
2 3 4
For a useful recent discussion of the process of abstraction in Marx’s method, a contribution that also has numerous parallels with the discussion set out below, see Ollman, 1993. And whatever the level, to the extent that there is an inconsistency between theory and evidence then prima facie some kind of explanatory investigation is required. See Ollman (1993) for an extremely good and extensive recent discussion of these and related issues. Thus Dobb (1972 [1937]) notes that if all that is postulated is simply that men ‘choose’, without anything being stated even as to how they choose or what governs their choice, it would seem impossible for economics to provide us with any more than a sort of algebra of human choice, indicating certain rather obvious forms of interrelationship between choices, but telling us little as to the way in which any actual situation will behave. (Dobb, 1972 [1937]):71) Of course, broad generalisations as descriptions of some manifest economic phenomenon can be useful as marking sites for analysis to begin—indeed this is the rationale for the notion of stylised facts noted above. But such generalisations will typically not be useful as explanatory devices, as statements of economic causes. As Dobb again observes, if an economic law is a statement of what actually tends to happen and not a mere statement of a relation between certain implicitly defined variables, then such propositions [as result from such high level abstractions] can surely be precious little guide to the ‘laws of motion of capitalist society’— or, indeed, to any of the other matters on which they are intended to pass an economic judgement. (ibid.: 42)
5
Specifically, this appears to be Cartwright’s recent (1989) position. Although Cartwright argues for an ontology of capacities, which at one point are likened to propensities or powers (1989:9), these capacities, or their associated causes, are nevertheless interpreted as ‘atomic’ (ibid.: 170, 174), and even as measurable, at least in principle, by probabilities (ibid.: 13). This metaphysical perspective,
© 1997 Tony Lawson
which she identifies with Mill, but recognises is explicitly rejected by Keynes (ibid.), does not, however, lead Cartwright to expect event regularities to be pervasive. Indeed, she repeatedly rejects such a scenario. For while the atomistic assumption guarantees that the intrinsic closure condition is automatically satisfied, Cartwright recognises that actual outcomes are generally produced by a shifting mixture of causes (ibid.: 175, 176). In other words, it is recognised that (what I am referring to as) the extrinsic closure condition cannot be guaranteed. Now, it is ultimately the accepted metaphysics, her belief that economic phenomena are atomic, that allows Cartwright to suppose that to conceptualise any cause quite independently of all context is a form of abstraction rather than a form of idealisation. She writes: Here is how I want to distinguish idealization and abstraction for the purposes of this book: in idealization we start with a concrete object and we mentally rearrange some of its inconvenient features—some of its specific properties—before we try to write down a law for it... By contrast, when we try to formulate Mill’s laws of tendencies, we consider the causal factors out of context altogether. It is not a matter of changing any particular features or properties, but rather of subtracting, not only the concrete circumstances but even the material in which the cause is embedded and all that follows from that. This means that the law we get by abstracting functions very differently from idealized laws. For example, it is typical in talking about idealizations to say...the ‘departure from truth’ is often ‘imperceptibly small’, or ‘if appreciably large’ then often ‘its effect on the associated model can be estimated and allowed for’. But where relevant features have been genuinely subtracted, it makes no sense to talk about the departure of the remaining law from truth, about whether this departure is small or not, or about how to calculate it. These questions, which are so important when treating of idealizations, are nonsense when it comes to abstractions. ... When my problem of abstraction is assimilated to the problem of idealization, it is easy to think, erroneously, that one can solve the combined problem by developing some notion of approximating truth. But that does not work. (Cartwright, 1989:187, 188)
6
It is also Cartwright’s metaphysics that conditions her support for the method of successive approximation (at least as formulated by Nowak, 1980), and encourages her to view it, with Nowak, as probably the principal method used throughout the sciences (ibid.: 204). Of course, this is equally true of natural science. As Hacking (1983) observes: We have the idea of numerous laws of nature adding up to a ‘resultant’. That metaphor comes from mechanics. You have this force and that force, this vector and that vector, and you can draw a pretty diagram with ruler and compass to see what results. John Stuart Mill remarked long ago that this fact about mechanics does not generalise. Most science is not mechanics. (Hacking, 1983:226)
© 1997 Tony Lawson
17 ON TRUTH IN ECONOMICS 1
2 3
4 5
Hendry (1983:70), for example, suggests that ‘A “true” model is virtually a selfcontradiction as it would have to be as complicated as whatever it is supposed to represent’. Elsewhere he elaborates: ‘I take it as self-evident that economic behaviour is sufficiently complex and evolutionary that it is not helpful to talk about economic theories or empirical models being “true” or of inferences yielding the “correct” results.’ And in the following paragraph he adds: ‘By their very nature, models are inherently simplifications and inevitably false...’ (Hendry, 1987:30, 31). More recently McAleer (1994) ties the term ‘abstraction’ to falsity more or less explicitly: ‘Models are abstractions from reality and, as such, there are no true models’ (1994:324). Further on he adds: ‘Econometric modelling is not intended to arrive at the truth, but at an adequate representation of the data’ (1994:328). See also Langlois and Csontos (1993:114–16). This conception is presumably influenced by the positivistic conflation of knowledge with direct experience and reduction of reality to events and states of affairs given in experience. Truth, of course, is a multifaceted notion and I cannot discuss all the possible aspects of it here. But see especially Bhaskar in Outhwaite and Bottomore (eds) (1993:680–1), and the numerous contributions on the topic by Mäki (for example, 1988, 1992a, 1992b, and 1994). See e.g. Hendry in Hendry et al. (1990:189). Strassman continues: The degree to which particular arguments ‘work’ in an intellectual community depends on who gets to evaluate them. In economics, a fairly narrowly defined demographic group in fact overwhelmingly predominates as the ‘evaluators’ of arguments. Situated knowledge predicts that if an intellectual community changes, such as its gender or racial composition, new accounts found more compelling by the restructured community will be produced, although such changes could be modulated by gatekeeping to ensure that those with different backgrounds or bodies conform to prevailing views. This is precisely what has happened in economics as more women have entered the profession. (Strassman, 1994:156)
6
Thus, for example, feminist economists have located numerous androcentric interests and values which appear to have shaped significantly the central concepts and methods of academic economics (see especially England, 1993; Folbre, 1982, 1991, 1993; Harding, 1995; Hartmann, 1981; Humphries, 1994; Nelson, 1993a, 1993b, 1996; Seiz, 1993 and Strassman, 1993a, 1993b). Humphries is especially interesting in locating the mainstream resistance to progressive developments in the methodologies of economists, along with their aversion to methodological critique. She writes: In other disciplines gender biases have long been acknowledged and solutions sought. In economics feminist revisionism has been laggard. Why? The answer lies in the uncompromising attachment of economists to their methodology. Economists can live with affirmative action where gender bias is identified with the under-representation of women. They might even tolerate a feminist empiricism according to which the biases of
© 1997 Tony Lawson
the discipline can be corrected by even stricter adherence to the alleged methodological norms of scientific inquiry. More dispas-sion, less involvement, a more scientific approach will purge androcentrism. What economists cannot permit is a deeper methodological critique. (Humphries, 1994:483)
18 ILLUSTRATION 1 I emphasise, however, that there can be no question of proposing the critical realist substantive account of anything. The most that may legitimately be said of any specific project, theory or illustration in this regard is that it is (or is not) consistent with the perspective here defended. There can be no expectation that researchers acting on the basis of critical realism will always reach the same conclusions. Given the inescapable openness of social reality in particular, researchers may even entertain a range of explanatory theories. And it is likely that any specific explanatory conception will undergo significant development over time. Of course, for those acting upon critical realism, the aim at each stage will be to bring empirical and other evidence to bear with the intention of identifying which of competing accounts is the most adequate. But there can be no guarantee that agreement will always be reached, or that, if and when agreement is reached, new evidence and thinking will not lead to revised accounts sooner or later. Substantive explanations, then, even when serving illustrative purposes, ought not to be tagged ‘critical realist’. Nor, incidently, should they be interpreted as constituting evidence by which the critical realist explanatory framework is itself to be assessed. Rather, the arguments for the latter have been set out in earlier chapters, taking the form of transcendental inferences from premises accepted as relatively uncontentious and sponsored by supporters of the more dominant alternative philosophy-of-science frameworks. In short, the examples and discussion which follow merely provide (1) an indication that explanatory endeavours consistent with critical realism are feasible in the social realm; and (2) illustrations in a more concrete form than hitherto of certain aspects of the sorts of explanatory goals and modes of reasoning that have been supported throughout. 2 Some connections between examples such as these and critical realism have already been made explicitly. See, for example, Fleetwood, 1995, 1996a, 1996b; C.Lawson, 1994; Peacock, 1993; Pratten, 1993, 1994, 1995; Rotheim, 1993 and Runde, 1993, 1997. 3 I do not, however, include in this category most of that diverse collection sometimes referred to as ‘new institutionalist economics’. Here I essentially agree with those (e.g. David, 1994; Hodgson, 1994; Posner, 1992; Pratten, 1995; Vanberg, 1994) who have identified it as a sub-set of the contemporary mainstream. Where the emphasis is on game theory and/or is highly formalistic this is clear (e.g. Schotter, 1981, 1983; Sugden, 1986). But it is also true of, say, the often more literary-styled transaction costs literature—as Williamson (1975:1) for example acknowledges. Posner (1992) seems correct in his assessment that ‘new institutional economics is a matter very largely of selecting from the tool bag of modern economic theory those tools (thought) most apt for the study of the narrow though important range of problems, involving © 1997 Tony Lawson
4
5
6 7
impediments to transactions, on which the new institutionalists have chosen to focus’. Or as Pratten (1995) concludes: ‘While TCE [transaction cost economics] does pay greater attention to institutions, etc., than other branches of the contemporary mainstream, what links TCE and economic orthodoxy is a shared essential method’ (1995a:38). Notice that there is also a ‘path-dependency’ (i.e., as well as a ‘path-dependence’) literature. This literature though is closer to the mainstream approach than the ‘path-dependence’ work associated with David. I focus on the latter here because I regard it as the best of its sort, being relatively free of the errors identified throughout the preceding discussion. It thus provides the strongest work to draw upon for both illustrative and comparative purposes There may appear to be a slight of hand here in that I focus first of all upon a specific technology to illustrate a mode of, or approach to, explanation that has been elaborated in the context of the need to identify and understand social structures. However, I think that technology must be understood as designating that realm of determinations where the content is natural but the form is social. And it is because the form is social, so that material objects constitute technological ones (medicines, toothbrushes, football pitches) in part because of human transformative activity, that the arguments so laid out carry through for, and can be illustrated by examples drawn from, technology. A development of this conception of technology must await a further occasion. See also Arthur (1989) and Gould (1987). Instruction in typewriters ‘began to be offered by private business colleges in New York City soon after the first Remington-built machines became available’ (David 1986:39). And in 1892 ‘the radical innovation of an eight-finger typing method was put forward by the proprietress of Longley’s Shorthand and Typewriter Institute, in Cincinnati. Her pamphlet happened to be “adapted to Remington’s perfected typewriters”’ (ibid.: 40). Also in that year, the New York firm of Wyckoff, Seamans & Benedict, having just bought the world-wide sales agency rights for the Remington Type Writer from the firm of E.Remington & Sons, began to promote their product by...[organising] an experimental class to teach eight young women to typewrite.... [Despite] critics’ predictions that typewriting was destined to remain a masculine occupation, every one of the female graduates had found employment quickly. Remington schools for typewriting soon joined the private business and stenographic ‘colleges’ that were now springing up in all the leading cities. (ibid.: 40) When Mrs Longley’s Typewriter Lessons were denounced in the pages of Cosmopolitan Shorthander and Mrs Longley was challenged to prove her case by one Louis Taub who was an advocate of four-finger typing on the Caligraph, the first public speed-typing competition was held, and Mrs Longley’s representative won. David continues: Frank E.McGurrin, the man who [won this competition]...already had won fame in demonstrations before gasping audiences throughout the West, because, in addition to deploying the ‘all-finger’ technique, he had memorized the QWERTY keyboard. We shall never know whether he could have managed the same feat with the 72 keys of the Caligraph machine. (ibid.: 40, 41)
© 1997 Tony Lawson
8
Thus, David elaborates how, amongst other things, the advent of ‘touch’ typing...gave rise to three features of the evolving production system that were crucially important in causing QWERTY to become ‘locked in’ as the dominant keyboard arrangement. These were technical interrelatedness, economics of scale and quasi-irreversibility of investment. (ibid.) Here, technical interrelatedness essentially means that those who purchased a ‘typewriter as an instrument of production’ needed to match this ‘hardware’ to the ‘“software” represented by the touch-typist’s memory of a particular arrangement of the keys’ (ibid.). (Similarly for those wishing to provide instruction in typewriting the incentive also existed to provide machines most related to needs and conditions elsewhere.) The economics of scale in question were basically those ‘exploited by private business colleges that taught young men and women to touch-type through the use of instruction manuals’ (ibid.: 44). Finally, the ‘quasi-irreversibility of investment’ refers to the durability of the tacit skills which typists developed on the basis of the QWERTY keyboard arrangement, and the costs of their being rendered obsolete and/or possibly of retraining for a different system. In David’s words The human capital formed in learning to touch-type is remarkably durable, for the skill resembles that of bicycle-riding or swimming in that once mastered it is long retained at some functional level and may be upgraded rapidly by practice. Moreover, once a specific touch-typing program has been ‘installed in memory’ it becomes quite costly (in retaining time and typing errors) to convert the afflicted typist to a different program. (David, 1986:44–5) At the same time technological developments were such that the cost of converting keyboards from one system to another was falling. Thus, as far as keyboard conversion costs are concerned, an important asymmetry had appeared between the software and the hardware components of the evolving typewriting system: the costs of typewriter software conversion were going up, whereas the costs of the typewriter hardware conversion were coming down. (ibid.: 45)
9
In elaborating this contrast David opposes his approach explicitly to ‘“new institutional economics” [which proposes] to explain many current features of organizations and economic institutions in a thoroughly neoclassical fashion, by citing the respects in which these represented presently efficient solutions to resource allocation problems’ (1994:206–7). Here he concludes that ‘the implicit presumption that institutional arrangements are perfectly malleable seems to be a persistent predilection on the part of many mainstream economists’ (ibid.: 207). And he also opposes his approach to that of some students of economic history [with the early North being singled out] who embraced the principle that to understand institutional change the key lay not in the past so much as in the present and the future, in the sense that institutional arrangements were plastic and, therefore, could
© 1997 Tony Lawson
and would be readily adapted to achieve efficiency wherever people saw that doing so would be to their economic advantage. (ibid.: 206–7) 10 Some conventions persist even long after they no longer serve their original function. It is often suggested, for example, that the shaking of hands on greeting a stranger, that occurs in certain places, is of this nature. David speculates that this behaviour derives from the need to show that one was not bearing a weapon, such as knife or club or a rock. We suppose it to be natural that the convention, once highly functional, at some point became purely customary and persists even though the original function is no longer important. (ibid.: 205) 11 For estimated average growth rates of productivity for various countries over different periods see especially Maddison (1964, 1977, 1989, 1994). Although the UK figure hits a peak, by its own standards, of a 2.8 per cent average annual growth rate of gross domestic product per man-hour over the post-war period of 1951–73, it is noticeable that, even over this period, only the USA of comparable countries is estimated as doing worse: with 2.3 per cent. These figures compare, for example, with Belgium: 4.5 per cent; Canada: 3.1 per cent; Denmark: 4.0 per cent; France: 4.9 per cent; Germany 5.8 per cent; Italy: 5.3 per cent; Japan: 7.5 per cent; Netherlands; 4.1 per cent; Norway: 4.4 per cent; and Sweden: 3.8 per cent. 12 Widely cited examples of the phenomenon in question include not merely the relatively recent developments in computer related technology but many revolutionising products of the last century or so, including the Bessemer, the Gilchrist Thomas, and the Solvey Processes, Mauve Dyes, the Parsons Steam Engine, the Ring Spindle, and so on. To expand merely on the latter, Lazonick (1978) finds that although the invention of the ‘ring frame’ in textile production started to replace the outdated ‘mule’ from the 1880s, and despite the fact that they were mainly supplied by British textile machinery manufactures, by 1907 ring spindles comprised less than 15 per cent of total spindles in the UK, compared with 40 per cent in Germany, 38 per cent in France, 36 per cent in Austria, 65 per cent in Italy, 56 per cent in Russia and about 80 per cent in the USA. Payne (1978:679) refers to a formidable list of new inventions, ideas, and developments more quickly put to practical use by Americans and Europeans than the British, compiled in 1916 by Gray and Turner. Amongst the bestknown accounts of the UK’s slow rate of adoption of new techniques and of comparatively high staffing levels, are Habakkuk (1967) and Landes (1969). For more recent evidence on high staffing levels see, for example, Pratten (1976), and for evidence of slow diffusion see OECD (1970) and Ray (1979) and references provided. 13 There is also much evidence which Stafford does not cite. Lazonick (1978), for example, details the way in which the ring spindle was introduced remarkably slowly in the UK compared to its speed of diffusion in Germany, France, Austria, Italy, Russia and the USA. In analysing this phenomenon Lazonick finds that American ‘employers retained virtually complete control over the labour process throughout the 19th century...’ (1978:50). By contrast, owing to the longer tradition of union organisation in England and the power of the English mule spinners union to perpetuate a division of labour which had © 1997 Tony Lawson
arisen in an earlier era...the English mule spinners were able to protect their particular positions within the labour process and to exercise control over the pace of work until the 1960s... (ibid.: 45) He continues by observing that any large scale attempt to replace mules by rings [in the UK] would have resulted in a large scale confrontation which capitalists could by no means be sure of winning. The emphasis of the era was on industrial conciliation. If there was ‘entrepreneurial failure’ in the Lancashire cotton spinning industry, the effective (if not always explicit) constraints which the sharing of control over work with strong unions placed on the possibilities for redivision of labour and the introduction of new technology had a lot to do with it. (ibid.: 60) Berg (1979) analyses a collection of over fifty documents on the organisation of work in British industry written by different people in the nineteenth century, and finds that ‘the desire of management to gain control over all the functions of the production process was met by working class militancy’ (1979:14). Resistance to new technology was found among type-setters and spinners, for example, and in the cutlery trades, the forge workers, file-making labourers, pattern makers, silk workers and farriers (ibid.). In the US shoe industry, mechanical techniques were introduced ‘early on’, while in the UK ‘the sharp resistance of shoemakers to mechanisation and a reasonably successful outworking system meant that the process took over fifty years’ (ibid.: 169). Craft union practices were also found to be extremely durable even with major changes in both product and process technology. For example, with the transformation from wood to iron shipbuilding in the UK, ‘neither the scale nor the new materials of the industry changed the basic organisation of work, which was still carried on along craft lines’ (ibid.). A further example of resistance is provided by the British engineering industry where: ‘craft orientated engineers were able to slow down...the transformation taking place under the system of Taylorism in the United States and on the continent’. Berg finds that The struggle within all these industries over the rationalisation of production was soon reflected in Britain’s comparative industrial performance. Other nations now took the technological lead. In Britain...workers resistance met the automation, division of labour, novel flow techniques, speed-up and productivity incentives which Marx had predicted would form the characteristics of the higher phases of modern industry. (ibid.: 170) More recent illustrations are provided, for example, in Elbaum and Wilkinson(1979:300–2), McKersie and Hunter (1973:especially 68–160) and Rubery (1978:32). 14 For a more recent comparative study which can be interpreted as reaching the same sort of conclusion see Kirby (1992). 15 I realise that there are risks involved in drawing upon substantive work to which I myself have contributed. Specifically I increase the likelihood of critical realism © 1997 Tony Lawson
(erroneously) being assessed according to a particular substantive account (i.e. by the one in question here). Even so the risks involved in avoiding my own work are probably of greater consequence. For then the door is left open to immediate dismissal on the grounds of that largely irrelevant, yet all too familiar, charge that engagement in methodo-ogy is a response to an incapacity for, or lack of experience of, substantive work. This charge is usually backed with an insistence that methodological arguments should not be taken seriously until their proponents have themselves put them into practice. Here we can recall the observation of Chapter 1 that only recently (October 1992:2) the UK Royal Economics Society Newsletter saw fit to repeat an assertion to this effect made in Irving Fisher’s 1932 Presidential Address to the American Statistical Association: It has long seemed to me that students of the social sciences, especially sociology and economics, have spent too much time in discussing what they call methodology. I have usually felt that the man who essays to tell the rest of us how to solve knotty problems would be more convincing if first he proved out his alleged method by solving a few himself. Apparently those would-be authorities who are forever telling others how to get results do not get any important results themselves. 16 Partly as a result of legal recognition there was immediately a rapid growth in the number of trades unions, with membership of those affiliated to the newly formed Trades Union Congress (TUC) more than doubling from 250,000 in 1869 to 510,000 in 1873. At first these were organisations of skilled workers only, defending acquired customs and their privileged status as crafts people. However, by the 1880s unskilled workers were also able to form stable organisations. These ‘new’ unions of unskilled workers were originally critical of the older craft unions for their inward-looking nature and elitism, and tended to concentrate their energies on national issues. By the turn of the century, however, they too had adopted craft union practices and were possessive of past gains and acquired customs (Clarke and Clements, 1977:12). 17 Notice that although the state, in its endeavours to achieve industrial peace, did not legislate in favour of collective bargaining, it facilitated the latter by way of sponsoring arbitration: the Board of Trade played a guiding role in edging employers into collective bargaining. During wartime, furthermore, employers were encouraged further in this by the Ministry of Munitions. 18 Many of the early British trades unions were primarily associations of subcontractors and it seems likely that the internal contract acted as a means of job control, and structural support, of craft-type organisations (Elbaum and Wilkinson, 1979:290–3; Hobsbawm, 1964:299). In so far as the subcontractor controlled hiring and firing, this power could be utilised to maintain exclusiveness and control of access to jobs. This structure of control may have been preserved, furthermore, because the elimination of contracting was brought about by unskilled workers rather than by employers. In regard to the steel industry, for example, Elbaum and Wilkinson (1979) report how the end of subcontracting was brought about by underhands in separate unions opposed to contracting, but who retained much of the contractors’ control over the pace of work. Scientific management made little or no impact and employers were unable to easily adjust manning levels and were obliged to take disputes through the
© 1997 Tony Lawson
disputes procedure where unions were as successful in establishing precedence over manning and work rules as they were over wage rates. (Elbaum and Wilkinson, 1979:300) 19 Unlike the continental industrial countries of Europe, Britain was neither occupied nor defeated, and it seems that the experience in the Second World War reinforced resistance to centralisation, and actually strengthened job-based worker organisation. For example, the ‘spirit of co-operation’ which the ‘war effort’ demanded led to the establishment of Joint Production Committees, particularly in the engineering industries. These committees were factory-based, with the union members (often shop stewards) elected by ballot, and concerned with all aspects of wartime production within the factory. These committees spread in factories throughout Britain, including in those with no previous tradition of shop stewards (Pelling, 1976:220). Thus, while temporarily performing a major role in ‘national interest corporatism’, these committees formalised and reinforced the institution of shop-floor trade unionism. Moreover, they gave workplace-based organisation an even greater independence from trades union officials than had previously been the case and eventually came to be regarded by workers as the custom or norm. 20 Although the concept of cherodic development (Waddington, 1957, 1972) might be more appropriate here in that it is more general and flexible in scope than the ‘hyper-selection’ associated with ‘path-dependence’ examples (for a discussion see Hodgson, 1993a: especially pp. 205–7). 21 In particular, when, after around 1880, competition was increasing in the UK domestic and export markets, protection was afforded by markets of the Empire. UK exporters found it to their advantage to avoid conflict over an accelerated remoulding of the structure of production, and redirected sales to these protected markets, often with the aid of capital exports. Thus, for example, between 1880 and 1914 the expansion of exports to Empire and other protected markets postponed the necessity to improve the compet-itiveness of such central industries as cotton, textiles and railways (Kindleberger, 1978:224). The Ottawa Agreements in 1932 provide a further example, when British colonies were instructed to give preferential treatment to British exports, and to apply restrictive quotas to others, especially to Japanese textiles, and when similar demands for preferential treatment were made on a number of other countries. Although trade increased, the long-run benefits of these actions are doubtful. As Lewis (1949) notes, the effect of the protection for Britain in these markets was to deflect foreign competition into other markets where on the whole, the British position was in any case less favourable. It seems likely that Britain would have fared better elsewhere if she had not fared so well in protected markets. (1949:84) This view is also borne out, for example, by the experience of the British motor vehicle industry which had been protected since 1915. Although it was successful abroad, this was mainly due to tariff preferences (Youngson, 1960). By 1937, 85 per cent of exported cars went to Empire countries and ‘the industry did not do much, nor did it try to do much, in European markets which it left chiefly to French and German producers’ (1960:107). Thus, in the post-war period when tariffs were reduced, UK producers found it difficult to succeed in the more
© 1997 Tony Lawson
competitive European markets. A further example is the protected steel industry which showed little signs of reducing the discrepancy between its costs and those of its rivals...lip service only was paid to the ideal of economy through more concentrated production...[and] there was still a dearth of modern blast furnaces in 1937. (ibid.: 106) 22 Wedderburn (1972 expresses the situation as follows: In truth there developed after 1871 a clearly enunciated legal framework for British industrial relations, a structure so intimately related to the surrounding social institutions engaged in bargaining and conflict that ‘framework’ suggests too remote a character for laws that grew as threads inherent in the social web. Whatever else, there was plenty of labour law. (Wedderburn, 1972:270) 23 For an account of related developments since 1980, including significant structural transformation and its causes, see Brown and Rea (1995).
19 ECONOMIC POLICY AND INTENDED CHANGE 1
2 3
This disturbance term is always specified to have ‘well-behaved’ (mathematically tractable) properties. The most widely adopted convention is to suppose that actual values are generated according to a normal probability distribution, with zero mean, and constant (or smoothly evolving) variance, with the disturbances themselves being uncorrelated either serially or with the included ‘explanatory’ variables. For an excellent discussion of the whole question of human need, see Doyal and Gough, 1991. See discussion in Lawson (1995) and Runde (1997).
20 ECONOMIC SCIENCE AND PREDICTION 1 2 3
Recent ‘post-modernist’, ‘pragmatist’, hermeneuticist and other tendencies achieve much the same sort of result, of course, through reducing reality to text, language, rhetoric, individual conceptions, or some such (see Sofianou, 1995). And I think we must ultimately include under this head the a priori attitude that methodology must be restricted to the descriptive history of economic method. For an insightful brief sketch of the relevant situation facing economics students I cannot improve upon Strassman’s recent portrayal: To a mainstream economist, theory means model, and model means ideas expressed in mathematical form. In learning how to ‘think like an economist,’ students learn certain critical concepts and models, ideas which typically are taught initially through simple mathematical analyses. These models, students learn, are theory. In more advanced courses, economic theories are presented in more mathematically elaborate models.
© 1997 Tony Lawson
Mainstream economists believe proper models—good models—take a recognizable form: presentation in equations, with mathematically expressed definitions, assumptions, and theoretical developments clearly laid out. Students also learn how economists argue. They learn that the legitimate way to argue is with models and econometrically constructed forms of evidence. While students are also presented with verbal and geometric masterpieces produced in bygone eras, they quickly learn that novices who want jobs should emulate their current teachers rather than deceased luminaries. Because all models are incomplete, students also learn that no model is perfect. Indeed, students learn that it is bad manners to engage in excessive questioning of simplifying assumptions. Claiming that a model is deficient is a minor feat—presumably anyone can do that. What is really valued is coming up with a better model, a better theory. And so, goes the accumulated wisdom of properly taught economists, those who criticize without coming up with better models are only pedestrian snipers. Major scientific triumphs call for a better theory with a better model in recognizable form. In this way economists learn their trade; it is how I learned mine. Therefore, imagine my reaction when I heard feminists from other disciplines apply the term theory to ideas presented in verbal form, ideas not containing even the remotest potential for mathematical expression. ‘This is theory?’ I asked. ‘Where’s the math?’ (Strassman, 1994:154)
© 1997 Tony Lawson
BIBLIOGRAPHY
Abreu, D. and Pearce, D.G. (1984) ‘On the Inconsistency of Certain Axioms on Solution Concepts for Non-Cooperative Games’, Journal of Economic Theory 34: 169–74. Allais, M. (1992) ‘The Economic Science of Today and Global Disequilibrium’, in Baldassarri M. et al. (eds) Global Disequilibrium in the World Economy, Basingstoke: Macmillan. Arestis, P. (1990) ‘Post-Keynesianism: A New Approach to Economics’, Review of Social Economy 48(3): 222–46. ——(1992) The Post-Keynesianism Approach to Economics: An Alternative Analysis of Economic Theory and Policy, Aldershot: Edward Elgar. ——(1996) ‘Post-Keynesian Economics: Towards Coherence’, Cambridge Journal of Economics (20)1: 111–35. Arestis, P. and Chick, V. (1992) Recent Developments in post-Keynesian economics, Aldershot: Edward Elgar. Arrow, K.J. (1968) ‘Economic Equilibrium’, International Encyclopedia of the Social Sciences 4: 376–89. London: Macmillan. Arrow, K.J. and Hahn, F.H. (1971) General Competitive Analysis, San Francisco, Cal.: Holden-Day. Arthur, W.B. (1989) ‘Competing Technologies, Increasing Returns, and Lock-in by Historical Events’, Economic Journal 99(1): 116–31. Backhouse, R.E. (1992) ‘Should we Ignore Methodology?’ Royal Economic Society Newsletter 78: 4–5. ——(1994) New Directions in Economic Methodology, London: Routledge. Baert, P. (1996) ‘Realist Philosophy of Social Science and Economics: a Critical Evaluation’, Cambridge Journal of Economics 20(5). Basu, K. (1990) ‘On the non-existence of a rationality definition for extensive games’, International Journal of Game Theory 19: 33–44. Bell, D. and Kristol, I. (eds) (1981) The Crisis in Economic Theory, New York: Basic Books. Berg, M. (ed.) (1979) Technology and Toil in Nineteenth-Century Britain, London: CSE Books. Bhaskar, R. (1978) A Realist Theory of Science, Hemel Hempstead: Harvester Press (1st edn Leeds, 1975). ——(1979) The Possibility of Naturalism, Hemel Hempstead: Harvester Press (republished with Postscript in 1989—page references to latter edition). ——(1986) Scientific Realism and Human Emancipation, London: Verso. ——(1989) Reclaiming Reality, London: Verso. ——(1993) Dialectic, London: Verso. ——(1994) Plato, etc., London: Verso.
© 1997 Tony Lawson
Binmore, K. (1990) Essays on the Foundations of Game Theory, Oxford: Basil Blackwell. Blackaby, F. (1979) ‘De-industrialization’, NIESR Economic Policy Papers 2, London: Heinemann. Blaug, M. (1980) The Methodology of Economics: Or How Economists Explain, Cambridge: Cambridge University Press. ——(1994) ‘Why I Am Not A Constructivist: Confessions of an Unrepentant Popperian’ in Backhouse, R.E. New Directions in Economic Methodology, London and New York: Routledge. Böhm, S. (1989) ‘Hayek on Knowledge, Equilibrium and Prices: Context and Impact’, Wirtschaftspolitische Blätter 36(2): 201–13. Boland, L.A. (1982) The Foundations of Economics Method, London: George Allen & Unwin. Boylan, T.A. and O’Gorman, P.F. (1995) Beyond Rhetoric and Realism in Economics: Towards a Reformulation of Economic Methodology, London: Routledge. Brown, W. and Rea, D. (1995) ‘The Changing Nature of the Employment Contract’, Scottish Journal of Political Economy 42(3): 363–77. Buchanan, J.M. (1979) What should Economists Do?, Indianapolis: Liberty Press. Bullock, A., Stallybrass, O. and Trombley, S. (eds) (1988) The Fontana Dictionary of Modern Thought (second edn), Hammersmith and London: Fontana Press. Caldwell, B.J. (1982) Beyond Positivism: Economic Methodology in the Twentieth Century, London: George Allen & Unwin. ——(1988a) ‘The Case for Pluralism’, in De Marchi, N. (ed.) The Popperian Legacy in Economics, Cambridge: Cambridge University Press: 231–44. ——(1988b) ‘Hayek’s Transformation’, History of Political Economy 20: 513–41. ——(1989) ‘Post-Keynesian Methodology: An Assessment’, Review of Political Economy 1(1) March 43–64. ——(1990) ‘Does Methodology Matter? How Should it be Practised?’, Finnish Economic Papers 3(1): 64–71. ——(1991) ‘Clarifying Popper’, in Journal of Economic Literature XXIX(1): 1–33. ——(1993) ‘Economic Methodology: Rationale, Foundations, Prospects’, in Mäki, U., Gustafsson, B. and Knudsen, C. (eds) Rationality, Institutions and Economic Methodology, London and New York: Routledge. ——(1994) ‘Two Proposals for the Recovery of Economic Practice’, in Backhouse, R.E. (ed.) New Directions in Economic Methodology. London and New York: Routledge. Canning, D. (1990) Equilibrium, Learning and Economic Dynamics, Cambridge: mimeo. Caplin, A. and Nalebuff, B. (1989) ‘Aggregation and Imperfect Competition: On the Existence of Equilibrium’, London School of Economics Discussion Paper. Carew, A. (1976) Democracy and Government in European Trade Unions, London: George Allen & Unwin. Cartwright, N. (1989) Nature’s Capacities and their Measurement, Oxford: Clarendon Press. Chalmers, A.F. (1978) What is This Thing Called Science?, Milton Keynes: Open University Press. Clarke, T. and Clements, L. (eds) (1977) Trade Unions under Capitalism, Glasgow: Fontana/Collins. Clegg, H.A. (1976) Trade Unionism under Collective Bargaining, Oxford: Basil Blackwell.
© 1997 Tony Lawson
——(1979) The Changing System of Industrial Relations in Great Britain, Oxford: Basil Blackwell. Colander, D. (1994) ‘The Art of Economics by the Numbers’, in Backhouse R.E. (ed.) New Directions in Economic Methodology, London and New York: Routledge. Collier, A. (1994) Critical Realism: An Introduction to Roy Bhaskar’s Philosophy, London: Verso. Commons, J.R. (1934) Institutional Economics—Its Place in Political Economy, New York: Macmillan. (Reprinted with new introduction by M.Rutherford (1990), New Brunswick: Transaction.) Cottrell, A. (1994) ‘Post-Keynesian Monetary Economics’, Cambridge Journal of Economics, 18(6): 587–607. Dangerfield, G. (1936) The Strange Death of Liberal England, London: Putnam. David, P.A. (1985) ‘Clio and the Economics of QWERTY’, American Economic Review (Papers and Proceedings) 75(2): 332–7. ——(1986) ‘Understanding the Economics of QWERTY: The Necessity of History’, in Parker, W.N. (ed.) Economic History and the Modern Economist, Oxford: Blackwell. ——(1992) ‘Path-dependency in economic processes: implications for policy analysis in dynamical system contexts’, Centre for Economic Policy Research, Stanford, Cal.: Stanford University, mimeo. ——(1993) ‘Path-dependence and Predictability in Dynamic Systems with Local Network Externalities: A Paradigm for Historical Economics’, in Foray, D. and Freeman, C. (eds) Technology and the Wealth of Nations, London: Francis Pinter. ——(1994) ‘Why Are Institutions The “Carriers Of History”: Path Dependence and the Evolution of Conventions, Organisations and Institutions’, Structural Change and Economic Dynamics 5(2): 205–20. Davis, J.B. (1987) ‘Three Principles of Post Keynesian Methodology’, Journal of Post Keynesian Economics, 9(4): 552–64. ——(1989) ‘Keynes on Atomism and Organicism’, Economic Journal 99: 1159–72. ——1992) ‘Atomism, Identity Criteria, and Impossibility Logic’, Methodus 4(1): 83– 7. Debreu, G. (1974) ‘Excess Demand Functions’, Journal of Mathematical Economics, 1. Dicey, A.V. (1914) Law and Opinion in England during the Nineteenth Century, London: Macmillan. Dierker, E. (1974) Topological Methods in Walrasian Economics, Berlin: SpringerVerlag. ——(1986) ‘Existence of a Nash Equilibrium in Pure Strategies in an Oligopoly with Price Setting Firms’, University of Bonn: Working Paper. Dobb, M. (1972 [1937]) ‘The Trend of Modern Economics’, in Hunt E.K. and Schwartz J.G. (eds) A Critique of Economic Theory, Harmondsworth: Penguin. (Originally in Political Economy and Capitalism, London: Routledge and Kegan Paul.) Donaldson, P. (1984) Economics of the Real World (third edn), Harmondworth: Penguin. Dopfer, K. (1989) ‘Causality and Consciousness in Economics: Concepts of Change in Orthodox and Heterodox Economics’, in Tool, M.R. and Samuels, W.J. (eds) The Methodology of Economic Thought, New Brunswick and Oxford: Transaction. Dow, S.C. (1990a) ‘Post-Keynesianism as Political Economy: a Methodological Discussion’, Review of Political Economy 2(3): 345–58. © 1997 Tony Lawson
——(1990b) ‘Post-Keynesian Methodology: A Comment’, Review of Political Economy 4(1): 111–13. ——(1990c) ‘Beyond Dualism’, Cambridge Journal of Economics 14(2): 143–57. ——(1991) ‘Post Keynesian School’, in Mair, D. and Miller, A. (eds) Comparative Schools of Economic Thought, Aldershot: Edward Elgar. ——(1995) ‘The Appeal of Neoclassical Economics: Some Insights from Keynes’s Epistemology’, Cambridge Journal of Economics, 16(6): 715–34. ——(1997) Macroeconomic Thought: a Methodological Approach (revised edn; first edn 1985), Oxford and New York: Basil Blackwell. Doyal, L. and Gough, I. (1991) A Theory of Human Need, Basingstoke: Macmillan Education Ltd. Drake, D. et al. (1920) Essays in Critical Realism: A Co-operative Study of the Problem of Knowledge, London: Macmillan & Co. Ltd. Drazen, A. (1980) ‘Recent Developments in Macroeconomic Disequilibrium theory’, Econometric 48(2). Duesenberry, J.S. (1960) ‘Comments’, in National Bureau of Economic Research Demographic and Economic Change in Developed Countries: 231–4. Dunlop, J.T. and Galenson, W. (eds) (1978) Labour in The Twentieth Century, New York: Academic Press. Eatwell, J., Milgate, M. and Newman, P. (eds) (1987) The New Palgrave: A Dictionary of Economics, London: Macmillan. Ekström, M. (1992) ‘Causal Explanation of Social action: the Contribution of Max Weber and of Critical Realism to a Generative View of Causal Explanation in Social Science’, Acta Sociologica, 35: 107–22. Elbaum, B. and Lazonick, W. (1986) The Decline of the British Economy, Oxford: Clarendon Press. Elbaum, B. and Wilkinson, F. (1979) ‘Industrial Relations and Uneven Development: a Comparative Study of the American and British Steel Industries’, Cambridge Journal of Economics 3(3): 275–305. England, P. (1993) ‘The Separative Self: Androcentric Bias in Neoclassical Assumptions’, in Ferber, M.A. and Nelson, J.A. (eds) Beyond Economic Man: Feminist Theory and Economics, 37–53, Chicago: Chicago University Press. Erikson, E.H. (1963) Childhood and Society, New York: Norton. Farmer, M.K. (1982) ‘Rational Action in Economic and Social Theory: Some Misunderstandings’, Archives européenes de sociologie, 23: 179–97. Farmer, M.K. and Matthews, M.L. (1991) ‘Cultural Difference and Subjective Rationality: Where Sociology Connects with the Economics of Technological Choice’, in Hodgson, G. and Screpanti, E. (eds) Rethinking Economics: Markets, Technology and Economic Evolution, 103–16, Aldershot: Edward Elgar. Feynman, R. (1988) Chapter 9, in Davies, P.C.W. and Brown, J. (eds) Superstrings: a Theory of Everything, Cambridge: Cambridge University Press. Fine, B. and Harris, L. (1985) The Peculiarities of the British Economy, London: Lawrence and Wishart. Fish, S. (1985) ‘Consequences’, in Mitchell, W.J.T. (ed.) Against Theory, 106–31, Chicago: University of Chicago Press. Fisher, I. (1932) ‘Presidential Address to American Statistical Association’, reprinted in Royal Economic Society Newsletter, 79 (1992): 2. Fleetwood, S. (1995) Hayek’s Political Economy: the Socio-economics of Order, London: Routledge. ——(1996a) ‘Order without Equilibrium: a Critical Realist Interpretation of Hayek’s Notion of Spontaneous Order’, Cambridge Journal of Economics forthcoming. ——(1996b) ‘Hayek III: The Necessity of Social Rules of Conduct’, in Frowen, S. (ed.) © 1997 Tony Lawson
Hayek the Economist and Social Philosopher, Basingstoke and London: Macmillan. Folbre, N. (1982) ‘Exploitation Comes Home: A Critique of the Marxian Theory of Family Labour’, Cambridge Journal of Economics 6: 317–29. ——(1991) ‘The Unproductive Housewife: Her Evolution in Nineteenth-Century Economic Thought’, Signs: Journal of Women in Culture and Society 16(3): 463–84. ——(1993) ‘How Does She Know? Feminist Theories of Gender Bias in Economics’, History of Political Economy 25(1): 167–84. ——(1994) Who Pays for the Kids?: Gender and the Structures of Constraint, London: Routledge. Foss, N.J. (1994) ‘Realism and Evolutionary Economics’, Journal of Social and Evolutionary Systems, 17(1): 21–40. Friedman, A.L. (1977) Industry and Labour, London: Macmillan. Friedman, M. (1953) ‘The Methodology of Positive Economics’, in Essays in Positive Economics, Chicago: University of Chicago Press. Frowen, S. (ed.) (1996) Hayek the Economist and Social Philosopher: A Critical Retrospect, Basingstoke: Macmillan. Gee, J.M.A. (1991) ‘The Neoclassical School’, in Mair, D. and Miller, A.G. (eds) A Modern Guide to Economic Thought: An Introduction to Comparative Schools of Thought in Economics, Aldershot: Edward Elgar. Gerschenkron, A. (1955) ‘Comment on Hoselitz, B.F. (ed.) Entrepreneurship and Capital Formation in France and Britain since 1700 ’, in NBER Capital Formation and Economic Growth, Princeton, NJ: Princeton University Press. Giddens, A. (1984) The Constitution of Society, Cambridge: Polity Press. Goffman, E. (1963) Behaviour in Public Places, New York: Free Press. Gould, S.J. (1987) ‘The Panda’s Thumb of Technology’, Natural History, 1: 14–23. Gould, J. and Kolb, W.L. (1964) A Dictionary of the Social Sciences, London: Tavistock Publications. Grandmont, J.M. (1987) ‘Distributions of Preferences and Law of Demand’, Econometrica, 55(1). Granick, D. (1972) Managerial Comparisons of Four Developed Countries: France, Britain, United States and Russia, Cambridge, Mass.: MIT Press. Green, M. (1988) Chapter 4, in Davies, P.C.W. and Brown, J. (eds) Superstrings: a Theory of Everything, Cambridge: Cambridge University Press. Gregory, D. and Urry, J. (1985) Social Relations and Spacial Structures, Basingstoke: Macmillan. Gunter, H. and Leminsky, G. (1978) ‘The Federal Republic of Germany’, in Dunlop, J.T. and Galenson, W. (eds) Labour in the Twentieth Century, New York: Academic Press. Haavelmo, T. (1944) ‘The Probability Approach in Econometrics’, Econometrica 12: 1–118, supplement. Habakkuk, H.J. (1967) American and British Technology in the Nineteenth Century, Cambridge: Cambridge University Press. Hacking, I. (1983) Representing and Intervening: Introductory Topics in the Philosophy of Natural Science, Cambridge: Cambridge University Press. Hahn, F. (1984) Equilibrium and Macroeconomics, Oxford: Basil Blackwell. ——(1985) ‘In Praise of Economic Theory’, The 1984 Jevons Memorial Fund Lecture, London: University College. ——(1991) ‘The Next Hundred Years’, Economic Journal 101(1): 47–50. ——(1992a) ‘Reflections’, Royal Economic Society Newsletter 77. © 1997 Tony Lawson
——(1992b) ‘Answer to Backhouse: Yes’, Royal Economic Society Newsletter 78: 5. ——(1994) ‘An Intellectual Retrospect’, Banca Nazionale del Lavoro Quarterly Review 245–58. Hall, P. (1986) ‘The State and Economic Decline’, 286–302 in Elbaum and Lazonick (eds) The Decline of the British Economy, Oxford: Clarendon Press. Hamilton, W.H. (1932) ‘Institution’, in Selighan, R.A. and Johnson, A. (eds), Encyclopedia of the Social Sciences, 8: 84–9. Hamouda, O. and Harcourt, G. (1989) ‘Post-Keynesianism: From Criticism to Coherence’, in Pheby, J. (ed.) New Directions in Post-Keynesian Economics, Aldershott: Edward Elgar. Hands, D.W. (1985) ‘Karl Popper and Economic Methodology: a New Look’, Economics and Philosophy 1(1): 83–99. ——(1993a) Testing, Rationality, and Progress: Essays on the Popperian Tradition in Economic Methodology, Lanham: Rowman and Littlefield. ——(1993b) ‘Popper and Lakatos in Economic Methodology’, in Mäki, U., Gustafsson, B. and Knudsen, C. (eds) Rationality, Institutions and Economic Methodology, London: Routledge. ——(1994) ‘The Sociology of Scientific Knowledge: Some Thoughts on the Possibilities’, in Backhouse, R.E. (ed.) New Directions in Economic Methodology, London and New York: Routledge. Hansen, G.D. (1985) ‘Indivisible Labour and the Business Cycle’, Journal of Monetary Economics 16: 309–28. Hansen, G.D. and Sargent, T.J. (1988) ‘Straight Time and Overtime in Equilibrium’, Journal of Monetary Economics 21: 281–308. Hanson, N.R. (1958) Patterns of Discovery: An Inquiry into the Conceptual Foundations of Science, Cambridge: Cambridge University Press. Harcourt, G. (1988) ‘Post-Keynesian economics’, entry for Eatwell J., Milgate M. and Newman P. (eds) The New Palgrave: A Dictionary of Economics, 924–8, London: Macmillan. ——(1992) ‘A Post-Keynesian Comment’, Methodus 4(1): 30. Harding, S. (1995) ‘Can Feminist Thought Make Economics More Objective?’, Feminist Economics 1(1): 7–32. Hartmann, H. (1981) ‘The Family as the Locus of Gender, Class, and Political Struggle: The Example of Housework’, Signs 6(3): 366–94. Hausman, D.M. (1989) ‘Economic Methodology in a Nut Shell’, Journal of Economic Perspectives 3(2): 115–28. ——(1992) The Inexact and Separate Science of Economics, Cambridge: Cambridge University Press. ——(1994) ‘Kuhn, Lakatos and the Character of Economics’, in Backhouse R.E. (ed.) New Directions in Economic Methodology, London and New York: Routledge. Hawkins, K. (1971) ‘The Decline of Voluntarism’, Industrial Relations Journal 2(2). Hayek, F.A. (1937) ‘Economics and Knowledge’, Economica IV: 33–54. ——(1942–4) ‘Scientism and the Study of Society’, Economica (reprinted in The Counter-Revolution of Science, Indianapolis: Liberty Press). ——(1943) ‘The Facts of the Social Sciences’, Ethics LIV(1): 1–13. ——(1967) Studies in Philosophy, Politics and Economics, London and Henley: Routledge and Kegan Paul. ——(1994) Hayek on Hayek: An Autobiographical Dialogue, Kresge, S. and Wenar, L. (eds), London: Routledge.
© 1997 Tony Lawson
Hempel, C. (1965) ‘The Function of General Laws in History’, in Aspects of Scientific Explanation, New York: Free Press. Hendry, D.F. (1980) ‘Econometrics—Alchemy or Science?’, Economica 47: 387–406. ——(1983) ‘On Keynesian Model Building and the Rational Expectations Critique: a Question of Methodology’, Cambridge Journal of Economics 7 (1): 69–76. ——(1987) ‘Econometric Methodology: a Personal Perspective’, in Bewley, T.F. (ed.) Advances in Economics: Fifth World Congress, Vol. II, Cambridge: Cambridge University Press: 29–48. Hendry, D.F., Leamer, E.E. and Poirier, D.J. (1990) ‘The ET dialogue; A Conversation on Econometric Methodology’, Econometric Theory 6: 171–261. Hesse, M. (1974) The Structure of Scientific Inference, Cambridge: Cambridge University Press. Hicks, G.D. (1938) Critical Realism: Studies in the Philosophy of Mind and Nature, London: MacMillan & Co. Ltd. Hicks, J. (1939) Value and Capital, Oxford: Oxford University Press. ——(1986) ‘Is Economics a Science?’, in Baranzini, M. and Scazzieri, R. (eds) Foundations of Economics (91–101 ) Oxford: Basil Blackwell. Hildenbrand, W. (1983a) ‘On the “Law of Demand”’, Econometrica 51(4): 997– 1020. ——(1983b) Introduction, in Mathematical Economics: Twenty Papers of Gérard Debreu, Cambridge: Cambridge University Press. Hobsbawm, E.J. (1964) Labouring Men, London: Weidenfeld and Nicolson. Hodgson, G.M. (1985) ‘Persuasion, Expectations and the Limits to Keynes’, in Lawson, T. and Pesaran, M.H. (eds) Keynes’ Economics: Methodological Issues, Beckenham: Croom Helm. ——(1986) ‘Behind Methodological Individualism’, Cambridge Journal of Economics 10(3): 21–4. ——(1987) ‘Economics and Systems Theory’, Journal of Economic Studies 14(4): 65–86. ——(1988) Economics and Institutions: A Manifesto for a Modern Institutional Economics, Cambridge: Polity Press. ——1989) ‘Post-Keynesianism and Institutionalism: The Missing Link’, 94–123 in Pheby, J. (ed.) New Directions in Post-Keynesian Economics, Aldershot: Edward Elgar. ——(1992) ‘Thorstein Veblen and Post-Darwinian Economics’, Cambridge Journal of Economics 16(3): 285–302. ——(1993a) Economics and Evolution: Bringing Life Back into Economics, Cambridge: Polity Press. ——(1993b) ‘Theories of Economic Evolution: a Preliminary Taxonomy’, Manchester School of Economic and Social Studies 125–43. ——(1993c) ‘Institutional Economics: Surveying the “Old” and the “New”’, Metroeconomica 44(1): 1–28. ——(1994) ‘Institutionalism, ‘Old’ and ‘New”, in Hodgson, G., Samuels, W.J. and Tool, M.R. (eds) The Elgar Companion to Institutional an Evolutionary Economics A-K: 397–402. Aldershot: Edward Elgar Hodgson, G. and Screpanti. E. (1991) (eds) Rethinking Economics: Markets, Technology and Economic Evolution, Aldershot: Edward Elgar. Hodgson, G., Samuels, W.J. and Tool, M. (1994) The Elgar Companion to Institutional and Evolutionary Economics, Aldershot: Edward Elgar.
© 1997 Tony Lawson
Hoover, K.D. (1988) The New Classical Macroeconomics: A Sceptical Inquiry, Oxford: Basil Blackwell. ——(1994) ‘Pragmatism, Pragmaticism and Economic Method’, in Backhouse R.E. (ed.) New Directions in Economic Methodology, London and New York: Routledge. ——(1995a) ‘Facts and Artifacts: Calibration and the Empirical Assessment of Realbusiness-cycle Models’, Oxford Economic Papers 47(1). ——(1995b) ‘Why does Methodology Matter for Economics?’, Economic Journal 105(430): 715–35. Humphries, J. (1994) ‘Review of Ferber, M. and Nelson, J. (eds) Beyond Economic Man: Feminist Theory and Economics’, Journal of History 54: 483, 484. ——(1995) Gender and Economics, The International Library of Critical Writings in Economics 45, Aldershot: Edward Elgar. Hutchison, T.W. (1981) The Politics and Philosophy of Economics: Marxians, Keynesians and Austrians, Oxford: Basil Blackwell. ——(1994) ‘Ends and Means in the Methodology of Economics’, in Backhouse, R.E. (ed.) New Directions in Economic Methodology, London and New York: Routledge. Ingrao, B. and Israel, G. (1985) ‘General Equilibrium Theory: A History of Ineffectual Paradigmatic Shifts’, Fundamenta Scientae 6(1): 1–45 and 6(2): 89–125. ——(1990) The Invisible Hand, Cambridge Mass.: MIT Press. International Labour Office (1974) Collective Bargaining in Industrialised Market Economics, Geneva: ILO. Janssen, M.C.W. (1993) Microfoundations: A Critical Inquiry, London and New York: Routledge. Jennings, A. and Waller, W. (1994) ‘Evolutionary Economics and Cultural Hermeneutics: Veblen, Cultural Relativism and Blind drift’, Journal of Economic Issues XXVIII(4): 997–1030. ——(1995) ‘Cultural: Core Concept Reaffirmed’, Journal of Economic Issues XXIX(2): 407–18. Jevons, W.S. (1965 [1871]) The Theory of Political Economy, New York: A.M.Kelly. Kaldor, N. (1978a) ‘Productivity and Job Opportunities’, Speech to the House of Lords, Hansard. ——(1978b) Further Essays on Applied Economics, London: Duckworth. ——(1985) Economics Without Equilibrium, Cardiff: University College Press. Kanth, R.K. (1991) ‘Economic Theory and Realism: Outlines of a Reconstruction’, Methodus 3(2): 37–45. ——(1992) Capitalism and Social Theory: The Science of Black Holes, Armonk, NY and London: Sharpe. Kapp, K.W. (1976) ‘The Nature and Significance of Institutional Economics’, Kyklos 29: 209–32 (reprinted in Samuels, W. (ed.) Institutional Economics 1, (1976): 68–91, Aldershot: Edward Elgar—page references are to the former). Kardiner, A. (1939) The Individual and His Society, New York: Columbia University Press. Kay, J. (1995) ‘Cracks in the Crystal Ball’, Financial Times 29 September. Kehoe, T.J., Levine D.K. and Romer P.M. (1990) ‘Determinacy of Equilibria in Dynamic Models with Finitely Many Consumers’, Journal of Economic Theory 50: 1–21. Kennedy, P. (1993) ‘British Economic Decline: Bourgeois Methodology versus Marxism’, Glasgow: mimeo. © 1997 Tony Lawson
——(1994) The Decline of Capitalism and the Rise of Labourism, Chapter 1, PhD, Glasgow University. Keuzenkamp, H.A. and McAleer, M. (1995) ‘Simplicity, Scientific Inference and Econometric Modelling’, The Economic Journal, 105(428): 1–21. Keynes, J.M. (1905) ‘A Theory of Beauty’, in the unpublished writings of J.M. Keynes, (1904–12), The Provost and Scholars of King’s College, Cambridge. ——(1973a) ‘The Collected Writings of John Maynard Keynes: The General Theory’, Royal Economic Society VII. ——(1973b) ‘The Collected Writings of John Maynard Keynes: The General Theory and After: Part II Defence and Development’, Royal Economic Society XIV. Keynes, J.N. (1980) The Scope and Method of Political Economy, London: Macmillan. Kilpatrick, A. and Lawson, T. (1980) ‘On the Nature of Industrial Decline in the UK’, Cambridge Journal of Economics 4: 85–102 (reprinted in Coates D. and Hillard J. (eds) The Economic Decline of Modern Britain, Harvester Press, 1986; also reprinted in Feinstein, C., (ed.) The Economic Development of Modern Europe Since 1980, Cheltenham: Edward Elgar (forthcoming)). Kindleberger, C.P. (1964) Economic Growth in France and Britain 1851–1950, Cambridge, Mass.: Harvard University Press. ——(1978) Economic Response: Comparative Studies in Trade, Finance and Growth, London: Harvard University Press. King, R.G. and Plosser, C.I. (1989) ‘Real Business Cycle Models and the Test of the Adelmans’, paper presented to the Federal Reserve Bank of San Francisco’s Fall Academic Conference. Kirby, M.W. (1992) ‘Institutional rigidities and economic decline: reflections on the British experience’, Economic History Review XLV(4): 637–60. Kirman, A. (1989) ‘The Intrinsic Limits of Modern Economic Theory: The Emperor has no clothes’, Economic Journal 99(395). Kirman, A. and Koch, K.J. (1986) ‘Market excess demand in exchange economies with identical preferences and collinear endowments’, Review of Economic Studies 53(3). Klein, L. (1946) ‘Macroeconomics and the Theory of Rational Behaviour’, Econometrica 14: 93–108. Knudsen, C. (1993) ‘Equilibrium, Perfect Rationality and the Problem of SelfReference in Economics’, in Mäki, U., Gustafsson, B. and Knudsen, C. (eds) Rationality, Institutions and Economic Methodology, London and New York: Routledge Koertge, N. (1975) ‘Popper’s metaphysical research program for the human sciences’, Inquiry 18: 437–62. Kydland, F.E. (1984) ‘Labour-force Heterogeneity and the Business Cycle’, in Brunner, K. and Meltzer, A.H. (eds) Essays on Macroeconomic Implications of Financial and Labour Markets and Political Processes, Carnegie-Rochester Conference Series on Public Policy, 21. Kydland, F.E. and Prescott, E.C. (1982) ‘Time to Build and Aggregate Fluctuations’, Econometrica 50: 1345–70. ——(1988) ‘The Workweek of Capital and Its Cyclical Implications’, Journal of Monetary Economics 21: 243–60. Landes, S. (1969) The Unbound Prometheus, Cambridge: Cambridge University Press. Langlois, R.N. (1986) ‘The New Institutional Economics: an Introductory Essay’, in
© 1997 Tony Lawson
Langlois, R.N. (ed.) Economics as a Process: Essays in the New Institutional Economics, Cambridge: Cambridge University Press. ——(1989) ‘What was Wrong with the Old Institutional Economics (and What is Still Wrong with the New)?’, Review of Political Economy 1(3): 270–98. Langlois, R.N. and Csontos, L. (1993) ‘Optimization, Rule-Following, and the Methodology of Situational Analysis’, in Mäki, U., Gustafsson, B. and Knudsen, C. (eds) Rationality, Institutions and Economic Methodology, London and New York: Routledge Latsis, S.J. (ed.) (1972) ‘Situational determinism in economics’, British Journal for the Philosophy of Science 23: 207–45. ——(1976a) Method and Appraisal in Economics, Cambridge: Cambridge University Press. ——(1976b) ‘A research programme in economics’, in Method and Appraisal in Economics, Cambridge: Cambridge University Press: 1–41. Lawson, C. (1994) ‘The transformation model of social activity and economic activity: a reinterpretation of the work of J.R.Commons’, Review of Political Economy 6(2). Lawson, C., Peacock, M. and Pratten, S. (1996), ‘Realism, Underlabouring and Institutions’, Cambridge Journal of Economics 20(1): 137–51. Lawson, T. (1981) ‘Paternalism and Labour Market Segmentation Theory’, in Wilkinson F.S. (ed.) Dynamics of Labour Market Segmentation, London: Academic Press. ——(1987) ‘The Relative/Absolute Nature of Knowledge and Economic Analysis’, Economic Journal 97(388): 951–70. ——(1992) ‘Realism, Closed Systems and Friedman’, Research in the History of Economic Thought and Methodology 10: 149–69. ——(1994a) ‘Realism and Hayek: A Case of Continuing Transformation’, in Colonna, M., Hagemann, H. and Hamouda, O.F. (eds) Capitalism, Socialism and Knowledge: The Economics of F.A.Hayek, Vol. II, Aldershot: Edward Elgar. ——(1994b) ‘The Nature of Post Keynesianism and its Links to other Traditions’, Journal of Post Keynesian Economics 16(4): 503–38 (reprinted in Prychitko D.L. (ed.) Why Economists Disagree: An Introduction to the Contemporary Schools of Thought, New York: State University of New York Press, 1996). ——(1995) ‘Expectations and Economics’, in Dow S. and Hillard J. (eds) Keynes, Knowledge and Uncertainty, Aldershot: Edward Elgar. ——(1996) ‘Development in Hayek’s Social Theorising’, in Frowen S. (ed.) Hayek, The Economist and Social Philosopher: A Critical Retrospect, Basingstoke and London: Macmillan Press. Lazonick, W. (1978) ‘The Division of Labour and Machinery: The Development of British and US Cotton Spinning’, Discussion Paper 620, Cambridge, Mass.: Harvard University. Leamer, E.E. (1978) ‘Specification Searches: Ad hoc Inferences with Non-experimental Data’, New York: John Wiley and Sons. ——(1983) ‘Let’s Take the Con out of Econometrics’, American Economic Review 73(1): 31–43. Leontief, W. (1982) Letter in Science 217: 104–7. Leplin, J. (ed.) (1984) Scientific Realism, Berkeley, Cal.: University of California Press. Levy, D.J. (1981) Realism: an Essay in Interpretation and Social Reality, Manchester: Carcanet Press.
© 1997 Tony Lawson
Lewis, R. (1976) ‘The Historical Development of Labour Law’, British Journal of Industrial Relations XIV(1): 1–17. Lewis, W.A. (1949) Economic Survey 1919–39, London: George Allen & Unwin. Lipton, P. (1991) Inference to the Best Explanation. London: Routledge. Lloyd, C. (1993) The Structures of History, Oxford: Blackwell Publishers Loasby, B.J. (1976) Choice, Complexity and Ignorance: An Enquiry into Economic Theory and the Practice of Decision Making, Cambridge: Cambridge University Press. Locke, J. (1985 [1690]) ‘An Essay Concerning Human Understanding’, An Abridgment Selected and Edited by Yolton, J.W., London and Melbourne: Dent. Lucas, R.E. (1976) ‘Econometric Policy Evaluation: A Critique’, in Brunner K. and Meltzer A.H. (eds.) The Phillips Curve and Labour Markets (reprinted in Lucas R. E. Studies in Business-Cycle Theory (1981) Oxford: Basil Blackwell). ——(1986) ‘Adaptive Behaviour Economic Theory’, Journal of Business 59(4). Luethy, H. (1955) France against Herself, New York: Praeger. McAleer, M. (1994) ‘Sherlock Holmes and the Search for Truth: A Diagnostic Tale’, Journal of Economic Surveys 8(4) December: 317–70, University of Western Australia. McCloskey, D.N. (1983) ‘The Rhetorics of Economics’, Journal of Economic Literature xxi (June): 481–517. ——(1986) The Rhetoric of Economics, Brighton: Wheatsheaf Books. McKersie, R.G. and Hunter, L.C. (1973) Pay, Productivity and Collective Bargaining, London: Macmillan. McLean, J.M. and Rush, H.J. (1978) ‘The Impact of Microelectronics on the UK: A Suggested Classification and Illustrative Case Studies’, SPRU Occasional Paper Series No.7, Sussex University. McMullin, E. (1984) ‘A Case for Scientific Realism’, in Leplin, J. (ed.) Scientific Realism, Berkeley, Cal.: University of California Press. Maddison, A. (1964) Economic Growth in the West, London: Allen & Unwin. ——(1977) ‘Phases of Capitalist development’, Banca Nazionale del Lavoro Review. ——(1989) The World Economy in the Twentieth Century, Paris: OECD Development Centre. ——(1994) Monitoring the World Economy 1820–1992, Paris: OECD Development Centre Madison, G.B. (1990) ‘Getting Beyond Objectivism: the Philosophical Hermeneutics of Gadamer and Ricoeur’, in Lavoi, D. (ed.) Economics and Hermeneutics, London: Routledge. Mäki, U. (1988) ‘How to Combine Rhetoric and Realism in the Methodology of Economics’, Economics and Philosophy, 4(1): 89–109. ——(1990a) ‘Scientific Realism and Austrian Explanation’, Review of Political Economy, 2(3): 310, 344. ——(1990b) ‘Methodology of Economics: Complaints and Guidelines’, Finnish Economic Papers 3(1): 77–84. ——(1992a) ‘On the Method of Isolation in Economics’, Poznan Studies in the Philosophy of the Sciences and the Humanities 26: 319–54. ——(1992b) ‘Friedman and Realism’, Research in the History of Economic Thought and Methodology 10: 171–95. ——(1993a) ‘Economics with Institutions: Agenda for Methodological Enquiry’, in Mäki, U., Gustafsson, B. and Knudsen, C. (eds) Rationality, Institutions and Economic Methodology, London and New York: Routledge. © 1997 Tony Lawson
——(1993b) ‘Social Theories of Science and the Fate of Institutionalism in Economics’, in Mäki, U., Gustafsson, B. and Knudsen, C. (eds) Rationality, Institutions and Economic Methodology, London and New York: Routledge. ——(1994) ‘Isolation, Idealization and Truth in Economics’, Poznan Studies in the Philosophy of the Sciences and the Humanities 38: 147–68. Mäki, U., Gustafsson, B. and Knudsen, C. (eds) (1993) Rationality, Institutions and Economic Methodology, London and New York: Routledge. Mantel, R. (1974) ‘On the Characterization of Aggregate Excess Demand’, Journal of Economic Theory 7(3). ——(1976) ‘Homothetic Preferences and Community Excess Demand Functions’, Journal of Economic Theory 12(2). ——(1977) ‘Implications of Microeconomic Theory for Community Excess Demand Functions’, in Intriligator, M.D. (ed.) Frontiers of Quantitative Economics Vol. IIIA, Amsterdam: North Holland. De-Marchi, N. (1992) Post-Popperian Methodology of Economics: Recovering Practice, Norwell, Mass, and Dordrecht: Kluwer Academic Press. Marshall, A. (1920 [1890]) Principles of Economics, London: Macmillan Press. ——(1898) ‘Distribution and Exchange’, Economic Journal 8: 37–59. Marx, K. (1973) Grundrisse: Foundations of the Critique of Political Economy, translated by Nicolaus, M., Harmondsworth: Penguin Books in association with New Left Review. ——(1977) The Eighteenth Brumaire of Louis Bonaparte, in McLellan, D. (ed.), Karl Marx: Selected Writings, Oxford: Oxford University Press. (Reprinted from Marx, K. and Engels, F., Selected Works, Vol. 1, Moscow: Lawrence and Wishart Ltd.) Meikle, S. (1985) Essentialism in the Thought of Karl Marx, London: Duckworth. Menger, C. (1985 [1883]) Investigations Into The Method Of The Social Sciences With Special Reference To Economics, version edited by Louis Schneider, translated by Francis J.Nock, with introduction by Lawrence White, New York: New York University Press. Mill, J.S. (1844) Essays on Some Unsettled Questions of Political Economy, London: J.W.Parker. Miller, R.W. (1987) Fact and Method: Explanation, Confirmation and Reality in the Natural and the Social Sciences, Princeton, NJ: Princeton University Press. Miller, E.S. (1989) ‘Comment of Boettke and Samuels: Austrian and Institutionalist Economics’, Research in the History of Economic Thought and Methodology Vol. 6: 151, 158. Minchenton, W.E. (1957) British Tinplate Industry, Oxford: Clarendon Press. Mirowski, P. (1989) More Heat Than Light: Economics as Social Physics, Physics as Nature’s Economics, Cambridge: Cambridge University Press. ——(1994) ‘What Are The Questions?’, in Backhouse R.E. (ed.) New Directions in Economic Methodology, London and New York: Routledge. Musgrave, A. (1981) ‘“Unreal assumptions” in Economic Theory: The F-Twist Untwisted’, Kyklos 34: 377–87. Nadeau, R. (1995) ‘Economics and Intentionality’, in Manon, M. and Cohen, R.S. (eds) Québec Studies in Philosophy of Science, Dordrecht/Boston/London: Kluwer Academic Publishers (page references to mimeo: Université du Québec à Montreal). Nash, J.F. Jr. (1950) ‘Equilibrium Points in n-person Games’, Proceedings of the National Academy of Sciences of the United States 46: 48–9. ——(1951) ‘Non-cooperative Games’, Annals of Mathematics 54: 289–95. © 1997 Tony Lawson
Neale, W.C. (1994) ‘Institutions’, in Hodgson, G.M., Samuels, W.J. and Tool, M.R. (eds), The Elgar Companion to Institutional and Evolutionary Economics, Vol. A-K: 402–6. Nelson, J. (1993a) ‘Value-Free or Valueless? Notes on the Pursuit of Detachment in Economics’, History of Political Economy 25(1): 121–45. ——(1993b) ‘The Study of Choice or the Study of Provisioning? Gender and the Definition of Economics’, in Ferber, M.A. and Nelson, J. (eds) Beyond Economic Man 23–36, Chicago University Press. ——(1996) Feminism, Objectivity and Economics, London and New York: Routledge. Northover, P. (1995) On Explaining Economic Growth: A Methodological Enquiry unpublished Ph.D. dissertation, Cambridge. Nowak, L. (1980) The Structure of Idealization, Towards a Systematic Interpretation of the Marxian Idea of Science, Dordrecht: Reidel. Oakeshott, M. (1962) Rationalism in Politics and Other Essays, London: Methuen. O’Driscoll, G.P. and Rizzo, M.J. (1996) The Economics of Time and Ignorance, London: Routledge, second edn (first edn 1985, Oxford: Basil Blackwell). OECD (1970) ‘Gaps on Technology: Comparisons between Member Countries in Education, Research and Development, Technological Innovation’, International Economics Exchanges, Paris: OECD. Ollman, B. (1993) Dialetical Investigations, New York and London: Routledge. O’Neill, J. (1996) ‘Hayek’, in Frowen, S. (ed.) Hayek the Economist and Social Philosopher: A Critical Retrospect, Basingstoke and London: Macmillan (page references to mimeo version). Ormerod, P. (1994) The Death of Economics, London: Faber and Faber. Outhwaite, W. (1987) New Philosophies of Social Science: Realism, Hermeneutics and Critical Theory, Basingstoke and London: Macmillan Education Ltd. Outhwaite, W. and Bottomore, T. (eds) (1993) The Blackwell Dictionary of TwentiethCentury Social Thought, Oxford: Basil Blackwell. Payne, P.L. (1978) ‘Industrial Entrepreneurship and Management in Great Britain’, in Mathias, P. and Postan, M.M. (eds) Cambridge Economic History in Europe Vol. 7, Part I, Cambridge: Cambridge University Press. Peacock, M. (1993) ‘Hayek: Realism and Spontaneous Order’, Journal for the Theory of Social Behaviour 23(3): 249, 264. Peck, M.J. (1968) ‘Science and Technology’, in Caves, R.E. et al. Britain’s Economic Prospects, Washington DC: Brookings Institution. Peirce, C.S. (1867) in Hartshorne, C. and Weiss, P. (eds) Collected Papers of Charles Sanders Peirce, 1931–35, Vols 1–6, Cambridge, Mass.: Harvard University Press. ——(1966) ‘Selected Writings: Values in a Universe of Chance’, ed. with an introduction and notes by Weiner, P.P., New York: Dover Publications. ——(1931–58) Collected Papers of Charles Saunders Peirce, 1–8, in Burks, A.W., Hartshorne, C. and Weiss, P. (eds) Cambridge, Mass.: Belknap Press. Pelling, H. (1976) A History of British Trade Unionism, London: Macmillan. Penn, R. (1978) ‘Skilled Manual Workers in the Labour Process’, presented at Nuffield Conference on the Labour Process, mimeo. Pesaran, M.H. (1987) ‘Econometrics’, in Eatwell et al. (eds) 2: 8–19. Pheby, J. (1988) Methodology and Economics: A Critical Introduction, Basingstoke: Macmillan Press. Popper, K. (1957) The Poverty of Historicism, New York: Harper & Row. ——(1965) Conjectures and Refutations: The Growth of Scientific Knowledge, New York: Harper & Row. © 1997 Tony Lawson
——(1966) The Open Society and its Enemies, fifth edn, Princeton, NJ: Princeton University Press. ——(1990) A World of Propensities, Bristol: Thoemmes Antiquarian Books. (1992 [1974]) Unended Quest: An Intellectual Autobiography, London: Routledge (first published as Autobiography by Karl Popper, 1974). Posner, R.A. (1992) ‘The New Institutional Economics Meets Law and Economics’, paper presented at the Tenth International Seminar on the New Institutional Economics, Wallerfanger/Saar, Germany. Pratt, A.C. (1995) ‘Putting Critical Realism to Work: the Practical Implications for Geographical Research’, Progress in Human Geography 19(1): 61–74. Pratten, C.F. (1976) Labour Productivity Differentials Within International Companies, Cambridge: Cambridge University Press. Pratten, S. (1993) ‘Structure, Agency and Marx’s Analysis of the Labour Process’, Review of Political Economy 5(4): 403–26. ——(1994) ‘Marshall on Tendencies, Equilibrium and the Statical Method’, Chapter 2 of PhD thesis: Forms of Realism, Conceptions of Science and Approaches to Industrial Organisations, Cambridge: forthcoming in History of Political Economy. ——(1995) ‘Operationalizing Transaction Cost Economics: Implications and Alternatives’, De Montfort University: mimeo. Quine, W.V.O. (1953) From a Logical Point of View, Cambridge, Mass.: Harvard University Press. Rattaggi, M.L. (1995) ‘Banking Safety and Regulation in the US: A Critical Realist Perspective’, Quaderni Di Ricerca, No. 19, Centro di Studi Bancari. Ray, G.F. (1979) Comment on Freeman, C., ‘Technical Innovation and British Trade Performance’, in Blackaby, F. (ed.) De-industrialisation, London: Heinemann. Reder, M.W. (1982) ‘Chicago Economics: Permanence and Change’, Journal of Economic Literature 35: 1–38. Reuten, G. and Williams, M. (1988) The Value-Form Determination of Economic Policy: a Dialectical Theory of Economy, Society and State in the Capitalist Epoch, Amsterdam: B.R.Grüner B.V. Publishing Company. Rizzo, M.J. (1990) ‘Hayek’s Four Tendencies Toward Equilibrium’, Cultural Dynamics 3(1): 12, 31. Rizvi, S.A.T. (1994a) ‘Game Theory To The Rescue?’, Contributions to Political Economy 13: 1–28. ——(1994b) ‘The Microfoundations Project in General Equilibrium Theory’, Cambridge Journal of Economics 18(4): 357–78. Robbins, L. (1932) An Essay on the Nature and Significance of Economic Science, London: Macmillan. Rosenberg, A. (1976) Microeconomic Laws: A Philosophical Analysis, Pittsburgh, Penn.: University of Pittsburgh Press. ——(1978) ‘The Puzzle of Economic Modelling’, The Journal of Philosophy 75(11): 679–83. ——(1983) ‘If Economics Isn’t Science, What Is It?’, The Philosophical Forum 14(3– 4): 296–314 (reprinted in Hausman, D.M. (ed.) The Philosophy of Economics: an Anthology (second edn): 376–94, Cambridge: Cambridge University Press (page references are to the latter). ——(1992) Economics—Mathematical Politics or Science of Diminishing Returns? Chicago: University of Chicago Press.
© 1997 Tony Lawson
——(1994) ‘What is the Cognitive Status of Economic Theory?’, in Backhouse, R. E. (ed.) New Directions in Economic Methodology, London: Routledge. Rotheim, R.J. (1988) ‘Keynes and the Language of Probability and Uncertainty’, Journal of Post Keynesian Economics 11(1): 82–99. ——(1989) ‘Organicism and the Role of the Individual in Keynes’ Thought’, Journal of Post-Keynesian Economics 12(2): 316–26. ——(1993) ‘On the Indeterminacy of Keynes’s Monetary Theory of Value’, Review of Political Economy 5(2): 197, 216. Rubery, J. (1978) ‘Structured Labour Markets, Worker Organisation and Low Pay’, Cambridge Journal of Economics. 2(1): 17–36 Rubinstein, A. (1991) ‘Comments on the Interpretation of Game Theory’, Econometrica 59(4): 909–24. ——(1995) ‘John Nash: the Master of Economic Modelling’, Scandinavian Journal of Economics 97(1): 9–13. Runde, J. (1990) ‘Keynesian Uncertainty and the Weight of Arguments’, Economics and Philosophy 6: 275–92. ——(1993) ‘Paul Davidson and the Austrians. Reply to Paul Davidson’, Critical Review, 7(2–3): 381–97. ——(1995a) ‘Risk, Uncertainty and Bayesian Decision Theory: A Keynesian View’: 197–210, in Dow, S. and Hillard, J. (eds) Keynes, Knowledge and Uncertainty, Aldershot: Edward Elgar. ——(1995b) ‘Review of Young Bak Choi’s Paradigms and Conventions’, Economics and Philosophy 11(2): 379–85. ——(1995c) ‘Chances and Choices: Some Notes on Probability and Belief in Economic Analysis’, The Monist July (78): 331, 352. ——(1995d) ‘Assessing Causal Explanation’, Cambridge: mimeo. ——(1997) ‘Keynesian Methodology’, in Harcourt, G.C. and Riach, P. (eds) Keynes General Theory: second edition, London: Routledge. Samuels, W.J. (1989) ‘Austrian and Institutional Economics: Some Common Elements’, Research in the History of Economic Thought and Methodology Vol. 6: 53, 72. Sargent, T.J. (1986) Rational Expectations and Inflation, New York: Harper & Row. Sargent, T.J. and Wallace, N. (1976) ‘Rational Expectations and the Theory of Economic Policy’, Journal of Monetary Economics 2. Sayer, A. (1984) Method in Social Science: A Realist Approach, London: Hutchison. ——(1985) ‘Realism and Geography’: 159–73 in Johnston, R.J. (ed.) The Future of Geography, London: Methuen. Schotter, A. (1981) The Economic Theory of Social Institutions, Cambridge: Cambridge University Press. ——(1983) ‘Why Take a Game Theoretical Approach to Economics? Institutions, Economics and Game Theory’, Economie Appliqué 36: 673–95. Seiz, J. (1993) ‘Feminism and the History of Economic Thought’, History of Political Economy 25(1): 185–201. Sellars, R.W. (1916) Critical Realism: a Study of the Nature and Conditions of Knowledge, New York and Chicago: Rand McNally and Company. Sellier, F. (1978) ‘France’, in Dunlop, J.T. and Galenson, W. (eds) Labour in the Twentieth Century, New York: Academic Press. Shackle, G.L.S. (1972) Epistemics and Economics: A Critique of Economic Doctrines, Cambridge: Cambridge University Press. ——(1989) What Did the ‘General Theory’ Do? 48–58
© 1997 Tony Lawson
in Pheby, J. (ed.) New Directions in Post-Keynsian Economics, Aldershot: Edward Elgar. Sims, C.A. (1980) ‘Macro-economics and Reality’, Econometrica 48 (1): 1–48. ——(1982) ‘Policy Analysis with Econometric Models’, Brookings Papers on Economic Activity (1): 107, 152. Sofianou, E. (1995) ‘Post-modernism and the Notion of Rationality in Economics’, Cambridge Journal of Economics 19(3): 373–90. Solow, R.E. (1986) ‘Economics: Is Something Missing?’, in Parker, W.N. (ed.) Economic History and the Modern Economist, Oxford: Basil Blackwell. Sonnenschein, H. (1972) ‘Market Excess Demand Functions’, Econometrica 40(3). ——(1973a) ‘Do Walras Identity and Continuity Characterize the Class of Community Excess Demand Functions?’, Journal of Economic Theory 6(4). ——(1973b) ‘The Utility Hypothesis and Market Demand Theory’, Western Economic Journal 11(4). Stafford, B. (1983) ‘The Class Struggle, the Multiplier and the Alterative Economic Strategy’, in Sawyer, M and Schott, K. (eds) Socialist Economic Review, London: Merlin Press. Stoker, T.M. (1993) ‘Empirical Approaches to the Problem of Aggregation over Individuals’, Journal of Economic Literature XXXI: 1827–74. Strassman, D.L. (1993a) ‘Not a Free Market: The Rhetoric of Disciplinary Authority in Economics’: 54–68 in Ferber, M.A. and Nelson, J. (eds) Beyond Economic Man: Feminist Theory and Economics, Chicago: Chicago University Press. ——(1993b) ‘The Stories of Economics and the Power of the Storyteller’, History of Political Economy 25(1): 147–65. ——(1994) ‘Feminist Thought and Economics; Or, What do the Visigoths Know?’, American Economic Review, Papers and Proceedings 84(2): 153–8. Strober, M.H. (1994) ‘Rethinking Economics through Feminist Lens’, American Economic Review, Papers and Proceedings 84(2): 143–7. Sugden, R. (1986) The Economics of Rights, Co-operation and Welfare, Oxford: Blackwell. Sullivan, H.S. (1955) The Interpersonal Theory of Psychiatry, London: Tavistock. Thompson, J.B. (1989) ‘The Theory of Structuration’, in Held, D. and Thompson, J.B. (eds) Social Theory of Modern Societies: Anthony Giddens and his Critics, Cambridge: Cambridge University Press. Vanberg, V. (1994) Rules and Choice in Economics, London: Routledge. Veblen, T.B. (1909) ‘The Limitations of Marginal Utility’, Journal of Political Economy XVII(9) (reprinted in The Place of Science in Modern Civilisation (1919), New York: Viking Press; the latter reprinted with new introduction by Samuels, W. (1990), New Brunswick: Transaction Publishers). ——(1919) The Place of Science in Modern Civilisation and Other Essays, New York: Huebsch (reprinted with a new introduction by Samuels, W.J., New Brunswick: Transaction (1990)). Von Neumann, J. and Morgenstern, O. (1944) Theory of Games and Economic Behaviour, Princeton, NJ: Princeton University Press. Vromen, J.J. (1995) Economic Evolution: an Enquiry into the Foundations of New Institutional Economics, London: Routledge Waddington, C.H. (1957) The Strategy of the Genes, London: George Allen & Unwin. ——(ed.) (1972) Towards a Theoretical Biology 4 vols, Edinburgh: Edinburgh University Press. © 1997 Tony Lawson
Webb, B. and Webb, S. (1920) The History of Trade Unionism, (revised edn), London: Longmans. Wedderburn, K.W. (1972) ‘Labour Law and Labour Relations in Britain’, British Journal of Industrial Relations X(2): 270–90. ——(1985) ‘The New Policies in Industrial Relations Law’, in Fosh, P. and Littler, C.R. (eds) Industrial Relations and the Law in the 1980s, Aldershot: Gower. Weintraub, E.R. (1989) ‘Methodology Doesn’t Matter, But the History of Thought Might’, Scandinavian Journal of Economics 91(2): 477–93. White, P.J. (1978) ‘Is Voluntarism in Decline?’, Industrial Relations Journal 9(3): 34– 43. Whitehead, A.N. (1926) Science and the Modern World, Cambridge: Cambridge University Press. Wiles, P. and Routh, E. (1984) Economics in Disarray, Oxford: Basil Blackwell. Wilkinson, F.W. (ed.) (1981) The Dynamics of Labour Market Segmentation, London: Academic Press. Williamson, O.E. (1975) Markets and Hierarchies, New York: Free Press. Winslow, E.G. (1986) ‘Keynes and Freud: Psychoanalysis and Keynes’s Account of the “Animal Spirits” of Capitalism’, Social Research 53(4): 549–78. Witten, E. (1988) Chapter 3, in Davies, P.C.W. and Brown, J. (eds) Superstrings: a Theory of Everything, Cambridge: Cambridge University Press. Youngson, A.J. (1960) The British Economy 1920–57, London: George Allen & Unwin.
© 1997 Tony Lawson