Annotations to Bhaskar’s Realist Theory of Science

Hans G. Ehrbar

May 12, 2005 Contents

Prefaces i 0.1 Bibliographical Remark ...... i 0.2 Preface to the First Edition ...... ii 0.3 Preface to the 2nd edition ...... xiii 0.4 Footnotes to the Preface ...... xiii

Introduction xv

2 CONTENTS 3

1 Philosophy and Scientific Realism 1 1.1 Two Sides of ‘Knowledge’ ...... 1 1.2 Three Traditions in the Philosophy of Science ...... 8 1.3 The Transcendental Analysis of Experiences ...... 19 1.3.A The Analysis of Perception ...... 21 1.3.B Analysis of Experimental Activity ...... 25 1.4 Status of Ontology and its Dissolution in Classical Philosophy ..... 31 1.5 Ontology Vindicated and the Real Basis of Causal Laws ...... 47 1.6 A sketch of a critique of empirical realism ...... 68 1.7 Footnotes to Chapter 1 ...... 76

2 Actualism and the Concept of a Closure 82 2.1 Introduction: On the Actuality of the Causal Connection ...... 82 2.2 Regularity Determinism and the Quest for a Closure ...... 90 2.3 The Classical Paradigm of Action ...... 103 2.4 Actualism and Transcendental Realism: the Interpretation of Normic Statements ...... 120 2.5 Autonomy and Reduction ...... 140 2.6 Explanation in Open Systems ...... 158 4 CONTENTS

2.7 Footnotes to Chapter 2 ...... 169 2.8 Appendix: Orthodox Philosophy of Science and the Implications of Open Systems ...... 173

3 The Logic of Scientific Discovery 196 3.1 Introduction: On the Contingency of the Causal Connection ...... 196 3.2 The Surplus-element in the Analysis of Law-like Statements: A Cri- tique of the Theory of Models ...... 204 3.3 Natural Necessity and Natural Kinds: The Stratification of Nature and the Stratification of Science ...... 222 3.4 The Social Production of Knowledge by Means of Knowledge ..... 249 3.5 Objections to the Account of Natural Necessity Proposed ...... 267 3.6 The Problem of Induction ...... 290 3.7 Footnotes to Chapter 3 ...... 308 3.8 Appendix: Natural Tendencies and Causal Powers ...... 316 3.9 Footnotes to Appendix of Chapter 3 ...... 330

4 Metaphysics and the Philosophy of Science 331 4.1 Footnotes to Chapter 4 ...... 348 CONTENTS 5

5 Postscript to the Second Edition 349 5.1 Footnotes for the Postscript ...... 361

Bhaskar’s Bibliography 367 6 CONTENTS Prefaces

0.1 Bibliographical Re- low follows these editions. [Bha97] is a mark reprint of the second edition. The collec- tion [ABC+98] has reprints of chapters The first edition [Bha75] and second edi- 1 and 2, page breaks of the form <16> tion [Bha78] of Bhaskar’s Realist The- come from this collection. The present ory of Science are identical until p. 250 pdf edition is fairly accurate concerning (except for a short remark at the end the text, but many diagrams are not yet of the preface). The second edition has transcribed. a postscript and an updated bibliogra- There is a mailing list about criti- phy. The paragraph numbering given be- cal realism and Bhaskar’s work, with

i ii PREFACES archives at http://archives.econ.utah. the proverbial frog at the bottom of the beer edu/archives/bhaskar. Between July 2, mug, they have always reappeared. 1996 and March 1998 this mailing list has There was a phase in recent philosophy been discussing the Realist Theory of Sci- when it was widely held that the problem ence. was the problems and not their solution. In practice, however, this interesting idea was usually coupled with the belief that termi- 0.2 Preface to the First nation of philosophical reflection of the tra- ditional kind would be in itself sufficient to Edition resolve the problems to which, it was held, philosophical reflection had given rise. The preface begins be mentioning a num- The next paragraph gives a number of ber of “problems of philosophy” that have examples where problems of philosophy eluded solution this far. A little later, in crop up in the substantive sciences. 10:1, Bhaskar says that this book will not 6:2 Whatever the merits of such a view in solve these problems of philosophy, but it general, it is quite untenable for any philoso- will prepare the ground for their solution. pher who is concerned with science. For 6:1 It has often been claimed, and per- in one science after another recent develop- haps more often felt, that the problems of ments, or in some cases the lack of them, philosophy have been solved. And yet, like have forced old philosophical problems to the 0.2. PREFACE TO THE FIRST EDITION iii fore. Thus the dispute between Parmenides by lacking tools and an inadequate con- and Heraclitus as to whether being or be- ceptual equipment. coming is ultimate lies not far from the cen- 6:3/o In this context one might have ex- tre of methodological controversy in physics; pected a ferment of creative activity within while the dispute between rationalists and the philosophy of science, and to a degree empiricists over the respective roles of the a this has occurred. But the latter’s capac- priori and the empirical continues to domi- ity for autonomous growth is limited. For nate methodological discussion in economics. the critical or analytical philosopher of sci- Sociologists are making increasing use of the ence can only say as much as the philosoph- allegedly discredited Aristotelian typology of ical tools at his disposal enable him to say. causes. And the problem of universals has And if philosophy lags behind the needs of re-emerged in an almost Platonic form in the moment then he is left in the position of structural linguistics, anthropology and de- a Priestley forced, by the inadequacy of his velopmental biology. The spectre of deter- conceptual equipment, to think of oxygen as minism continues to haunt many of the sci- ‘dephlogisticated air’;1 or, of a Winch baffled ences; and the problem of ‘free-will’ is still a by an alien sociology.2 problem for psychology. Question 1 In the Preface of A Real- ⇓ These problems should have given ist Theory of Science, Bhaskar writes that rise to a ferment of activity in philoso- philosophy has not kept up with modern phy, but philosophy has been held back sciences. Modern sciences are plagued iv PREFACES by a number of ancient unresolved “pi- focus has been almost exclusively on how losophical problems.” Philosophy has not we know, marginalizing consideration of been able to solve these problems because what has to be the case ontologically for its framework is too limited. Bhaskar us to know at all. proposes in his book to expand this frame- The very word “objective” is generally work and thus enable philosophy to better assumed to be an outdated, untenable, understand science and also be helpful to epistemological category – as in “there it. What does the great advance in phi- is no objective or value-neutral” point of losophy consist of, which RB alludes to view. But “objective” is also an ontolog- here? ical category: While all our ideas about the world are our own creation (episte- Here is Doug Porpora’s answer to mologically), we still make an ontologi- this question on the Bhaskar list, on cal distinction among our ideas between http://archives.econ.utah.edu/archives/bhaskar/1996-those that refer to what is dependent on 07-26.024/msg00052.htm: our consciousness and those that refer to I think Bhaskar makes the crucial what is independent of our consciousness. point that contemporary philosophical Today, these two senses of “objective” are or meta-theoretical discussion has “priv- conflated at the expense of ontology. ileged” epistemology over ontology. The Bhaskar intends to recover the dis- 0.2. PREFACE TO THE FIRST EDITION v tinction. And he is going to do so too late.3 Yet there can be little doubt that by focusing on how we conceptualize our theory of knowledge has scarcely come to that ontologically objective quality we terms with, let alone resolved the crises in- call causality. Positivism conceptualized duced by, the changes that have taken place causality as event-regularities, a concep- across the whole spectrum of scientific (and tion that presupposes a certain ontology. one might add social and political) thought. In this respect our present age contrasts un- By examining that ontology, we can de- favourably with both Ancient Greece and termine whether the conception is ten- Post-Renaissance Europe, where there was able. Already, Bhaskar is suggesting that a close and mutually beneficial relationship this conception will not be tenable and between science and philosophy. It is true that a post-positivist account of causal- that in the second of these periods there was ity is needed, which presupposes a dif- a progressive ‘problem-shift’ within philos- ferent, presumably more defensible ontol- ophy from the question of the content of ogy. That ontology will presumably also knowledge to the meta-question of its sta- 4 make better sense of what science is ac- tus as such. This shift was in part a re- tually doing. sponse to the consolidation of the Newtonian world-view, until by Kant’s time its funda- 7:1 mental axioms could be regarded as a priori Hegel may have exaggerated when he said conditions of the possibility of any empiri- that philosophy always arrives on the scene vi PREFACES cal knowledge. However, those philosophers a Beethoven symphony or the perfume of a of the present who insist upon their total au- rose, as a mere effect of those primary pow- tonomy from the natural and human sciences ers. Suppose further that philosophers had not only impoverish, but delude themselves. taken biology or economics as their paradigm For they thereby condemn themselves to liv- of a science rather than physics; or 16th not ing in the shadow cast by the great scientific 17th century physics as their paradigm of sci- thought of the past. entific activity. Would not our philosophical inheritance have been vastly different? As 7:2/o Anyone who doubts that scientific this is primarily a problem for the philoso- theories constitute a significant ingredient in phy of philosophy rather than the philosophy philosophical thought should consider what of science, I shall not dwell on this point fur- the course of intellectual history might have ther here. Its significance for our story will been if gestalt psychology had been estab- emerge in due course. lished in place of Hartley’s principle of the After some preliminaries, the following association of ideas; or if the phenomena of paragraphs give an interesting overview: electricity and magnetism had come to be regarded as more basic than those of im- ⇓ Primary aim of this book: devel- pact and gravity; or if sounds and smells opment of systematic realist account of had been taken as constitutive of the basic science, which is an alternative to pos- stuff of reality and the rich tapestry of the itivism. Other positivism critiques are visual-tactile world had been regarded, like vulnerable to counterattack because they 0.2. PREFACE TO THE FIRST EDITION vii do not go all the way as Bhaskar does. then showing that these conditions (sig- 8:1 The primary aim of this study is the nificance of experience) are not valid gen- development of a systematic realist account erally but a special case (experience is of- of science. In this way I hope to pro- ten insignificant). vide a comprehensive alternative to the pos- 8:2 itivism that has usurped the title of sci- ence. I think that only the position devel- My subsidiary aim is thus to show once- oped here can do full justice to the rational- and-for-all why no return to positivism is ity of scientific practice or sustain the intelli- possible. This of course depends upon my gibility of such scientific activities as theory- primary aim. For any adequate answer to construction and experimentation. And that the critical meta-question ‘what are the con- while recent developments in the philosophy ditions of the plausibility of an account of sci- of science mark a great advance on posi- ence?’ presupposes an account which is ca- tivism they must eventually prove vulnerable pable of thinking of those conditions as spe- to positivist counter-attack, unless carried to cial cases. That is to say, to adapt an image the limit worked out here. of Wittgenstein’s, one can only see the fly in the fly-bottle if one’s perspective is different ⇓ Subsidiary aim is to refute posi- from that of the fly.5 And the sting is only tivism once-and-for-all. Bhaskar does removed from a system of thought when the this by describing the conditions un- particular conditions under which it makes der which positivism seems plausible and sense are described. In practice this task is viii PREFACES simplified for us by the fact that the condi- of vigorous critical activity in the philoso- tions under which positivism is plausible as phy of science. In the course of this the an account of science are largely co-extensive twin templates of the positivist view of sci- with the conditions under which experience ence, viz. the ideas that science has a certain is significant in science. This is of course an base and a deductive structure, have been important and substantive question which subjected to damaging attack. With a de- we could say, echoing Kant, no account of gree of arbitrariness one can separate this science can decline, but positivism cannot critical activity into two strands. The first, ask, because (it will be seen) the idea of in- represented by writers such as Kuhn, Pop- significant experiences transcends the very per, Lakatos, Feyerabend, Toulmin, Polanyi bounds of its thought.6 and Ravetz, emphasises the social charac- The next paragraph positions this ter of science and focusses particularly on study in two strands of contemporary the phenomena of scientific change and de- criticism of the positivist view that “sci- velopment. It is generally critical of any monistic interpretation of scientific develop- ence has a certain base and deductive ment, of the kind characteristic of empiri- structure”: one calling attention to the cist historiography and implicit in any doc- social character and the other to the trine of the foundations of knowledge. The stratification of science. This study is a second strand, represented by the work of synthesis of both. Scriven, Hanson, Hesse and Harre among 8:3/o This book is written in the context 0.2. PREFACE TO THE FIRST EDITION ix others, calls attention to the stratification of not constant conjunctions of events. science. It stresses the difference between ex- planation and prediction and emphasises the 9:1/o To see science as a social activity, role played by models in scientific thought. and as structured and discriminating in its It is highly critical of the deductivist view of thought, constitutes a significant step in our the structure of scientific theories, and more understanding of science. But, I shall argue, generally of any exclusively formal account without the support of a revised ontology, of science. This study attempts to synthe- and in particular a conception of the world as sise these two critical strands; and to show in stratified and differentiated too, it is impos- particular why and how the realism presup- sible to steer clear of the Scylla of holding the posed by the first strand must be extended structure dispensable in the long run (back to cover the objects of scientific thought pos- to empiricism) without being pulled into the tulated by the second strand. In this way Charybdis of justifying it exlusively in terms I will be describing the nature and the de- of the fixed or changing needs of the scientific velopment of what has been hailed as the community (a form of neo-Kantian pragma- ‘Copernican Revolution’ in the philosophy of tism exemplified by e.g. Toulmin and Kuhn). science.7 In this study I attempt to show how such a revised ontology is in fact presupposed by ⇓ But the new element Bhaskar brings the social activity of science. The basic prin- in is: epistemology presupposes an ontol- ciple of realist philosophy of science, viz. that ogy. Causal laws are tendencies of things, perception gives us access to things and ex- x PREFACES perimental activity access to structures that tution instance of some more general theory exist independently of us, is very simple. Yet of knowledge. This is a situation which has the full working out of this principle implies a worked to the disadvantage of both philoso- radical account of the nature of causal laws, phy and knowledge. If, however, we reverse viz. as expressing tendencies of things, not the customary procedure and substitute the conjunctions of events. And it implies that more specific ‘science’ (or even better ‘sci- a constant conjunction of events is no more ences’) for ‘knowledge’, considerable illumi- a necessary than a sufficient condition for a nation of many traditional epistemological causal law. problems can, I think, be achieved. And ⇓ This study focuses on science be- some even, in so far as the ‘knowledge’ we cause one needs a philosophy of science are concerned with is that produced by ‘sci- ence’, become susceptible of definitive solu- first, before one can have a theory of tion. The result of this reversal will also be knowledge. This also makes philosophy a philosophy which has a greater relevance relevant for science, its underlaborer. than is the case at present for scientific prac- 10:1 I do not claim in this book to solve tice. In this sense my objective could be said any general problems of philosophy. It is my to be a ‘philosophy for science’. intention merely to give an adequate account My comment about this passage on the of science. Philosophers, including philoso- bhaskar mailing list: phers of science, have for too long regarded the philosophy of science as a simple substi- Bhaskar does not say here why the re- 0.2. PREFACE TO THE FIRST EDITION xi versal will yield “considerable illumina- also shape its intra-social communica- tion.” This is a preface, and many things tions, but which it would not be able to are only hinted at. Perhaps we should, acquire by such communications alone. in this collective reading, try to fill in Have I made myself clear? To me, the some of the gaps, i.e., inform each other primacy of a theory of science within a of what we are reading into the text. theory of knowledge is closely related to Bhaskar claims that one should do the- the primacy of the relations of produc- ory of science before a more general the- tion within the broader circle of all social ory of knowledge. Why should one? My relations. take on it is the following, and I am not Doug Porpora replied as follows: sure how close this is to what Bhaskar Hans, my reaction to the passage you would argue: Science is a process by cite was different. I thought Bhaskar which society acquires knowledge about was saying that scientific knowledge has the world. It is “between” society and been subsumed under the larger question the outside world. Much of nonscientific of knowledge per se and that progress knowledge is communication “within” so- on that larger question has stagnated ciety. In order to gain knowledge about around such formulations as “knowledge the world, society has to overcome ob- as justified true belief.” stacles and acquire tools which will then So what Bhaskar is proposing is to be- xii PREFACES gin fresh by examining only what knowl- late the philosophy of science to the more edge is for science. That is also a very general historical concerns of philosophy; strategic point of entry because scien- and at the same time to indicate more pre- tific knowledge is a paradigmatic form of cisely than is usual the consequences for sci- knowledge. Perhaps, all would agree that entific practice of the methodological strate- science is accomplishing at least some- gies implied by different philosophies of sci- ence. thing we might call knowledge. Thus, what we learn about knowledge from sci- ⇓ This preface concludes with the re- ence may help us know about knowledge mark that science is frail, but human ex- more generally. That is how I understood istence is generally frail, and science may Bhaskar’s philosophical “reversal.” still be our best hope. ⇓ Now some important remarks re- 10:2/o We are too apt to forget the frailty of both our science and our philosophy. garding “underlaboring”: There can be no certainty that they will sur- For I willingly confess to Lockean motives. vive and flourish; or, if they do, that they That is to say, I believe it to be an essential will benefit mankind. Civilisation is, like (though not the only) part of the business of man himself, perhaps nothing more than a philosophy to act as the under-labourer, and temporary rupture in the normal order of occasionally as the mid-wife, of science.8 I things.9 It is thus also part of the job of have therefore tried in this study both to re- the philosopher to show the limits of science. 0.4. FOOTNOTES TO THE PREFACE xiii

And, in this broader sense, to seek to ensure ically comment on the book. The index fills a that the Owl of Minerva takes flight before major lacuna in the first edition of the work. the final falling of the dusk. Francis Roberts and Robin Kinross helped 11:1 I would like to take this opportunity me to compile it. to express my thanks to Alan Montefiore and 11:4 ROY BHASKAR University of Edin- Rom Harre for reading earlier versions of burgh September 1977 this work; to Rom Harre and Hilary Wain- wright for their continual encouragement; to many other colleagues and friends for their 0.4 Footnotes to the help; and to Mrs E. Browne for typing the manuscript. Preface 11:2 ROY BHASKAR University of Edin- burgh April 1974 1 See e.g. S. E. Toulmin, ‘Crucial Experiments: Priestley and Lavoisier’, The Journal of the His- tory of Ideas, Vol. XVIII (1957), pp. 205–20; and J. B. Conant, The Overthrow of the Phlogiston 0.3 Preface to the 2nd Theory. 2 P. Winch, The Idea of a Social Science, p. edition 114. 3 G. W. F. Hegel, Philosophy of Right, Pref- 11:3 This edition includes a postscript and ace. an index. The postscript enables me to crit- 4 Cf. G. Buchdahl, Metaphysics and the Phi- xiv PREFACES losophy of Science, p. 2 5 L. Wittgenstein, Philosophical Investiga- tions, 309. 6 I. Kant, Critique of Pure Reason, Preface to the 1st Edition. 7 R. Harre, Principles of Scientific Thinking, p. 15. 8 J. Locke, Essay Concerning Human Under- standing, Epistle to the Reader. 9 Cf. M. Foucault, The Order of Things, p. XXIII. Introduction

In order to get away from get away even necessary. from the positivistic view of science, a 12:1 The aim of this book is the devel- new understanding of causal laws is nec- opment of a systematic realist account of essary. Hume’s theory identifying causal science. Such an account must provide a laws with constant conjunctions of events comprehensive alternative to the positivism is still widely accepted. Modern cri- which since the time of Hume has fashioned tiques of Hume say that constant con- our image of science. Central to the posi- junctions are not sufficient for causal laws tivist vision of science is the Humean theory (they distinguish correlation and causa- of causal laws. It is a principal concern of this study to develop some new arguments tion). Bhaskar will argue they are not and show how they relate to more familiar

xv xvi INTRODUCTION ones against this still widely accepted the- Question 3 Bhaskar says: it can be ory. In particular I want to argue that not shown by a transcendental argument from only is a constant conjunction of events not a the nature of experimental activity that a sufficient, it is not even a necessary condition constant conjunction of events is not a for a scientific law; and that it is only if we necessary condition for a scientific law. can establish the latter that we can provide Give this argument. an adequate rationale for the former. It has often been contended that a constant con- 12:2 It is a condition of the intelligibility junction of events is insufficient but it has of experimental activity that in an experi- not so far been systematically argued that ment the experimenter is a causal agent of a it is not necessary. This can, however, be sequence of events but not of the causal law shown by a transcendental argument from which the sequence of events enables him to the nature of experimental activity. identify. This suggests that there is a on- tological distinction between scientific laws Question 2 Bhaskar says that constant and patterns of events. conjunctions of events are neither neces- ⇑ Since it takes so much work on the sary nor sufficient for causal laws. Give part of the experimenter to translate a both halves of this argument. Which half law into a constant conjunction of events, is Bhaskar’s innovation? laws and constant conjunctions cannot be one and the same. xvii

Obviously this creates a prima facie problem is only if it is backed by a theory, containing for any theory of science. a model or conception of a putative causal or This prima face problem is: which real- explanatory ‘link’, that a law can be distin- ity does a law have if it is not a constant guished from a purely accidental concommi- conjunction of events? tance. The possibility of saying this clearly depends upon a non-reductionist conception I think that it can be solved along the fol- of theory. Now at the core of theory is a con- lowing lines: ception or picture of a natural mechanism or Look at a law. Implicit in a law is structure at work. Under certain conditions always a theory. A theory is usually some postulated mechanisms can come to be the description of some natural mecha- established as real. And it is in the working nism at work. Bhaskar claims now that of such mechanisms that the objective basis some of these mechanisms are not sim- of our ascriptions of natural necessity lies. ply as-if scenarios allowing one to orga- ⇓ Here I propose to look at the sec- nize one’s thoughts but that they corre- ond and third sentence first, and then spond to real natural mechanisms going at the first sentence. Positivism, with on in the world. Whenever we talk about all its emphasis on the experiment, can- a natural necessity we refer to the work- not answer the following question: if it ing of such a mechanism. takes so much effort by the experimenter To ascribe a law one needs a theory. For it to elicit the constant conjunctions, why xviii INTRODUCTION should his findings have any relevance for But it is only if we are justified in assum- someone outside such controlled experi- ing this that the idea of the universality of a mental situations? To answer this, you known law can be sustained or that experi- need indeed the independence of the law mental activity can be rendered intelligible. from constant conjunctions: experiments Hence one of the chief objections to posi- are only relevant because the laws, which tivism is that it cannot show why or the con- ditions under which experience is significant the experimenter has to spend so much in science. Most critics have emphasized its care to expose, remain active outside the depreciation of the role of theory; this ar- experimental setup. Outside the experi- gument shows its inadequacy to experience. mental situation, they do not give rise to Moreover it is only because it must be as- constant conjunctions, but they are still sumed, if experimental activity is to be ren- active. dered intelligible, that natural mechanisms endure and act outside the conditions that 13:1 It is only if we make the assump- enable us to identify them that the applica- tion of the real independence of such mech- bility of known laws in open systems, i.e. in anisms from the events they generate that systems where no constant conjunctions of we are justified in assuming that they en- events prevail, can be sustained. This has dure and go on acting in their normal way the corollary that a constant conjunction of outside the experimentally closed conditions events cannot be necessary for the assump- that enable us to empirically identify them. xix tion of the efficacy of a law. ful also outside the experiment. This is The first two sentences in the next Bhaskar’s distinction of the real, i.e., the paragraph go over the same argument underlying mechanisms, and the actual, again: i.e., the events that actually happen. 13:2 This argument shows that real struc- ⇓ Then Bhaskar does some handwav- tures exist independently of and are often ing to derive the difference between the out of phase with the actual patterns of actual and the empirical: he says that a events. Indeed it is only because of the lat- similar distinction must be made between ter that we need to perform experiments and the actual, i.e., that what happens, and only because of the former that we can make the empirical, i.e., that what is perceived sense of our performances of them. (this is why scientists need training). ⇑ Real structures (or, as Bhaskar also calls them, generative mechanisms) are Similarly it can be shown to be a condition of the intelligibility of perception that events often out of phase with the actual pat- occur independently of experiences. And ex- terns of events: this is why experimen- periences are often (epistemically speaking) tal work is needed to expose them. And ‘out of phase’ with events - e.g. when they these generative mechanisms exist in- are misidentified. It is partly because of this dependently of the actual patterns of possibility that the scientist needs a scientific events: this is why their knowledge is use- education or training. xx INTRODUCTION

This is how he derives that the domains factual statements about the world. of the real, actual, and empirical are dis- tinct. 14:1 The real basis of causal laws are pro- Thus I will argue that what I will call the vided by the generative mechanisms of na- domains of the real, the actual and the em- ture. Such generative mechanisms are, it is pirical are distinct. This is represented in argued, nothing other than the ways of act- Table 0.1 below: – ing of things. And causal laws must be an- 13:3 alyzed as their tendencies. Tendencies may be regarded as powers or liabilities of a thing which may be exercised without being man- Question 4 Re-phrase in your own ifest in any particular outcome. The kind of words Bhaskar’s derivation of the separa- conditional we are concerned with here may tion of the world into the three domains: be characterised as normic. They are not the empirical, the actual, and the real. counter-factual but transfactual statements. Nomic universals, properly understood, are transfactual or normic statements with fac- ⇓ Now Bhaskar goes one step further in tual instances in the laboratory (and perhaps his theory of what laws really are: they a few other effectively closed contexts) that are not just mechanisms hanging in the constitute their empirical grounds; they need air, but tendencies inherent in the way not, and in general will not, be reflected in things act. They lead to normic or trans- an invariant pattern or regularly recurring xxi

Table 0.1 Domain of Real Actual Empirical Mechanisms X Events X X Experiences X X X

sequence of events. the laws for practical activity that takes place in open systems? And how can ⇓ Back to a critique of Hume. It seems one explain what the experimenter does to be a summary. A closed system is a with all the necessary preparations be- system which has no input from the out- fore the constant conjunction can actu- side, so that constant conjunctions are ally be observed? Note the parentheti- possible in it. Now I will do Bhaskar’s cal remark Bhaskar makes in paragraph second point first because it seems eas- 15:2 that experimental activity presup- ier to understand for me. Weakness of poses open systems! To resolve this, what Hume is: if laws are valid only in closed one has to do is to interpret laws in a systems, then why is it relevant to know xxii INTRODUCTION transfactual way. But once one goes be- open systems can be sustained. Once we al- hind the facts, this is where one has to low for open systems then laws can only be locate a natural necessity in the world, universal if they are interpreted in a non- not just in our brains. Perhaps one can empirical (trans-factual) way, i.e. as desig- say: as soon as one no longer identifies nating the activity of generative mechanisms empirical regularities with laws, but says and structures independently of any partic- ular sequence or pattern of events. But once that empirical regularities are the man- we do this there is an ontological basis for ifestation of some transfactual tendency, a concept of natural necessity, that is neces- then there must be something that has sity in nature quite independent of men or this tendency. Such a tendency cannot human activity. just hang in empty space, it must be the ⇓ After this summary he continues the way of acting of something (see preceding argument, goes with diagram 0.1, which paragraph 14:1). is not yet typed in here, but a copy of 14:2 The weakness of the Humean concept this diagram is also used in chapter 2 and of laws is that it ties laws to closed systems, therefore can be found in the Essential viz. systems where a constant conjunction of Readings p. 50. events occurs. This has the consequence that 14:3/o In science there is a kind of di- neither the experimental establishment nor alectic in which a regularity is identified, a the practical application of our knowledge in plausible explanation for it is invented, and xxiii the reality of the entities and processes pos- Question 5 Why is it necessary to tulated in the explanation is then checked. check the reality of the underlying mech- This dialectic is illustrated in Diagram 0.1 anism which a theory postulates as expla- below. nation of certain phenomena? Is it not 15:1 Diagram 0.1. The Logic of Scientific enough to verify that the postulated un- Discovery derlying mechanism leads to the phenom- Three steps, Bhaskar calls it dialectic ena to be explained? because it leads from empirics to some real entities which become the empirics in the next step. ⇓ In the remainder of this paragraph 15:2 If a classical empiricist tradition in Bhaskar rounds off and concludes his ar- the philosophy of science stops at the first gument. Most glaring weakness of or- stage, a rival neo-Kantian or transcendental thodox philosophy of science is: if one idealist tradition (discernible in the history takes Hume’s definition of causal laws se- of the philosophy of science) stops at the sec- riously, one must conclude that there are ond. If and only if the third step is taken and no causal laws in the sciences. There developed in the way indicated above can are no univeral empirical regularities, but there be an adequate rationale for the use of philosophers postulate those and there- laws to explain phenomena in open systems, fore get into troubles whenever they try where no constant conjunctions prevail. to confirm a theory or define what pro- xxiv INTRODUCTION cedures are scientifically rational. Or- most glaring weakness of orthodox philoso- thodox philosophy thinks the world is a phy of science: viz. the nonexistence in sci- closed system (which means that every- ence of Humean causal laws, i.e. of universal thing has been predetermined since the empirical generalizations, and hence the in- Big Bang), and they ignore the fact which adequacy of the criteria of explanation, con- everybody knows that usually, constant firmation (or falsification), scientific ratio- nality etc., that are based on the assump- conjunctions of events will only come tion that a closure is the universal rule rather about as the result of experimental activ- than the rare and (for the most part) artifi- ity. The experimenter’s activity, like hu- cially generated exception that I contend it man activity in general, is a causal agent is. It is because our activity is (normally) a which changes the course of the world necessary condition of constant conjunctions (but this kind of activity is so relevant of events that the philosophy of science needs exactly because it leads to a display of an ontology of structures and transfactually a natural law which is not of the experi- active things. menter’s making!) ⇓ The next paragraph is still about di- agram 0.1: How transcendental realism It is the unthinking presupposition of closed systems together with the failure to anal- differs from empiricism and transcenden- yse experimental activity (which presup- tal idealism. poses open systems) that accounts for the 15:3 The position advanced here is charac- xxv terized as transcendental realism, in opposi- ⇑ Here the slashed expressions re- tion to the empirical realism common to the sult/regularity and imagined/imaginary other two traditions. Both the neo-Kantian in the diagram are significant: empiri- or transcendental idealist tradition and tran- cism considers step (1) as a regularity, scendental realism see the step between (1) TR as a result of a causal law that may and (2) in Diagram 0.1 as involving creative not necessarily be regular. Transcenden- model building, in which plausible genera- tal idealism sees (2) as necessarily imagi- tive mechanisms are imagined to produce the phenomena in question. But transcendental nary, while TR says it is at first imagined realism sees the need for the step between but may come to be known as real. (2) and (3) also, in which the reality of the ⇓ The last sentence in this paragraph mechanisms postulated are subjected to em- introduces a theme that will discussed pirical scrutiny. Transcendental realism dif- further in the next paragraph: with- fers from empirical realism in interpreting out all this one cannot explain scientific (1) as the invariance of an (experimentally growth and change. I.e., the necessity produced) result rather than a regularity; of step (3) is therefore not only linked and from transcendental idealism in allow- ing the possibility that what is imagined in into the stylized facts about experimental (2) need not be imaginary but may be (and activity, but also into the stylized facts come to be known as) real. about the growth and change of science. Without such an interpretation it is impos- xxvi INTRODUCTION sible to sustain the rationality of scientific new sense-extending equipment. Once the growth and change. explanation is discovered science then moves on to the construction and testing of possi- Question 6 Bhaskar claims that scien- ble explanations for it. At each level of real- tific growth and change cannot be ex- ity law-like behaviour has to be interpreted plained without transcendental realism. normically, i.e. as involving the exercise of Why not? tendencies which may not be realised. ⇓ Now Bhaskar introduces the concept ⇓ This dialectic has no end. of epistemic fallacy. The dogma that 15:4/o A conception of science is argued statements about being can always be re- for in which it is seen as a process-in-motion, formulated so that they become state- with the dialectic mentioned above in prin- ciple having no foreseable end. Thus when ments about our knowledge about be- a new stratum or level of reality has been ing denies a second-order, transcendental discovered and adequately described science knowledge about the world that can be moves immediately to the construction and gained from observing scientists at work. testing of possible explanations for what 16:1 Empirical realism is underpinned by happens at that level. This will involve a metaphysical dogma, which I call the epis- drawing on whatever cognitive equipment is temic fallacy, that statements about being available and perhaps the design of new ex- can always be transposed into statements perimental techniques and the invention of about our knowledge of being. As ontology xxvii cannot, it is argued, be reduced to episte- the recognition of, an implicit empiricist mology this mistake merely covers the gen- ontology. eration of an implicit ontology based on the (These presumptions can, I think, only category of experience; and an implicit re- be explained in terms of the need felt alism based on the presumed characteristics by philosophers for certain foundations of of the objects of experience, viz. atomistic knowledge.) This in turn leads to the gener- events, and their relations, viz. constant con- ation of a methodology which is either con- junctions. sistent with epistemology but of no relevance to science; or relevant to science but more or ⇑ Let me try to explain it this way: less radically inconsistent with epistemology. One argument against the primacy of the So that, in short, philosophy itself tends to question “what must the world be like for be out of joint with science. science to be possible” is that everything ⇑ The sentence in parentheses is inter- we know goes through science and there- esting: philosophers feel the need to base fore what we say about the world is only knowledge on some certainty, therefore a subset of what we may say about sci- they postulate atomistic events related ence, epistemology crowds out ontology. by constant conjunctions. This false and Bhaskar maintains that one has to look unreflected ontology leads to misguided at the ontology separately, and that the methodology (but he does not elaborate epistemic fallacy only covers up, prevents here). xxviii INTRODUCTION

Question 7 What is the epistemic fal- passive recipients of given facts and recorders lacy? What are its implications for sci- of their given conjunctions. ence and for philosophy? Because, look in chapter 1, 27:3/o or around there. One reason is: experiences The next paragraph explains why is may not be significant, it is the result of it wrong to speak about the “empirical prior scientific activity to sort out which world.” A more detailed answer to this experiences are significant. (This is rele- question will be given in 27:3/o. vant for class consciousness, also explains 16:2 It is argued in Chapter 1 that the why sciences can progress in a completely very concept of the empirical world embodies different direction for a long time.) Here a category mistake, which depends upon a he speaks of epistemological individual- barely concealed anthropomorphism within ism, to be distinguished from method- philosophy; and leads to a neglect of the ological individualism, i.e., neglect of an- important question of the conditions under tecedent social activity. (Leads to an im- which experience is in fact significant in sci- ence. In general this depends upon an- plicit sociology.) tecedent social activity. Neglect of this activ- 16:3/o Against this it is argued that ity merely results in the generation of an im- knowledge is a social product, produced by plicit sociology, based on an epistemological means of antecedent social products; but individualism in which men are regarded as that the objects of which, in the social ac- xxix tivity of science, knowledge comes to be pro- structure or mechanism that exists and acts duced, exist and act quite independently of quite independently of men and the condi- men. tions which allow men access to it. These ⇑ The “of which” construction is per- dimensions are related in Chapter 3. Two haps difficult to understand. I think he criteria for the adequacy of an account of sci- means that the objects whose knowledge ence are developed: (i) its capacity to sustain comes to be produced in the social activ- the idea of knowledge as a produced means of production; and (ii) its capacity to sustain ity of science exist and act quite indepen- the idea of the independent existence and ac- dently of men. tivity of the objects of scientific thought. ⇓ Next he introduces the transitive di- mension. Two criteria for adequacy of an ⇓ Laws do not refer to the empirical or account of science. actual but the real: These two aspects of the philosophy of sci- 17:1 It is the overall argument of this ence justify our talking of two dimensions study then that knowledge must be viewed and two kinds of ‘object’ of knowledge: a as a produced means of production and transitive dimension, in which the object is science as an ongoing social activity in a the material cause or antecedently estab- continuing process of transformation. But lished knowledge which is used to gener- the aim of science is the production of the ate the new knowledge; and an intransitive knowledge of the mechanisms of the pro- dimension, in which the object is the real duction of phenomena in nature that com- xxx INTRODUCTION bine to generate the actual flux of phenom- that it is a special case. ena of the world. These mechanisms, which are the intransitive objects of scientific en- 17:2/o Although the primary aim of this quiry, endure and act quite independently of book is constructive, it is an important sub- men. The statements that describe their op- sidiary aim to situate the conditions of the erations, which may be termed ‘laws’, are plausibility of empirical realism and to show not statements about experiences (empirical it as depending upon what is in effect a statements, properly so called) or statements special case. These conditions are briefly: about events. Rather they are statements a naturally occurring closure, a mechanis- about the ways things act in the world (that tic conception of action and the model of is, about the forms of activity of the things man referred to earlier. The attempt to re- of the world) and would act in a world with- duce knowledge to an individual acquistion out men, where there would be no experi- in sense-experience and to view the latter as ences and few, if any, constant conjunctions the neutral ground of knowledge that (lit- of events. (It is to be able to say this inter erally) defines the world results in the gen- alia that we need to distinguish the domains eration of an ontology of atomistic discrete of the real, the actual and the empirical.) events, which if they are to be related at all (so making general knowledge possible) must ⇓ The next paragraph gives a sub- be constantly conjoined. (Hence the pre- sidiary aim: we want to explain why em- supposition of a closure.) On this view the pirical realism is so plausible and show causal connection must be contingent and xxxi actual; by contrast I want to argue that it velops in detail the conditions required for is necessary and real. the Humean analysis of laws and provides an Conditions: naturally occuring closure, analysis of normic statements (see esp. 2.4). Determinism is shown to be an immensely mechanistic conception of action, and the implausible thesis; and the central tenets of model of man as individualistic receptor orthodox philosophy of science—such as the of sense-experience which is the neutral principle of instance-confirmation (or falsifi- ground of knowledge. Here he derives the cation), the Humean theory of causality, the necessity of the implicit empiricist ontol- Popper-Hempel theory of explanation, the ogy: if events are to be related at all, they thesis of the symmetry between explanation must be constantly conjoined. What does and prediction, the criterion of falsifiability, he mean with last sentence in the above etc.—to be manifestly untenable. Chapter 3 paragraph? sets out to give a rational account of the pro- cess of scientific discovery; in which both na- ⇓ The next paragraph goes through ture and our knowledge of nature are seen as content of the book. stratified, as well as differentiated (see esp. 18:1 Chapter 1 establishes the necessity 3.3). A theory of natural necessity is devel- for an ontological distinction between causal oped which it is claimed is capable of resolv- laws and patterns of events (see esp. 1.3) ing inter alia the problems of induction and and contains a sketch of a critique of em- of subjunctive conditionals and Goodman’s pirical realism (see esp. 1.6). Chapter 2 de- xxxii INTRODUCTION and Hempel’s paradoxes (see 3.6). Chapter this study is natural science, something is 4 rounds off the argument and summarises said about the social sciences and about the some of the main themes of this study. characteristic pattern of explanation in his- ⇓ After this brief table-of-contents-like tory. summary a summary is given of the main 18:3/o If the first half of this work is con- ideas, and a brief sketch how they are pre- cerned with establishing the necessity for an sented in the book. Starts with general ontological distinction between causal laws remarks about concepts of tendencies and and patterns of events and tracing the im- powers. plications of the distinction between open 18:2 Moving towards a conception of sci- systems and closed, that is, of the differ- ence as concerned essentially with possi- entiation of our world, the second is con- bilities, and only derivatively with actuali- cerned principally with showing how science ties, much attention is given to the analy- can come to have knowledge of natural neces- sis of such concepts as tendencies and pow- sity a posteriori. The differentiation of the ers. Roughly the theory advanced here is world implies its stratification, if it is to be that statements of law are tendency state- a possible object of knowledge for us. If gen- ments. Tendencies may be possessed unexer- erative mechanisms and structures are real cised, exercised unrealised, and realized un- then there is a clear criterion for distinguish- perceived (or undetected) by men; they may ing between a necessary and an accidental also be transformed. Although the focus of sequence: a sequence Ea.Eb is necessary if xxxiii and only if there is a generative mechanism part, then we do have a reason independent or structure which when stimulated by the of its behaviour as to why it behaves the way event described by ‘Ea’ produces Eb. If we it does. Now such a reason may be discov- can have empirical knowledge of such gener- ered empirically. And if we can deduce the ative mechanisms or structures then we can thing’s tendency from it then the most strin- have knowledge of natural necessity a poste- gent possible (or Lockean) criterion for our riori. In showing how this is possible a non- knowledge of natural necessity is satisfied. Kantian ‘sublation’ of empiricism and ratio- For example, we may discover that copper nalism is achieved. has a certain atomic or electronic structure and then be able to deduce its dispositional 19:1 In the transitive process of science properties from a statement of that struc- three levels of knowledge may be distin- ture. We may then be said to have knowl- guished. At the first (or Humean) level we edge of natural necessity a posteriori. At the just have the invariance of an experimentally third (or Leibnizian) level we may seek to produced result. Given such an invariance express our discovery of the electronic struc- science moves immediately to the construc- ture of copper in an attempted real defini- tion and testing of possible explanations for tion of the thing. This is not to put an end it. If there is a correct explanation, located to enquiry, but a stepping stone to a new in the nature of the thing whose behaviour process of discovery in which we attempt to is described in the putative law or the struc- discover the mechanisms responsible for elec- ture of the system of which the thing is a xxxiv INTRODUCTION tronic structure. sons only in as much as they cast light on things. A realist theory of the universals 19:2/o In 3.5 the grounds for inductive of interest to science complements the real- scepticism are examined and shown to be ist theory of scientifically significant invari- fundamentally mistaken and in 3.6 the prob- ances, i.e. invariances generated under con- lem, which arises from the ontology of atom- ditions which are artificially produced and istic events (and closed systems), resolved. controlled. Dynamic realist principles of substance and 20:1 It is the argument of this book that causality are shown to be a condition of the if science is to be possible the world must intelligibility of experimental activity and consist of enduring and transfactually active the stratification of science. Science, it is ar- mechanisms; society must consist of an en- gued, is concerned with both taxonomic and semble of powers irreducible to but present explanatory knowledge: with what kinds of only in the intentional actions of men; and things there are, as well as how the things men must be causal agents capable of act- there are behave. It attempts to express the ing self-consciously on the world. They do former in real definitions of the natural kinds so in an endeavour to express to themselves and the latter in statements of causal laws, in thought the diverse and deeper structures i.e. of the tendencies of things. But it is con- that account in their complex manifold de- cerned with neither in an undiscriminating terminations for all the phenomena of our way. It is concerned with things only in as world. much as they cast light on reasons; and rea- Chapter 1

<16> Philosophy and Scientific Realism

1.1 Two Sides of ‘Knowledge’

Distinguishes between transitive and in- transitive objects of science (example 1 2 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM

Darwin) and in a very elegant way ar- light propagation. None of these ‘objects of rives at the transcendental question by knowledge’ depend upon human activity. If going through all possibilities: men ceased to exist sound would continue to travel and heavy bodies fall to the earth in Starting point is an interesting “central exactly the same way, though ex hypothesi paradox of sciences”: there would be no-one to know it. Let us call 21:1/o Any adequate philosophy of science these, in an unavoidable technical neologism, must find a way of grappling with this cen- the intransitive objects of knowledge. tral paradox of science: that men in their After introducing the intransitive ob- social activity produce knowledge which is a jects of knowledge, Bhaskar remarks that social product much like any other, which is there is also a different type of objects of no more independent of its production and knowledge, the “transitive” ones, which the men who produce it than motor cars, armchairs or books, which has its own crafts- are comparable to the means of produc- men, technicians, publicists, standards and tion in the production process. Transitive skills and which is no less subject to change objects of knowledge are prior knowledge than any other commodity. This is one side and the knowledge of other things. of ‘knowledge’. The other is that knowl- The transitive objects of knowledge are edge is ‘of’ things which are not produced by Aristotelian material causes.1 They are the men at all: the specific gravity of mercury, raw materials of science—the artificial ob- the process of electrolysis, the mechanism of jects fashioned into items of knowledge by 1.1. TWO SIDES OF ‘KNOWLEDGE’ 3 the science of the day.2 They include the too slow and complex to be perceived, which antecedently established facts and theories, had been going on for millions of years before paradigms and models, methods and tech- him. But he could not, at least if his theory niques of inquiry available to a particular sci- is correct, have produced the process he de- entific school or worker. scribed, the <17> intransitive object of the Bhaskar uses Darwin as an example il- knowledge he had produced: the mechanism lustrating the difference between intran- of natural selection. sitive and transitive objects. Darwin’s process of developing a scientific theory Question 8 What is the central paradox is compared with a production process. of science and how does Bhaskar propose Darwin produced knowledge albout the to deal with it? selection mechanism but he did not pro- Now Bhaskar goes through a tedious duce the mechanism itself. elimination argument which you may not The material cause, in this sense, of Dar- want to bother with. The next important win’s theory of natural selection consisted of the ingredients out of which he fashioned his high point in his argument is his question theory. Among these were the facts of natu- in 23:1/o ‘what must the world be like for ral variation, the theory of domestic selection science to be possible?’ and Malthus’ theory of population.3 Darwin ⇓ Intransitive world without science is worked these into a knowledge of a process, possible. I.e., intransitive objects do not 4 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM depend on science. drogen would continue to combine with one atom of oxygen and in favourable circum- 22:1 We can easily imagine a world simi- stances osmosis would continue to occur. In lar to ours, containing the same intransitive short, the intransitive objects of knowledge objects of scientific knowledge, but without are in general invariant to our knowledge of any science to produce knowledge of them. them: they are the real things and struc- In such a world, which has occurred and tures, mechanisms and processes, events and may come again, reality would be unspoken possibilities of the world; and for the most for and yet things would not cease to act part they are quite independent of us. They and interact in all kinds of ways. In such a are not unknowable, because as a matter of world the causal laws that science has now, fact quite a bit is known about them. (Re- as a matter of fact, discovered would pre- member they were introduced as objects of sumably still prevail, and the kinds of things scientific knowledge.) But neither are they that science has identified endure. The tides in any way dependent upon our knowledge, would still turn and metals conduct elec- let alone perception, of them. They are the tricity in the way that they do, without a intransitive, science-independent, objects of Newton or a Drude to produce our knowl- scientific discovery and investigation. edge of them. The Wiedemann-Franz law ⇓ But science without transitive ob- would continue to hold although there would be no-one to formulate, experimentally es- jects is not possible: tablish or deduce it. Two atoms of hy- 22:2/o If we can imagine a world of in- 1.1. TWO SIDES OF ‘KNOWLEDGE’ 5 transitive objects without science, we can- economists sought explanations of phenom- not imagine a science without transitive ob- ena which would conform to the paradigm jects, i.e. without scientific or pre-scientific of a decision-making unit maximizing an antecedents. That is, we cannot imag- objective function with given resources un- ine the production of knowledge save from, til marginalism became discredited in the and by means of, knowledge-like materi- 1930’s. No doubt at the back of economists’ als. Knowledge depends upon knowledge- minds during the period of the paradigm’s like antecedents. Harvey thought of blood hegemony was the cosy picture of a house- circulation in terms of an hydraulic model. wife doing her weekly shopping subject to Spencer, less successfully perhaps, used an a budget constraint; just as Rutherford dis- organic metaphor to express his idea of so- armingly confessed in 1934, long after the ciety. W. Thomson (Lord Kelvin) declared paradigm was hopelessly out of date, to in 1884 that it seemed to him that ‘the test a predilection for corpuscularian models of of “do we understand a particular topic in atoms and fundamental particles as ‘little physics [e.g. heat, magnetism]?” is “can we hard billiard balls, preferably red or black’.5 make a mechanical model of it?”.’4 And as Von Helmont’s concept of an arche <18> is well known this was the guiding maxim was the intellectual ancestor of the concept of physical research until the gradual dis- of a bacterium, which furnished the model integration of the Newtonian world-view in for the concept of a virus. The biochemi- the first decades of this century. Similarly cal structure of genes, which were initially 6 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM introduced as the unknown bearers of ac- ical as opposed to the scientific study of quired characteristics, has been explored un- the intransitive objects of science. Only der the metaphor of a linguistic code. In afterwards can one ask: What must sci- this way social products, antecedently estab- ence be like to give us knowledge of in- lished knowledges capable of functioning as transitive objects (of this kind)? the transitive objects of new knowledges, are used to explore the unknown (but knowable) 23:1/o If we cannot imagine a science intransitive structure of the world. Knowl- without transitive objects, can we imagine edge of B is produced by means of knowledge a science without intransitive ones? If the of A, but both items of knowledge exist only answer to this question is ‘no’, then a philo- in thought. sophical study of the intransitive objects of science becomes possible. ⇓ And science without intransitive ob- jects? If that is considered impossible, The answer to the transcendental question i.e., if one is a realist, then two ques- ‘what must the world be like for science to be possible?’ deserves the name of ontol- tions must be asked. This is important ogy. And in showing that the objects of sci- now, and unfortunately RB does not give ence are intransitive (in this sense) and of enough explanation how these two ques- a certain kind, viz. structures not events, it tions follow. First question: What must is my intention to furnish the new philoso- the world be like for science to be possi- phy of science with an ontology. The par- ble? This is what he calls the philosoph- allel question ‘what must science be like to 1.1. TWO SIDES OF ‘KNOWLEDGE’ 7 give us knowledge of intransitive objects (of ing both aspects of science; that is, of show- this kind)?’ is not a petitio principii of the ing how science which is a transitive process, ontological question, because the intelligibil- dependent upon antecedent knowledge and ity of the scientific activities of perception the efficient activity of men, has intransitive and experimentation already entails the in- objects which depend upon neither. That transitivity of the objects to which, in the is, it must be capable of sustaining both (1) course of these activities, access is obtained. the social character of science and (2) the That is to say, the philosophical position de- independence from science of the objects of veloped in this study does not depend upon scientific thought. More specifically, it must an arbitrary definition of science, but rather satisfy both: upon the intelligibility of certain universally 24:2 (1)’ a criterion of the non- recognized, if inadequately analysed, scien- spontaneous production of knowledge, tific activities. In this respect I am taking viz. the production of knowledge from and it to be the function of philosophy to anal- by means of knowledge (in the transitive yse concepts which are ‘already given’ but dimension), and ‘as confused’.6 24:3 (2)’ a criterion of structural and es- From this he derives four criteria of ad- sential realism, viz. the independent exis- equacy of philosophy of science. tence and activity of causal structures and 24:1 Any adequate philosophy of science things (in the intransitive dimension). must be capable of sustaining and reconcil- (Non-spontaneous production of 8 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM knowledge means: the world does not knowledge is map of world. Science is reveal itself automatically. Science is “automatic,” a behavioral response to work.) conjunction of events, epiphenomenon of ⇓ The last paragraph gives a succinct nature. formulation of results. 24:4 For science, I will argue, is a social 24:5/o Viewed historically, three broad activity whose aim is the production of the positions in the philosophy of science may be knowledge of the kinds and ways of acting of distinguished. According to the first, that of independently existing and active things. classical empiricism, represented by Hume and his heirs, the ultimate objects of knowl- edge are atomistic events. Such events con- 1.2 <19> Three Tradi- stitute given facts and their conjunctions ex- haust the objective content of our idea of tions in the Philoso- natural necessity. Knowledge and the world may be viewed as surfaces whose points are phy of Science in isomorphic correspondence or, in the case of phenomenalism, actually fused. On this ⇓ (1) The first tradition is classical em- conception, science is conceived as a kind piricism: ultimate objects of knowledge of automatic or behavioural response to the are atomistic events, natural necessity stimulus of given facts and their conjunc- = conjunction of events. World is flat, tions. Even if, as in logical empiricism, such 1.2. THREE TRADITIONS IN THE PHILOSOPHY OF SCIENCE 9 a behaviourism is rejected as an account of 25:1 The second position received its clas- the genesis of scientific knowledge, its valid sical though static formulation in Kant’s content can still in principle be reduced to transcendental idealism, but it is susceptible such facts and their conjunctions. Thus sci- of updated and dynamized variations. Ac- ence becomes a kind of epiphenomenon of cording to it, the objects of scientific knowl- nature. edge are models, ideals of natural order etc. ⇓ (2) Kant’s transcendental idealism: Such objects are artificial constructs and objects of knowledge are socially gener- though they may be independent of particu- ated “ideals,” i.e., human constructs im- lar men, they are not independent of men or posed on the phenomena. This discus- human activity in general. On this concep- sion here is much too brief! Presumably, tion, a constant conjunction of events is in- sufficient, though it is still necessary, for the that part of nature which is not a prod- attribution of natural necessity. Knowledge uct of mankind (things in themselves) is is seen as a structure rather than a surface. not knowable. But the natural world becomes a construc- tion of the human mind or, in its modern Question 9 Both Bhaskar and Marx are versions, of the scientific community. critical of idealism, but Bhaskar is also critical of empiricism. Define empiricism and explain why Bhaskar is critical of it. ⇓ (3) Transcendental Realism, which is What is Marx’s stance on empiricism? nowadays usually called Critical Realism 10 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM

(CR), uses correct elements of both: to this view, both knowledge and the world are structured, both are differentiated and 25:2 The third position, which is advanced changing; the latter exists independently of here, may be characterized as transcendental the former (though not our knowledge of this realism. It regards the objects of knowledge fact); and experiences and the things and as the structures and mechanisms that gen- causal laws to which it affords us access are erate phenomena; and the knowledge as pro- normally out of phase with one another. On duced in the social activity of science. These this view, science is not an epiphenomenon objects are neither phenomena (empiricism) of nature, nor is nature a product of man. nor human constructs imposed upon the ⇓ This typology is not exhaustive, phenomena (idealism), but real structures it leaves out rationalism and absolute which endure and operate independently of idealism, which are in-between states. our knowledge, our experience and the condi- Bhaskar is describing an extreme posi- tions which allow us access to them. Against empiricism, the objects of knowledge are tion here, but he will argue that this is structures, not events; against idealism, they the only position “that can do justice to are intransitive (in the sense defined). On science.” this conception, a constant conjunction of 26:1 A word of caution is necessary here. events is no more a necessary than it is In outlining these positions, I am not offer- a sufficient condition for the assumption of ing them as a complete typology, but only as the operation of a causal law. According one which will be of some <20> significance 1.2. THREE TRADITIONS IN THE PHILOSOPHY OF SCIENCE 11 in illuminating current issues in the philos- only postion that can do justice to science. ophy of science. Thus I am not concerned ⇓ Transcendental realism distinguished with rationalism as such, or absolute ideal- from empirical realism, realism with re- ism. Moreover, few, if any, modern philoso- phers of science could be unambiguously lo- spect to causal laws etc. cated under one of these banners. Nagel for 26:2 Transcendental realism must be dis- example stands somewhere along the contin- tinguished from, and is in direct opposition uum between Humean empiricism and neo- to, empirical realism. This is a doctrine to Kantianism; Sellars nearer the position char- which both classical empiricism and tran- acterized here as transcendental realist; and scendental idealism subscribe. My reasons so on. One could say of such philosophers for rejecting it will be elaborated in a mo- that they combine, and when successful in ment. ‘Realism’ is normally associated by an original way synthesize, aspects of those philosophers with positions in the theory of philosophical limits whose study we are un- perception or the theory of universals. In the dertaking. It is my intention here, in working former case the real entity concerned is some out the implications of a full and consistent particular object of perception; in the latter realism, to describe such a limit; in rather case some general feature or property of the the way Hume did. As an intellectual ex- world. The ‘real entities’ the transcendental ercise alone this would be rewarding, but I realist is concerned with are the objects of believe, and hope to show, that it is also the scientific discovery and investigation, such as causal laws. Realism about such entities will 12 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM be seen to entail particular realist positions above. Moreover in its most consistent forms in the theory of perception and universals, it involves both solipsism and phenomenal- but not to be reducible to them. ism; so that neither (1) nor (2) can be up- held. In particular not even the idea of the Question 10 What is the criterion for independence of the event from the experi- being real in transcendental realism? ence that grounds it, i.e. the intransitivity How does it differ from the criterion for of events, can be sustained; and, in the last being real in empirical realism? instance, events must be analysed as sensa- tions or in terms of what is epistemologically Adequacy criteria applied to the above equivalent, viz. human operations. three positions: 26:3 Only transcendental realism, I will Question 11 Bhaskar claims that Clas- argue, can sustain the idea of a law-governed sical empiricism can sustain neither tran- world independent of man; and it is this con- cept, I will argue, that is necessary to under- sitive nor intransitive dimensions. Why stand science. not? What is his argument? ⇓ Classical empiricism: 26:4/o Classical empiricism can sustain ⇓ Transcendental idealism: neither transitive nor intransitive dimen- 27:1 Transcendental idealism attempts to sions; so that it fails both the criteria of ad- uphold the objectivity (intersubjectivity) of equacy (1)’ and (2)’ advanced on page 24 facts, i.e. (1). And, if given a dynamic gloss, 1.2. THREE TRADITIONS IN THE PHILOSOPHY OF SCIENCE 13 it can allow a transitive dimensions and sat- that there could be no knowledge without isfy criterion (1)’; so that, in this respect, it the social activity of science. They disagree is an improvement on empiricism. According over whether in this case there would be no to such a dynamized transcendental idealism nature also. Transcendental realism argues knowledge is given structure by a sequence that it is necessary to assume for the intel- of models, rather than a fixed set of a pri- ligibility of science that the order discovered ori rules. However in neither its static nor in nature exists independently of men, i.e. its dynamic form can it sustain the intransi- of human activity in general. Transcenden- tive dimension. For in both cases the objects tal idealism maintains that this order is ac- of which knowledge is obtained do not exist tually imposed by men in their cognitive ac- independently of human activity in general. tivity. Their differences should thus be clear. And if there are things which do (things-in- According to transcendental realism, if there themselves), no scientific knowledge of them were no science there would still be a nature, can be obtained. and it is this nature which is investigated by science. Whatever is discovered in na- ⇓ Transcendental realism: ture must be expressed in thought, but the 27:2 <21> Both transcendental realism structures and constitutions and causal laws and transcendental idealism reject the em- discovered in nature do not depend upon piricist account of science, according to thought. Moreover, the transcendental real- which its valid content is exhausted by atom- ist argues, this is not just a dogmatic meta- istic facts and their conjunctions. Both agree 14 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM physical belief; but rather a philosophical po- cist account of being. This ontological legacy sition presupposed by key aspects of the so- is expressed most succinctly in its commit- cial activity of science, whose intelligibility ment to empirical realism, and thus to the the transcendental idealist cannot thus, any- concept of the ‘empirical world’. For the more than the empiricist, sustain. transcendental realist this concept embodies Now he starts arguing that the most a sequence of related philosophical mistakes. fundamental difference between these The first consists in the use of the category three approaches is their underlying on- of experience to define the world. This in- volves giving what is in effect a particular tology. epistemological concept a general ontologi- ⇓ Three errors implied in the concept cal function. The second consists in the view of the “empirical world.” that its being experienced or experienciable 27:3/o Neither classical empiricism nor is an essential property of the world; whereas transcendental idealism can sustain the idea it is more correctly conceived as an acciden- of the independent existence and action of tal property of some things, albeit one which the causal structures and things investigated can, in special circumstances, be of great sig- and discovered by science. It is in their nificance for science. The third thus consists shared ontology that the source of this com- in the neglect of the (socially produced) cir- mon incapacity lies. For although transcen- cumstances under which experience is in fact dental idealism rejects the empiricist account epistemically significant in science. of science, it tacitly takes over the empiri- 1.2. THREE TRADITIONS IN THE PHILOSOPHY OF SCIENCE 15

Question 12 Which three interrelated instead a proposition about the way men un- mistakes are buried in the concept of the derstand it. Structure becomes a function of “empirical world.”? human needs; it is denied a place in the world of things. ⇓ For the transcendental idealist, the But if one says that science can only difference between necessary and acci- help humans to sort out our experiences, dental sequences is not real but merely that natural laws do not exist in reality describes how humans understand the but only in our brains, then one cannot world. Therefore it cannot explain how understand the actual process by which theories are constructed and tested. science is gained (experiments): 28:1 If the bounds of the real and the em- But just because of this, I shall argue, the pirical are co-extensive then of course any transcendental idealist cannot adequately ‘surplus-element’ which the transcendental describe the principles <22> according to idealist finds in the analysis of law-like state- which our theories are constructed and em- ments cannot reflect a real difference be- pirically tested; so that the rationality of tween necessary and accidental sequences of the transitive process of science, in which events. It merely reflects a difference in our knowledge of the world is continually ex- men’s attitude to them. Saying that light tended and corrected, cannot be sustained. travels in straight lines ceases then to express ⇓ From these criticisms one can see a proposition about the world; it expresses that science cannot be ontologically neu- 16 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM tral. Some say that Bhaskar “has put discover them. Such a philosopher is then ontology back into philosophy.” committed to the belief that, given that sci- ence occurs, there are such conjunctions. As 28:2/o To say that the weaknesses of both Mill put it, that ‘there are such things in the empiricist and idealist traditions lie in nature as parallel cases; that what happens their commitment to empirical realism is of once will, under a sufficient degree of simi- course to commit oneself to the impossibility larity of circumstance, happen again’.7 of ontological neutrality in an account of sci- One of Bhaskar’s central arguments is ence; and thus to the impossibility of avoid- ing ontological questions in the philosophy of however that such constant conjunctions science. The sense in which every account of are extremely rare; they occur in astron- science presupposes an ontology is the sense omy, but they rarely occur spontaneously in which it presupposes a schematic answer anywhere else. It is still important to dis- to the question of what the world must be cover natural laws, not because they pro- like for science to be possible. Thus suppose duce constant conjunctions, but because a philosopher holds, as both empiricists and they continue to be active even in envi- transcendental idealists do, that a constant ronments that are not conducive to con- conjunction of events apprehended in sense- stant conjunctions. experience is at least a necessary condition for the ascription of a causal law and that it ⇓ Distinction between philosophical is an essential part of the job of science to and scientific ontology. Gives examples 1.2. THREE TRADITIONS IN THE PHILOSOPHY OF SCIENCE 17 of second-order conclusions: world must that the world is a certain way. It is contin- be stratified and differentiated, because gent that the world is such that science is science would not be possible otherwise. possible. And, given that it is possible, it 29:1/o There are two important points to is contingent upon the satisfaction of certain register about such ontological beliefs and social conditions that science in fact occurs. commitments. But given that science does or could occur, the world must be a certain way. The first point takes up the rest of this paragraph; the second is the next para- Now Bhaskar brings some examples of graph 30:1. such second-order arguments: The first is that they should only be inter- Thus, the transcendental realist asserts, preted hypothetically, viz. as entailing what that the world is structured and differenti- must be the case for science to be possible; ated can be established by philosophical ar- on which interpretation it is a contingent fact gument; though the particular structures it that the world is such that science can occur. contains and the ways in which it is differen- It is only in this relative or conditional sense tiated are matters for substantive scientific that an account of science presupposes an investigation. The necessity for categorical ontology. The status of propositions in on- distinctions between structures and events tology may thus be described by the follow- and between open systems and closed are in- ing formula: It is not necessary that science dices of the stratification and differentiation occurs. But given that it does, it is necessary of the world, i.e. of the transcendental realist 18 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM philosophical ontology. comes up with depends on one’s concept of science. But this does not mean that Question 13 What does Bhaskar mean the ontological statements one comes up when he says that the world is “stratified” with are in the last analysis statements or “structured and differentiated”? How about science. They are not. The can this be shown by a philosophical ar- mistaken belief that everything one says gument? about ontology is really something about science is called the “epistemic fallacy.” These distinctions are presupposed, it will 30:1 The second point to stress is that be shown, by the intelligibility of experimen- propositions in ontology cannot be <23> es- tal activity. Whenever there is any danger of tablished independently of an account of sci- confusion between an ‘ontology’ in the sense ence. On the contrary, they can only be es- of the kind of world presupposed by a philo- tablished by reference to such an account, or sophical account of science and in the sense at least to an account of certain scientific ac- of the particular entities and processes pos- tivities. However, it will be contended that tulated by some substantive scientific theory this essential order of analysis, viz. science I shall explicitly distinguish between a philo- → being, reverses the real nature of depen- sophical and a scientific ontology. dency (or, we could say, the real burden of The flip side of this definition of philo- contingency). For it is not the fact that sci- sophical ontology is: the ontology one ence occurs that gives the world a structure 1.3. THE TRANSCENDENTAL ANALYSIS OF EXPERIENCES 19 such that it can be known by men. Rather, Question 14 What is the definition of it is the fact that the world has such a struc- ontology? of epistemology? ture that makes science, whether or not it actually occurs, possible. That is to say, it Question 15 Explain the difference be- is not the character of science that imposes tween philosophical and scientific ontol- a determinate pattern or order on the world; ogy. but the order of the world that, under cer- tain determinate conditions, makes possible the cluster of activities we call ‘science’. It 1.3 The Transcendental does not follow from the fact that the nature of the world can only be known from (a study Analysis of Experi- of) science, that its nature is determined by (the structure of) science. Propositions in ences ontology, i.e. about being, can only be es- ⇓ Goal is to show that empiricism fails tablished by reference to science. But this does not mean that they are disguised, veiled in its most favored case: it cannot even or otherwise elliptical propositions about sci- do justice to the role of experience in sci- ence. What I shall characterize in a moment ence. This is called an Achilles-heel cri- as the ‘epistemic fallacy’ consists in assum- tique: criticize something where it seems ing that, or arguing as if, they are. strongest. 30:2/o The empiricist ontology is consti- 20 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM tuted by the category of experience. What nor transcendental idealism can reveal the transcendental arguments can be produced true significance of experience in science. to show its inadequacy to science; and, on the other hand, to demonstrate the intran- Question 16 What is empiricism? sitivity and structured character of the ob- jects of scientific knowledge? Now the oc- Question 17 What is the implicit ontol- currence of experience in science would be ogy of empiricism? Which transcendental agreed upon by all three combatants. More- arguments suggest that the implicit ontol- over, it is generally assumed that, what- ogy of empiricism cannot be right? ever its other inadequacies, empiricism can at least do justice to the role of experience Question 18 Bhaskar’s Realist Theory in science. Now I want to argue that the intelligibility of experience in science itself of Science gives an Achilles-heel critique presupposes the intransitive and structured of empiricism. Explain. character of the objects to which, in scientific experience, ‘access’ is obtained. This estab- ⇓ One will not get scientifically signifi- lishes the inadequacy, in its most favoured cant experiences if one just watches; one case, of the empiricist ontology. Further I has to set up experiments first: want to argue that, in virtue of their shared 31:1 Scientifically significant experience ontological commitment, neither empiricism normally depends upon experimental activ- ity as well as sense-perception; that is, upon 1.3. THE TRANSCENDENTAL ANALYSIS OF EXPERIENCES 21 the role of men as causal agents as well as have been and could be <24> deployed to perceivers. I will consider the two indepen- demonstrate this, which there is no space dently. here to rehearse. For our purposes, it is suf- ficient merely to note that both the possi- bility of scientific change (or criticism) and 1.3.A The Analysis of Per- the necessity for a scientific training presup- ception pose the intransitivity of some real objects; which, for the empirical realist at least, can 31:2/o The intelligibility of sense-perception only be objects of perception. presupposes the intransitivity of the object Bhaskar says that the arguments are perceived. known. One is change or criticism: one ⇑ One can only make sense of percep- cannot revise or criticize one’s perception tion if the perceived objects are intransi- unless it refers to something. tive. This first sentence summarizes the whole paragraph. Question 19 Explain why scientific For it is in the independent occurrence or ex- istence of such objects that the meaning of change or criticism would not be pos- ‘perception’, and the epistemic significance sible, and scientific training would not of perception, lies. Among such objects are be necessary, if there would not be events, which must thus be categorically in- intransitive objects. dependent of experiences. Many arguments 22 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM

If changing experience of objects is to be At the end, Bhaskar says, he does not possible, objects must have a distinct being think that events and momentary states in space and time from the experiences of are the primary objects of perception, which they are the objects. For Kepler to see which are probably processes and things, the rim of the earth drop away, while Tycho but Bhaskar does not want to argue this Brahe watches the sun rise, we must suppose here because it anticipates results still to that there is something that they both see be derived. (in different ways).8 Similarly when modern sailors refer to what ancient mariners called a Events and momentary states do not of sea-serpent as a school of porpoises, we must course exhaust the objects of perception. In- suppose that there is something which they deed, I do not think they are even the pri- are describing in different ways.9 The intel- mary objects of perception, which are prob- ligibility of scientific change (and criticism) ably processes and things, from which events 10 and scientific education thus presupposes the and states are then ‘reconstructed’. How- ontological independence of the objects of ever I do not wish to argue the point here— experience from the objects of which they as it depends upon a prior resolution of the are the experiences. problems of causality and induction, upon which their status as objects of experience I think this is the concept of “referen- must, at least for the empiricist, depend.11 tial detachment” which Bhaskar uses in 32:1 Events then are categorically inde- his later works. pendent of experiences. There could be a 1.3. THE TRANSCENDENTAL ANALYSIS OF EXPERIENCES 23 world of events without experiences. Such ceived. events would constitute actualities unper- Therefore an ontology which defines ceived and, in the absence of men, unper- things that are real by their being per- ceivable. There is no reason why, given the possibility of a world without perceptions, ceivable is wrong. which is presupposed by the intelligibility And of such events theoretical knowledge of actual scientific perceptions, there should may or may not be possessed, and may or not be events in a world containing percep- may not be achievable. Clearly if at some tions which are unperceived and, given our particular time I have no knowledge of an current or permanent capacities, unperceiv- unperceived or unperceivable event, I cannot able. say that such an event occurred (as a pu- Interesting derivation of the existence tative piece of substantive knowledge). But of unperceived events in our world (i.e., that in itself is no reason for saying that such the empirical is a subset of the actual). an occurrence is impossible or that its sup- position is meaningless (as a piece of philos- From intransitivity of the objects of per- ophy). To do so would be to argue quite ception follows: there would also be illicitly from the current state of knowledge things if people did not exist, i.e., if ev- to a philosophical conception of the world. erything is unperceived. From there one Indeed, we know from the history of science must conclude that also in the presence that at any moment of time there are types of of people things exist which are unper- events never imagined, of which theoretical, 24 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM and sometimes empirical, knowledge is even- ertheless it is possible to make meaningful tually achieved. For in the transitive pro- statements, even if one can never know cess of science the possibilities of perception, whether they are true. How does this re- and of theoretical knowledge, are continu- late to the main thrust of the argument? ally being extended. Thus unless it is dog- matically postulated that our present knowl- edge is complete or these possibilities ex- ⇓ Later he will show that the domain hausted, there are good grounds for holding of actualities consists not only of events that the class of unknowable events is non- but also of things. empty, and unperceivable ones non-emptier; and no grounds for supposing that this will ever not be so. 32:2 <25> Later, I will show how the do- Another argument in favor of the ex- main of actualities, whose categorical inde- istence of unperceived actualities is also pendence from experiences is presupposed by that the range of perceived things con- the intelligibility of sense-perception, may be extended to include things as well as events. tinually expands, and one should not as- sume that this process of expansion has reached its limits. Where and how does he show that, and He also talkes about the unperceivable does he imply here that empiricism only and the unknowable, about which it nev- limit itself to events? 1.3. THE TRANSCENDENTAL ANALYSIS OF EXPERIENCES 25

1.3.B Analysis of Experi- under experimental conditions would not be mental Activity forthcoming without it. Thus in an experi- ment we are a causal agent of the sequence ⇓ In an experiment we are a causal agent of events, but not of the causal law which the sequence of events, because it has been of the sequence of events, but not of the produced under experimental conditions, en- causal law which the sequence of events ables us to identify. enables us to identify. ⇓ Two implications of this: (1) there 33:1 The intelligibility of experimental ac- must be a difference between causal laws tivity presupposes not just the intransitivity and sequences of events, and (2) experi- but the structured character of the objects investigated under experimental conditions. ments only make sense if the causal laws Let me once again focus on the empiricist’s prevail also outside the context under favourite case, viz. causal laws, leaving aside which they generate sequences of events. for the moment such other objects of inves- 33:2 Two consequences flow from this. tigation as structures and atomic constitu- First, the real basis of causal laws cannot be tions. A causal law is analysed in empiri- sequences of events; there must be an onto- cist ontology as a constant conjunction of logical distinction between them. Secondly, events perceived (or perceptions). Now an experimental activity can only be given a experiment is necessary precisely to the ex- satisfactory rationale if the causal law it en- tent that the pattern of events forthcoming ables us to identify is held to prevail out- 26 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM side the contexts under which the sequence and patterns of events conforms with our of events is generated. In short, the intelli- intuitions. Regular occurences are often gibility of experimental activity presupposes disturbed. Everyone knows that it is easy that a constant conjunction is no more a nec- to muck up an experiment, but nobody essary than a sufficient condition for a causal thinks this changes the causal laws. law. And it implies that causal laws endure and continue to operate in their normal way 33:3/o Both Anscombe and von Wright under conditions, which may be character- have recently made the point that our active ized as ‘open’, where no constant conjunc- interference in nature is normally a condi- tion or regular sequence of events is forth- tion of empirical regularities.12 But neither coming. It is worth noting that in general, have seen that it follows from this that there outside astronomy, closed systems, viz. sys- must be an ontological distinction between tems in which constant conjunctions occur, the empirical regularity we produce and the must be experimentally established. causal law it enables us to identify. Although it has yet to be given an adequate philosoph- Question 20 How can experimental ac- ical rationale, the distinction between causal tivity be used to prove that constant con- laws and patterns of events is consistent with junctions of events are neither necessary our intuitions. Thus supposing a nuclear ex- nor sufficient for causal laws? plosion were to destroy our planet no-one would hold that it violated, rather than ex- ⇓ This distinction between causal laws emplified, Newton’s laws of motion;13 just 1.3. THE TRANSCENDENTAL ANALYSIS OF EXPERIENCES 27 as if something were to affect Mercury’s per- systems. ihelion it would not be regarded as falsifying 34:1 Thus the intelligibility of experimen- Einstein’s theory of relativity. Similarly it tal activity presupposes the categorical inde- lies within the power of every reasonably in- pendence of the causal laws discovered from telligent schoolboy or moderately clumsy re- the patterns of events produced. For, to re- search worker to upset the results of even the peat, in an experiment we produce a pattern 14 best designed experiment, but we do not of events to identify a causal law, but we do thereby suppose they have the power to over- not produce the causal law identified. Once turn the laws of nature. I can <26> quite the categorical independence of causal laws easily affect any sequence of events designed and patterns of events is established, then to test say Coulomb’s or Gay-Lussac’s law; we may readily allow that laws continue to but I have no more power over the relation- operate in open systems, where no constant ships the laws describe than the men who conjunctions of events prevail. And the ra- discovered them had. In short, laws cannot tional explanation of phenomena occurring be the regularities that constitute their em- in such systems becomes possible. pirical grounds. ⇓ Now he brings in the other point already mentioned in 33:2, that sponta- ⇓ To sum up: causal laws independent neously occurring patterns of events are of constant conjunctions of events, this rare. Regular patterns of events are usu- allows us to explain phenomena in open ally generated by human activity. There- 28 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM fore both experiences and regular pat- ways is empiricism anthropocentric? terns of events presuppose humans, i.e., empiricist ontology is anthropocentric. ⇓ If one identifies causal laws with em- 34:2 In a world without men there would pirical regularities one makes two cate- be no experiences and few, if any, constant gory mistakes: (a) One identifies events conjunctions of events, i.e. had they been ex- with experiences and (b) constant con- perienced Humean ‘causal laws’. For both junctions of events with causal laws experiences and invariances (constant con- (which presupposes the ubiquity of closed junctions of events) depend, in general, upon systems). This double identification pre- human activity. But causal laws do not. vents the empirical realist from seeing the Thus in a world without men the causal laws conditions under which experiences are that science has now as a matter of fact dis- significant for science, (a) the perceiver covered would continue to prevail, though there would be few sequences of events and must be theoretically informed, and (b) no experiences with which they were in cor- the system in which the events occur respondence. Thus, we can begin to see how must be closed. the empiricist ontology in fact depends upon 34:3/o The concept of causal laws be- a concealed anthropocentricity. ing or depending upon empirical regulari- ties involves thus a double identification: of Question 21 Explain the word events and experiences; and of constant con- “antrhropocentric.” In which two junctions (or regular sequences) of events 1.3. THE TRANSCENDENTAL ANALYSIS OF EXPERIENCES 29 and causal laws. This double identification lying causal structures which are the objects involves two category mistakes, expressed of his theory. And not until the categori- most succinctly in the concepts of the empir- cal independence of causal laws, patterns of ical world and the actuality of causal laws. events and experiences has been philosophi- The latter presupposes the ubiquity of closed cally established and the possibility of their systems. Both concepts, I shall argue, are disjuncture thereby posed can we appreciate profoundly mistaken and have no place in the enormous effort—in experimental design any philosophy of science. and scientific training—that is required to The empirical realist cannot explain make experience epistemically significant in science. why scientific experiments take so much effort. ⇓ The next paragraph summarizes the transcendental implications of exper- This double identification prevents the em- imental activity: the intransitive char- pirical realist from examining the important acter of the objects of scientific knowl- question of the conditions under which expe- rience is in fact significant in science. In gen- edge; (hence by extension?) possibility of eral this requires both that the perceiver be a nonhuman world. (Secondly?,) the pos- theoretically informed15 and that the system sibility of open systems. in which the events occur be closed.16 Only 35:1 The intelligibility of experimental ac- under such conditions can the experimental tivity presupposes then the intransitive and scientist come to have access to those under- structured character of the objects of scien- 30 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM tific knowledge, at least in so far as these tive definition will be given in 52:1/o. are causal laws. And this presupposes in turn the possibility of a non-human world, i.e. causal laws without invariances and expe- riences, and in particular of a non-empirical world, i.e. causal laws and events without <27> experiences; and the possibility of open systems, i.e. causal laws out of phase 35:2/o In saying that the objects of sci- with patterns of events and experiences, and entific discovery and investigation are ‘in- more generally of epistemically insignificant transitive’ I mean to indicate therefore that experiences, i.e. experiences out of phase they exist independently of all human activ- with events and/or causal laws. ity; and in saying that they are ‘structured’ that they are distinct from the patterns of ⇓ Causal laws are therefore not empir- events that occur. The causal laws of nature are not empirical statements, i.e. statements ical statements (they are not even state- about experiences; nor are they statements ments about events), neither are they about events; nor are they synthetic a priori synthetic a priori statements. Bhaskar statements. For the moment I merely style does not say what they are, he merely them negatively as ‘structured intransitive’, rules out the only two possibilities the postponing a positive analysis of them until empirical realist considers. A more posi- Section 5. 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY31

1.4 Status of Ontology ological mistakes within the tradition of empirical realism. For if the intelligibility and its Dissolution of experimental activity entails that the in Classical Philoso- objects of scientific understanding are intransitive and structured then we can es- phy tablish at one stroke: (i) that a philosophical ontology is possible; (ii) some propositions ⇓ Our argument that experimental activ- in it (causal laws are distinct from patterns ity only then makes sense if the world of events, and events from experiences); and is structured and if causal laws are not (iii) the possibility of a philosophy which constant conjunctions of events shows the is consistent with (and has some relevance possibility of a philosophical ontology, for), i.e. which is itself ‘in phase with’, the realist practice of science. Ontology, which talks about the same world as the it should be stressed, does not have as its scientific ontology, and which has been subject matter a world apart from that discredited because its early attempts investigated by science. Rather, its subject were “treating a mysterious underlying matter just is that world, considered from physical realm.” the point of view of what can be established 36:1 This analysis of experimental about it by philosophical argument. The episodes enables us to isolate a series of idea of ontology as treating of a mysterious metaphysical, epistemological and method- 32 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM underlying physical realm, which owes a or analysed in terms of statements about lot to Locke and some of his rationalist knowledge; i.e. that ontological questions contemporaries (particularly Leibniz), has can always be transposed into epistemologi- done much to discredit it; and to prevent cal terms. The idea that being can always be metaphysics from becoming what it ought to analysed in terms of our knowledge of being, be, viz. a conceptual science. Philosophical that it is sufficient for philosophy to ‘treat ontology asks what the world must be like only of the network, and not what the net- for science to be possible; and its premises work describes’,17 results in the systematic are generally recognized scientific activities. dissolution of the idea of a world (which I Its method is transcendental; its premise shall here <28> metaphorically characterize science; its conclusion the object of our as an ontological realm) independent of but present investigation. investigated by science. And it is manifest in ⇓ Now the epistemic fallacy is intro- the prohibition of any transcendent entities. duced (apart from the earlier brief men- How does the epistemic fallacy result in tions in 16:1 and 30:1). the systematic dissolution of the intran- sitive dimension? 36:2/o The metaphysical mistake the ar- gument of the previous section allows us Here is an attempt of an answer: to pinpoint may be called the ‘epistemic Bhaskar’s transitive-intransitive distinc- fallacy’. This consists in the view that tion not only entails that there is a world statements about being can be reduced to out there independent of science (and 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY33 also independent of the existence of hu- as the naturalistic fallacy prevents us from mans), but moreover that the objects of saying what is good about e.g. maximizing scientific knowledge are located in this utility in society, so the epistemic one pre- world. Ther epistemic fallacy says that vents us from saying what is epistemically science is not about the outside world, significant about e.g. experience in science. but that in the last analysis science is What is the naturalistic fallacy in only about what people do in their heads. moral philosophy? How does the natu- Someone who commits the epistemic fal- ralistic fallacy prevents us from asking: lacy will, if he or she is consistent, be led what is good about maximizing utility in to hold that a world independent of sci- society? ence does not exist. Question 22 How does the epistemic Furthermore, Bhaskar says that the fallacy prevent us from asking which ex- epistemic fallacy manifests itself in the periences are relevant for science and prohibition of any transcendent entities. which aren’t? What is a transcendent entity? Is it an entity of whose existence we can only in- If you consider the epistemic fallacy fer by transcendental argument? difficult to understand, don’t worry, it is It might be usefully compared with the nat- indeed a difficult subject. Tracing the ef- uralistic fallacy in moral philosophy. For just fects of the epistemic fallacy is one of the 34 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM principal subjects of this book. [Col94, temic fallacy is indeed a fallacy? pp. 76–85] has an extra section devoted Attempt at an answer: Bhaskar al- to it. ready showed that it is a fallacy by the To show that it is a fallacy and to trace argument that causal laws cannot be re- its effects are two of the principal objectives duced to empirical regularities, i.e., to of this study. In showing that the intelli- what can be known about them. gibility of experimental activity entails that ⇓ The next paragraph is very dense, the objects of scientific knowledge, in so far brings many examples how the tradition as they are causal laws, are intransitive I of empirical realism commits the epis- have already succeeded in the first of these temic fallacy: aims. For this means that a statement of a 37:1/o The epistemic fallacy is most causal law cannot now be reduced to or anal- marked, perhaps, in the concept of the em- ysed in terms of a statement about anyone’s pirical world. But it is manifest in the cri- knowledge of it or knowledge in general. On teria of significance and even the problems the contrary, its assertion now entails that a associated with the tradition of empirical re- causal law would operate even if unknown, alism. Kant committed it in arguing that the and even if there were no-one to know it. So categories ‘allow only of empirical employ- that knowledge ceases to be, as it were, an ment and have no meaning whatsoever when essential predicate of things. not applied to objects of possible experience; How has Bhaskar shown that the epis- that is to the world of sense.’18 (For us on 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY35 the other hand if the Kantian categories were tween natural and epistemic possibility. Fur- adequate to the objects of scientific thought ther there is no need to assume that the or- then they would continue to apply in a world der of dependence of being must be the same without sense, and have a meaning in rela- as the order of dependence of our knowledge tion to that possibility.) Similarly, the logi- of being. Thus we can allow that experience cal positivists committed it when arguing, in is in the last instance epistemically decisive, the spirit of Hume, that if a proposition was without supposing that its objects are onto- not empirically verifiable (or falsifiable) or logically ultimate, in the sense that their ex- a tautology, it was meaningless.19 Verifica- istence depends upon nothing else. Indeed tionism indeed may be regarded as a partic- if science is regarded as a continuing pro- ular form of the epistemic fallacy, in which cess of discovery of ever finer and in an ex- the meaning of a proposition about reality planatory sense more basic causal structures, (which cannot be designated ‘empirical’) is then it is rational to assume that what is at confused with our grounds, which may or any moment of time least certain epistemi- may not be empirical, for holding it. Once cally speaking is most basic from the onto- this doctrine is rejected there is no need to logical point of view.20 More generally, the identify the necessary and the a priori, and epistemic fallacy is manifest in a persistent the contingent and the a posteriori; or, to tendency to read the conditions of a partic- put it another way, one can distinguish be- ular concept of knowledge into an implicit tween natural and logical necessity, and be- concept of the world. Thus the problem 36 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM of induction is a <29> consequence of the knowledge, if, as in some varieties of Pla- atomicity of the events conjoined, which is a tonism and rationalism, he were to define function of the necessity for an epistemically the world in terms of the possibility of non- certain base. empirical knowledge of it. For the transcen- dental realist it is not a necessary condition Question 23 What is the problem of in- for the existence of the world that science oc- duction? How is the epistemic fallacy curs. But it is a necessary condition for the connected with the problem of induction? occurrence of science that the world exists and is of a certain type. Thus the possibility of our knowing it is not an essential property, ⇓ The next paragraph gives then the and so cannot be a defining characteristic, of significance of the epistemic fallacy in the world. Rather on a cosmic scale, it is an other traditions, namely, for rationalism. historical accident; though it is only because Rationalism says, in brief, that the world of this accident that we can establish in sci- ticks the same way as our brains tick. ence the way the world is, and in philosophy 38:1/o Although the epistemic fallacy is of the way it must be for science to be possible. most interest to us as it is manifest in the tra- dition of empirical realism, it is worth men- Apparently Bhaskar argues here as fol- tioning that a philosopher who rejected em- lows: the rationalist doctrine implies pirical realism might still commit the epis- that everything about the world can be temic fallacy, i.e. analyse being in terms of known. But the possibility of know- 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY37 ing the world is not an essential prop- exist’ refers to a possible content of a scien- erty of the world, since we humans are tific theory then that it is merely a part of only a contingent outcome and the world what we can know is an uninteresting tru- could also have gone on without humans. ism. But a philosophical ontology is devel- Therefore it can also not be a defining oped by reflection upon what must be the characteristic of the world. case for science to be possible; and this is in- dependent of any actual scientific knowledge. ⇓ A strong argument in favor of the Moreover, it is not true, even from the point epistemic fallacy is: what can be known of view of the immanent logic of a science, to exist is part of what we can know, and that what we can know to exist is just a part therefore must be subsumed to epistemol- of what we can know. For a law may exist ogy. Let’s see how Bhaskar refutes this: and be known to exist without our knowing 39:1 The view that statements about be- the law. Much scientific research has in fact ing can be reduced to or analysed in terms the same logical character as detection. In of statements about knowledge might be de- a piece of criminal detection, the detective fended in the following way: ontology is de- knows that a crime has been committed and pendent upon epistemology since what we some facts about it but he does not know, or can know to exist is merely a part of what at least cannot yet prove, the identity of the we can know.21 But this defence trades upon criminal. a tacit conflation of philosophical and scien- tific ontologies. For if ‘what we can know to This denies second-order conclusions, 38 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM that we know things to exit which we do world be like so that production is possi- not know themselves. Like a detective ble? who knows that murder has been com- ⇓ The thesis that to be is not to be the mitted, but does not know who did it. value of a variable is another implication In all this discussion of the epistemic of the epistemic fallacy: fallacy I am missing the following argu- 39:2 To be is not to be the value of a vari- 22 ment which seems convincing and simple able; though it is plausible (if, I would ar- to me: the epistemic fallacy cannot be gue, incorrect) to suppose that things can only be known as such. For if to be were true because knowledge is not the only just to be the value of a variable we could channel which gives us information about never make sense of the complex processes the world. We also learn about the world of identification and measurement by means by our practical activity in the world. of which we can sometimes represent some We are able to act in the world without things as such. Knowledge follows existence, knowing why we can act in this way. For in logic and in time; and any philosophical Marx, the most fundamental human ac- position which explicitly or implicitly denies tivity is production. Bhaskar is asking: this has got things upside down. what must the world be like that science I guess Marx was also arguing against is possible? I think it would have been the thesis that “to be is to be the value clearer if he had asked: what must the of a variable” with his emphasis of the 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY39 qualitative as opposed to the quantitative epistemic fallacy thus covers or <30> dis- aspects of things: one should not confuse guises an ontology based on the category of the quantitative dimension of something experience, and a realism based on the pre- with the thing itself. sumed characteristics of the objects of expe- ⇓ Epistemic fallacy denies the possi- riences, viz. atomistic events, and their rela- tions, viz. constant conjunctions. (Such pre- bility of philosophical ontology, therefore sumptions can, I think, only be explained in gives implicit ontology. The false implicit terms of the needs of a justificationist epis- ontology gives a methodology irrelevant temology, e.g. for incorrigible foundations of to science. That is why philosophy came knowledge.) This in turn leads to the gener- into such disrepute! ation of a methodology which is either con- 39:3/o The metaphysical mistake the sistent with epistemology but of no relevance analysis of experimental episodes pinpoints, to science; or relevant to science but more or viz. the epistemic fallacy, involves the denial less radically inconsistent with epistemology. of the possibility of a philosophical ontol- So that, in short, philosophy itself is ‘out of ogy. But if transcendental realism is correct, phase’ with science. Let us see how this hap- and ontology cannot in fact be reduced to pens. epistemology, then denying the possibility of an ontology merely results in the generation of an implicit ontology and an implicit re- ⇓ Next paragraph shows the arguments alism. In the empirical realist tradition the of a philosophy which is out of phase with 40 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM science. this way an implicit ontology, crystallized in the concept of the empirical world, is gen- 40:1 First, the general line of Hume’s cri- erated. And it is this ontology which sub- tique of the possibility of any philosophical sequent philosophers of science have uncrit- ontology or account of being, and in par- ically taken over. For whether they have ticular his denial that we can philosophi- agreed with Hume’s epistemology or not, cally establish the independent existence of they have accepted his critique of ontology, things or operation of natural necessities, is which contains its own implicit ontology, as accepted. Now it is important to see what valid. Hume has in fact done. He has not really succeeded in banishing ontology from his ac- (1) it accepts Hume’s denial that the count of science. Rather he has replaced independence of things can be established the Lockean ontology of real essences, pow- philosophically. Hume did not get rid of ers and atomic constitutions with his own ontology but replaced it with his ontology ontology of impressions. To say that ev- of impressions. (2) this implicit ontology ery account of science, or every philosophy is uncritially taken over by others, even in as much as it is concerned with ‘science’, those who do not agree with Hume’s epis- presupposes an ontology is to say that the temology. philosophy of science abhors an ontological vacuum. The empiricist fills the vacuum he ⇓ Next paragraph describes the gener- creates with his concept of experience. In ation of this implicit ontology in more de- 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY41 tail: instead of asking whether constant in fact occur. That is, philosophy’s concern conjunctions of events do in fact occur, is with whether our knowledge of the world philosophy only asks whether our knowl- can be reduced to sense-experience as so con- edge of the world exhausts itself in empir- ceived or whether it must include an a priori ical knowledge, or whether there is also or theoretical component as well; not with an a priori or theoretical component. whether experience can adequately consti- tute the world. 40:2/o Let us examine the generation ⇓ Humean empiricism not only reduces of this implicit ontology in greater de- knowledge to sense experience but also tail. In Hume’s positive analysis of percep- reduces the world to empirical phenom- tion and causality experiences constituting ena. Through this double reduction it atomistic events and their conjunctions are seen as exhausting our knowledge of nature. can get an isomorphism of the two. The Now, adopting a realist meta-perspective result has been a continuing ontological this means that such events and their con- tension. junctions must occur in nature, if science is 41:1 But in Humean empiricism two to be possible. But from Hume onwards the things are done. First, knowledge is re- sole question in the philosophy of science is duced to that of atomistic events appre- whether our knowledge is exhausted by our hended in sense-experience. Secondly, these knowledge of such events and their conjunc- events are then identified as the particulars tions; it is never questioned whether they of the world. In this way our knowledge of 42 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM reality is literally identified, or at best taken world to the subjective necessity of causally- to be in isomorphic correspondence, with the determined or the inter-subjective necessity reality known by science. From Hume on- of rule-governed minds. wards philosophers have thus allowed, for the ⇑ This is another remark relevant for sake of avoiding ontology, a particular con- induction. cept of our knowledge of reality, which they ⇓ Next paragraph gives examples of may <31> wish to explicitly reject, to in- form and implicitly define their concept of this ontological tension: Hume’s realist the reality known by science. The result methodology does not fit together with has been a continuing ‘ontological tension’ his own epistemology. Similar incon- induced by the conflict between the ratio- gruity in Kant. nal intuitions of philosophers about science 41:2/o Now if transcendental realism is and the constraints imposed upon their de- true, and scientists act as if the objects of velopment by their inherited ontology. This their investigation are intransitive and struc- has led to a nexus of interminably insoluble tured, then any adequate methodology must problems, such as how we can reason from be consistent with the realist practice of sci- one experience to another, and to a displace- ence, and so inconsistent with the epistemol- ment of these rational intuitions whereby, ogy of empirical realism. It is instructive to for example, the locus of necessity is shifted look at Hume here. One finds in the Trea- from the objective necessity of the natural tise an eminently sensible realist methodol- ogy in almost total dislocation from, and cer- 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY43 tainly lacking any foundation in, his radical tific criticism: epistemology. Thus one might be forgiven for wondering what has become of his phe- 42:1/o It might be argued in defence of nomenalism and the doctrine of impressions Hume that he is concerned to show that our when Hume allows that the ‘understanding realist intuitions cannot be justified; that corrects the appearances of the senses’.23 Or his point is precisely that there is a dis- what has happened to the idea of the contin- location between what can be shown and gency of the causal connection and the prob- what must be believed (that ‘there is a di- lem of induction when he argues that scien- rect and total opposition twixt our reason tists, when faced with exceptions to estab- and our senses’);25 and that he leaves the lished generalizations, quite properly search latter intact. But the matter is not so sim- for the ‘secret operation of contrary causes’ ple as this. Humean empiricism is not neu- rather than postulate an upset in the uni- tral in its consequences for scientific prac- formity of nature.24 This is typical. There tice. Taken consistently, it does generate a is a similar dislocation between Kant’s Cri- methodology; not indeed Hume’s (or New- tique of Pure Reason and his Metaphysical ton’s), but Mach’s. For in the absence of the Foundations of Natural Science. concept of an ontological realm, the implicit realism generated implies that whatever is ⇓ Consistent application of Hume’s experienced in sense-experience is an event epistemology is Mach’s methodology. and whatever constant conjunctions are ex- “Squeezes out” possibility of any scien- perienced are causal laws. In this way our 44 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM current knowledge fills the vacuum left by ⇓ Realism is therefore less dogmatic the dissolution of the ontological realm; and than empiricism. In order to be a falli- in so doing it squeezes out, metaphorically bilist about knowledge it is necessary to speaking, the possibility of any substantive be a realist about things: scientific criticism. In the methodology of Humean empiricism facts, which are social 43:1 It is thus quite incorrect to suppose products, usurp the place of the particulars that realist as opposed to non-realist in- of the world; and their conjunctions, which terpretations of scientific theory have con- are doubly social products (once qua fact, sequences for science which are in practice once qua event-conjunction), the place of more dogmatic;27 or to suppose that the con- causal laws. The result is the generation of a cept of natural necessity is a kind of sur- conservative ideology which serves to ratio- vival from the bad old days of scientific cer- nalize the practice of what <32> Kuhn has tainty.28 On the contrary, the converse is called ‘normal science’.26 Descriptivist, in- the case. For it is only if the working sci- strumentalist and fictionalist interpretations entist possesses the concept of an ontolog- of theory do not do away with e.g. scien- ical realm, distinct from his current claims tific laws, but by reducing their ontological to knowledge of it, that he can philosophi- import to a given self-certifying experience, cally think out the possibility of a rational they serve to exempt our current claims to criticism of these claims. To be a fallibilist knowledge of them from criticism. about knowledge, it is necessary to be a re- alist about things. Conversely, to be a scep- 1.4. STATUS OF ONTOLOGY AND ITS DISSOLUTION IN CLASSICAL PHILOSOPHY45 tic about things is to be a dogmatist about cal acceptance of it, that accounts for the on- knowledge. tological tension within philosophy and the dislocation of epistemology from methodol- Question 24 Bhaskar says that his ogy, of philosophy from science. It must be depth realism is less dogmatic than em- accounted for in part by the conditions of piricism. How does he make this point? science, as well as philosophy. For the pe- riod in which Humean ontology became em- ⇓ Also the condition of science between bedded in philosophy (1750–1900) was, at 1750 and 1900 is responsible for ontolog- least in physics, a period of scientific con- ical tension. This was not a period of solidation rather than change. The role of scientific change but of scientific consoli- philosophy was seen more and more to be dation, and philosophy became more and that of showing how our knowledge is jus- tified as distinct from showing how it was more concerned with the question how produced, can be criticized and may come knowledge can be justified, instead of to be changed. Thus whereas transcendental asking how knowledge can be produced. realism asks explicitly what the world must But these attempts to justify knowledge be like for science to be possible, classical failed; the critical realist on the other philosophy asked merely what science would hand can criticize scientific activity. have to be like for the knowledge it yielded to 43:2/o Now it is not only the doctrine of be justified. It was presumed that our knowl- empirical realism, and philosophers’ uncriti- 46 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM edge was justified; science was not viewed as of battleground for internecine warfare be- a process in motion; and doing away with tween opposed concepts of justified belief. ontology left philosophy without any criti- Moreover, when the idea of scientific cer- cal purchase on science. The transcendetal tainty eventually collapsed, the absence of an realist, on the other hand, allows a limited ontological dimension discouraged anything critical role for philosophy. For by restor- other than a purely voluntaristic reaction— ing the idea of an ontological realm distinct in which it was supposed that because our from science, he makes it possible for us to beliefs about the world were not causally say that in a particular field, say social psy- determined by the <33> world then they chology, science is not being done, although must be completely ‘free creations of our own as a philosopher he cannot say dogmatically minds, the result of an almost poetic intu- whether or not a science of social psychology ition’.30 29 is possible. (An ontological dimension is in ⇓ (1) There is also a strong anthro- this way necessary not only to render intelli- pocentric current, and (2) the dissolution gible scientific criticism, but to make possi- ble philosophical critcism of the practice of of ontology made philosophy irrelevant. a science.) Increasingly then it was the log- 44:1/o Behind this state of affairs there ical structure of justificatory argument that ran a strong anthropocentric current in clas- defined philosophy’s concept of science; and sical and subsequent philosophy,31 which the philosophy of science itself became a kind sought to rephrase questions about the world as questions about the nature or behaviour 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS47 of men. One aspect of this is the view, which turned inside out in an attempt to confine it I have characterized as the epistemic fal- within sentience. An inevitable ‘involution’ lacy, that ontological questions can always in the philosophy of science occurred. With- be rephrased as epistemological ones. The out a concept of a reality unknown, but at anthropocentric and epistemic biases of clas- least in part knowable, philosophy could not sical philosophy led to the dissolution of the display the creative and critical activity of concept of the ontological realm, which we science, and ceased to be of any practical need to render intelligible the transitive pro- relevance for it. This was the price paid for cess of science. In this way the world, which the dissolution of ontology. A philosophy for ought to be viewed as a multi-dimensional science depends upon its reconstitution. structure independent of man, came to be squashed into a flat surface whose charac- teristics, such as being constituted by atom- 1.5 Ontology Vindi- istic facts, were determined by the needs of a particular concept of knowledge. This led cated and the Real to a barrage of problems and an impossible Basis of Causal account of science. For from now on any structure, if it was allowed at all, had to Laws be located in the human mind or the scien- tific community. Thus the world was literally After arguing at length that there must be a difference between causal laws and 48 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM constant conjunctions of events, Bhaskar Bhaskar will argue that the ontology of sets out to explain now what causal laws structures, generative mechanisms, and are. active things is really a much less mys- ⇓ The first paragraph uses the word terious explanation of it all. “mysterious” like Marx in the fetish-like ⇓ Both transcendental idealists and character of the commodity. We know transcendental realists would agree that what causal laws are not: they are not some generative mechanism is at work sequences of events. But what are they? to distinguish necessary from accidental 45:1 In Section 3 I argued that only if sequences. But the idealist locates this causal laws are not the patterns of events generative mechanism only in the mind, that enable us to identify them can the in- while the realist says that the theory con- telligibility of experimental activity be sus- struction and change can only be ex- tained. But causal laws are, or have seemed plained if those generative mechanisms to philosophers to be, pretty mysterious en- tities. What can it mean to say that they are real. have a real basis independent of events? 45:2/o The ontological status of causal The answer to this question will be seen laws can best be approached by considering to necessitate the development of a non- the divergent responses of transcendental re- anthropocentric ontology of structures, gen- alism and idealism to the problem of distin- erative mechanisms and active things. guishing a necessary from a purely accidental 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS49 sequence of events. Both may agree, in their change. modern versions, that without some concep- ⇓ If it is granted that these genera- tion of a generative mechanism at work no tive mechanisms are real, then the realist attribution of necessity is justified. For the transcendental idealist, however, this neces- can explain theory construction by exper- sity is imposed by men on the pattern of imental activity: events; the generative mechanism is an ir- 46:1 Now once it is granted that mecha- reducible figment of the imagination. For nisms and structures may be said to be real, the transcendental realist, on the other hand, we can provide an interpretation of the in- the generative mechanism may come to be dependence of causal laws from the patterns established as real in the course of the on- of events, and a fortiori of the rationale of going activity of science. Indeed he will ar- experimental activity. For the real basis of gue that it is only <34> if existential ques- this independence lies in the independence tions can be raised about the objects of sci- of the generative mechanisms of nature from entific theory that the rationality of theory the events they generate. Such mechanisms construction can be sustained. For without endure even when not acting; and act in their them science would remain, as in empiricism, normal way even when the consequents of a purely internal process—with the familiar- the law-like statements they ground are, ow- ity of image replacing the reinforcement of ing to the operation of intervening mecha- sensation, still lacking a rational dynamic of nisms or countervailing causes, unrealized. It is the role of the experimental scientist to 50 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM exclude such interventions which are usual; ⇓ Is the following paragraph now the and to trigger the mechanism so that it is epistemological conclusion from this on- active. The activity of the mechanism may tology? Or is it a summary of what he is then be studied without interference. And going to prove? it is this characteristic pattern of activity or mode of operation that is described in the 46:2/o Only if causal laws persist through, statement of a causal law. It is only un- which means they must be irreducible to, der closed conditions that there will be a the flux of conditions can the idea of the one-to-one relationship between the causal universality of a known law be sustained. law and the sequence of events. And it is And only if they have a reality distinct from normally only in the laboratory that these that of events can the assumption of a natu- enduring mechanisms of nature, whose op- ral necessity be justified. On this view laws erations are described in the statements of are not empirical statements, but statements causal laws, become actually manifest and about the forms of activity characteristic of empirically accessible to men. But because the things of the world. And their neces- they endure and continue to act, when stimu- sity is that of a natural connection, not that lated, in their normal way outside those con- of a human rule. There is a distinction be- ditions, their use to explain phenomena and tween the real structures and mechanisms of resistence to pseudo-falsification in open sys- the world and the actual patterns of events tems can be rationally justified. that they generate. And this distinction in turn justifies the more familiar one between 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS51 necessary and accidental sequences. For a ate the flux of phenomena that constitute the necessary sequence is simply one which cor- actual states and happenings of the world. responds to, or is in phase with, a real con- They may be said to be real, though it is nection; that is, it is a real connection actu- rarely that they are actually manifest and ally manifest in the sequence of events that rarer still that they are empirically identi- occurs. fied by men. They are the intransitive ob- jects of scientific theory. They are quite inde- ⇓ The world consists of mechanisms, pendent of men—as thinkers, causal agents not events. These mechanism combine to and perceivers. They are not unknowable, produce the flux of phenomena. In other although knowledge of them depends upon a words: the empirical realist says: “be- rare blending of <35> intellectual, practico- hind phenomena are atomistic events,” technical and perceptual skills. They are not while the realist says: “behind phenom- artificial constructs. But neither are they ena are enduring mechanisms.” These Platonic forms. For they can become mani- mechanisms are not unknowable, but fest to men in experience. Thus we are not the process of gaining knowledge about imprisoned in caves, either of our own or of them is arduous. It requires intellecual, nature’s making. We are not doomed to ig- norance. But neither are we spontaneously practico-technical, and perceptual skills. free. This is the arduous task of science: 47:1 The world consists of mechanisms not the production of the knowledge of those en- events. Such mechanisms combine to gener- 52 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM during and continually active mechanisms of sense. nature that produce the phenomena of our 47:3 Thus, it might be objected that the world. very idea of a world without men is unintel- ⇓ Those objecting to this reconstitu- ligible because the conditions under which it tion of ontology question first the intran- is true would make its being conceived im- sitivity and then the structured character possible. But I can think of a world without men; and I can think of a world without my- of the postulated objects of scientific in- self. No-one can truly say ‘I do not exist’ quiry: but that does not mean that ‘I do not exist’ 47:2 Objections may be made to my pro- is unintelligible; or that it cannot be mean- posed reconstitution of an ontological realm, ingfully, just because it cannot be truly said. which question in turn the intransitivity and It is no objection to the intelligibility of a the structured character of the postulated statement that it is counter-factual. Indeed objects of scientific inquiry, i.e. the ideas of it is only because it is intelligible that we can their categorical independence from men and say that it is counter-factual. events respectively. I will consider the two ⇓ Second objection: it is impossible to kinds of objections in turn. think of a world without men. Some- ⇓ First objection: the idea of a world how the process of thinking is insepara- without men is unintelligible. Bhaskar ble from men, Bhaskar assumes the ar- says it is counterfactual, but it does make gument is: we cannot think anything 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS53 without thinking about us thinking it or thinking the thought. Clearly if this were about ourselves. Bhaskar tries to refute so, an infinite regress would be impossible to this by distinguishing betwen being aware avoid. However, to be aware of the fact that of thinking, thinking about thinking, and I am thinking of a particular topic x, it is not thinking about the thinker. I guess he necessary for me to be thinking of that fact. means: every human activity is charac- Such awareness may be expressed in thought, but when it is the topic is no longer x but terized by awareness of what we are do- my thought of x. It is possible for A to think ing, but this does not mean we are think- epsilon and to be aware of thinking epsilon ing about what we are doing or thinking without thinking about thinking epsilon; and about ourselves. unless this were so no-one could ever intelli- 47:4/o Someone might hold that to think gently think. Moreover it is possible for A to of a world without men is not so much un- think about thinking epsilon without think- intelligible as impossible; that we must pic- ing about his (A’s) thinking epsilon. Think- ture ourselves in any picture. Now it is a ing about thinking about a particular topic fact about human beings that we can do must be distinguished from thinking about this. But we do not have to do it, any more the thinker of the topic.32 than an artist must initial his work. The idea may be perhaps that a thought must ⇓ After having said that there is no always contain, or at least be accompanied absurdity in the supposition of a world by, a thought of the thinker of the thought without men, he says now how important 54 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM it is to ask what a world would be like in would still act, be subject to laws and pre- which there are no men. “For example, serve their identity through certain changes. things would still act, be subject to laws, ⇓ Second objection says: there is a dif- and preserve their identity throug certain ference between laws and constant con- changes.” junctions, however from this does not fol- 48:1 There is no absurdity in the supposi- low that laws are produced by real gen- tion of a world without men. Rather it is a erative mechanisms. possibility presupposed by the social activity 48:2/o A second kind of objection might of science. It is important to establish this focus on the structured character of the pos- fact. For we are too liable to underestimate tulated objects of scientific inquiry, question- the power of the pictures, often unconscious, ing not so much the idea itself but the inter- which underpin philosophical theories. Such pretation I have given to it; and in particu- pictures indeed often hold our philosophical lar the explanatory value of the particular imagination ‘captive’.33 Our philosophy of ontology proposed. Thus it might be ob- science is heavily anthropocentric, which is jected that, while the transcendental argu- why it is important to consider what it would ment from experimental activity in Section be possible to say about our world if there 3 establishing the distinctiveness of causal were no men, given that we know that our laws and patterns of events, is sound, the world is one in which science is as a <36> introduction of the concept of generative matter of fact possible. For example things mechanisms to provide a real basis for causal 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS55 laws is gratuitous. determine the manifest outcome of their ac- ⇓ Expands what a generative mecha- tivity. That is to say, it entails that gen- nism is: it continues to act in its normal erative mechanisms endure even when inac- tive and act even where, as in open systems, way even if the results of its operations there is no one-to-one relationship between are modified. And it matters even less for the causal law representing the characteris- this mechanism whether things are per- tic mode of operation of the mechanism and ceived by man or not. the particular sequence of events that occurs. 49:1 What does it mean to say that a gen- In particular, it entails that mechanisms act erative mechanism endures and acts in its in their normal way outside the closed condi- characteristic way? It does not mean, we tions that enable us to experimentally iden- have seen, that a regular sequence of events tify them and whether or not we do so; i.e. occurs or is experienced; though the occur- whether or not the results of their operations rence of such a sequence may, in special cir- are modified, and whether or not these re- cumstances, provide empirical grounds for sults are perceived by men. (In the former the hypothesis of the existence of the mech- case we could talk of a disjuncture between anism. For the intelligibility of experimen- the domains of the real and the actual; in tal activity entails that the particular mech- the latter case of a disjuncture between the anism endures and at least some mecha- domains of the real and the empirical.) nisms act through the flux of conditions that determine whether they are active and co- ⇓ Now he goes from generative mecha- 56 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM nisms to active things. outside the conditions that enable us to em- pirically identify them. Now accepting this 49:2/o Now the reason why the concept non-empirical interpretation means that ref- of a causal law cannot itself be taken as on- erence to causal laws involves centrally ref- tologically basic is because its analysis pre- erence to causal agents; that is, to things supposes a ‘real something’ over and above endowed with causal powers. and independent of patterns of events; and it is for the status of this real something From the fact that generative mecha- that the concept of a generative mechanism nisms persist even if they do not have is groomed. But then does to say that a gen- actual results he concludes that genera- erative mechanism endures and acts in its tive mechanisms must be the powers of characteristic way mean anything more than things. Otherwise they could not be said to say that a thing goes on acting in a certain to persist without making an illicit reifi- way? As stated the reformulation is ambigu- cation of the generative mechanism. I ous. For the continuance of a form or pattern take it, this means, it cannot just be a se- of activity can be interpreted in an empiri- cal or a non-empirical way. The intelligibil- quence of actions and reactions, because ity of experimental activity requires the lat- if it were this, then it would not start up ter non-empirical interpretation. For it en- again after it dies down. Since it per- tails, as we have seen, that causal laws per- sists through periods in which it does not sist and are efficacious in open systems, i.e. manifest itself, even without needing a 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS57 new trigger (it is a tendency, not just a tive mechanisms of nature <37> exist as power), this indicates that it is anchored the causal powers of things. We now have in something that persists, i.e., in a thing. a perfectly acceptable ontological basis for It is central to this argument that laws causal laws. For if it is wrong to reify are not merely the power of things, but causal laws, and it is wrong to reify gen- it is the tendency of things. erative mechanisms, it cannot be wrong to reify things! However, the fact that the tran- Here is my definition of tendencies, li- scendental analysis of experimental activity abilities, powers: powers is something showed that generative mechanisms must go which it can do, liabilities are things on acting (i.e. that causal laws must be ef- which can be done to it, things to which ficacious) outside the closed conditions that it is susceptible. The human body is sus- permit their identification means that causal ceptible to cancer or to virus infections. laws cannot be simply analysed as powers. Tendencies are things which it has to do. Rather they must be analysed as tendencies. Even if it is barred from doing it for a For whereas powers are potentialities which may or may not be exercised, tendencies are while, afterwards it will spontaneously potentialities which may be exercised or as it start up doing it again, and perhaps one were ‘in play’ without being realized or man- can say it has been trying to do it all ifest in any particular outcome. They are along. therefore just right for the analysis of causal On this interpretation then the genera- 58 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM laws.34 ized unperceived by men). It is by reference ⇓ Laws are not enduring powers, not just to the enduring powers but the un- i.e., they endure between experiments in realized activities or unmanifest (or incom- pletely manifest) actions of things that the which they are triggered, but laws refer phenomena of the world are explained. It to the continuing activity of things. This is the idea of continuing activity as distinct is why experiments are relevant for non- from that of enduring power that the concept experimental situations. of tendency is designed to capture. In the 50:1 If the analysis of causal laws (and concept of tendency, the concept of power is generative mechanisms) is to be given by the thus literally dynamized or set in motion. concept of things and not events (a possibil- ⇓ Therfore not subjunctive, but normic ity which I have already rejected by demon- statements. Not statements: “if this then strating in Section 3 their categorical in- dependence from events), the consideration that,” but statements: “normally so and that they not only persist but are efficacious so,” not counterfactual but transfactual. in open systems, which is presupposed by the 50:2/o In the full analysis of law-like state- intelligibility of experimental activity, entails ments we are thus concerned with a new that causal laws must be analysed as tenden- kind of conditional: which specifies the exer- cies. For tendencies are powers which may cise of possibilities which need not be mani- be exercised without being fulfilled or actu- fest in any particular outcome. Such condi- alized (as well as being fulfilled or actual- tionals are normic,35 rather than subjunc- 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS59 tive. They do not say what would hap- ties, and tendencies. The scientist is not pen, but what is happening in a perhaps saying what is going to happen, but what unmanifest way. Whereas a powers state- kind of things there are, and what they ment says A would ψ in appropriate circum- tend to do. stances, a normic statement says that A re- ally is ψ’ing, whether or not its actual (or 51:1 The world consists of things, not perceivable) effects are counteracted. They events. Most things are complex objects, in are not counter-factuals, but transfactuals; virtue of which they possess an ensemble of they take us to a level at which things are tendencies, liabilities and powers. It is by really going on irrespective of the actual out- reference to the exercise of their tendencies, come. To invoke a causal law is to invoke a liabilities and powers that the phenomena normic conditional. A full analysis of normic of the world are explained. Such continu- and tendency statements will be provided ing <38> activity is in turn referred back for later. For the moment, it should be noted explanation to the essential nature of things. that normic statements provide the correct On this conception of science it is concerned analysis of the nomic indicative form. A essentially with what kinds of things they nomic statement is a transfactual statement, are and with what they tend to do; it is only with actual instances in the laboratory that derivatively concerned with predicting what constitute its empirical grounds. is actually going to happen. It is only rarely, and normally under conditions which are ar- ⇓ Then he talks about powers, liabili- tificially produced and controlled, that scien- 60 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM tists can do the latter. And, when they do, ing of a thing. It endures, and under ap- its significance lies precisely in the light that propriate circumstances is exercised, as long it casts on the enduring natures and ways of as the properties that account for it per- acting of independently existing and trans- sist. Laws then are neither empirical state- factually active things. ments (statements about experiences) nor statements about events. Rather they are Question 25 According to transcenden- statements about the ways of acting of inde- tal realism, explanation and prediction pendently existing and transfactually active are not the same thing. What is the role things. of prediction in science? Compare that with his earlier state- ment that causal laws are held to be mys- ⇓ There is nothing mysterious about terious. I guess what he means here is: he the concept of a generative mechanism, is coming up with a very simple explana- they are simply the ways of acting of tion of something that otherwise can only things. be explained by a “mystery.” 51:2/o There is nothing esoteric or mys- terious about the concept of the generative What Bhaskar accords with Collier’s mechanisms of nature, which provide the real definition of generative mechanisms as basis of causal laws. For a generative mech- those aspects of a thing which cause it anism is nothing other than a way of act- to act in a certain way. 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS61

⇓ After his negative interpretation in quiry of metaphysics. Rather, it is the end to 35:2/o, he now can give a positive inter- which all the empirical efforts of science are pretation of what structured and intran- directed. Ontology has been vindicated not sitive means: there are active structures as providing a set of necessary truths about which cause the regularities. a mysterious underlying physical realm, but as providing a set of conditionally necessary 52:1/o It is now possible to give a posi- truths about our ordinary world as inves- tive interpretation of our characterization in tigated by science. It is important to be Section 3 of the objects of scientific investiga- clear about what philosophical argument can tion, at least in so far as they are causal laws, achieve. Thus as a piece of philosophy we as ‘structured intransitive’. ‘Structured’ in can say (given that science occurs) that some so far as it is the activities of mechanisms real things and generative mechanisms must and causal structures, not the occurrence of exist (and act). But philosophical argument events, that are designated in statements of cannot establish which ones actually do; or, causal law. ‘Intransitive’ in so far as the to put it the other way round, what the real mechanisms and causal structures, whose ac- mechanisms are. That is up to science to dis- tivity is designated, endure and act quite cover. That generative mechanisms must ex- independently of men. To discover the in- ist and sometimes act independently of men dependently existing and transfactually ac- and that they must be irreducible to the pat- tive machinery of nature is not, it should terns of events they generate is presupposed be stressed, the aim of an independent in- 62 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM by the intelligibility of experimental activ- do: it can give second-order arguemnts, ity. But is up to actual experiments to tell but then he says somethign else: the sci- us what the mechanisms of nature are. Here, entists in their conclusions use concepts as elsewhere, it is the task of philosophy to like thing, law, existence, etc., and it analyse notions which in their substantive is a task of the philosopher to analyze employment have only a syncategorematic the concept as such. This is legitimate use. Thus whenever a scientist refers to a and necessary even though “there are no thing or event, structure or law, or says that something exists or acts in a certain way he things in general.” I guess, it is an ab- must refer to it under some particular de- straction, and the philosopher character- scription; he is using the notion of thing, izes this abstraction. law, existence, etc. But it is the task of the ⇓ “Tidies up” the analysis of experi- philosopher to analyse the concept as such. mental activity. To argue that this task is both <39> legiti- mate and necessary is not to populate the 53:1 I am now in a position to tidy up world with (or to suppose that there is a my analysis of experimental activity. The world of) things without names or events- experimental scientist must perform two es- in-general. sential functions in an experiment. First, he must trigger the mechanism under study Greater part of this paragraph is a to ensure that it is active; and secondly, he characterization of what philosophy can must prevent any interference with the op- 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS63 eration of the mechanism. These activities conclusion before, he now gives a positive could be designated ‘experimental produc- and more precise interpretation. Experi- tion’ and ‘experimental control’. The for- ment consists of two things: experimental mer is necessary to ensure the satisfaction of production (setting the antecedent, the the antecedent (or stimulus) conditions, the stimulus in motion), und experimental latter to ensure the realization of the conse- control (to make sure that the consequent quent, i.e. that a closure has been obtained. can be realized, is not thwarted). Here But both involve changing or being prepared to change the ‘course of nature’, i.e. the se- he uses the word “realization” where per- quence of events that would otherwise have haps actualization would be better. But occurred.36 In a simple electrical experiment one should even talk about empiricaliza- designed to illustrate say Ohm’s Law, the tion: it must not only occur but also be wiring of an electric circuit and the genera- observed. tion of an electric current would constitute But another question: if laws are ten- ‘experimental production’; maintaining the dencies of things, why do they need stim- appropriate resistance levels, ensuring that uli? no new magnetic field is suddenly placed in the neighbourhood of the circuit, etc. would Footnote: experimental production φ then constitute ‘experimental control’. changes the state of the world α to a, which is the stimulus which tends to cre- I guess again, instead of a “can only” ate b. And experimental control (there is 64 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM a typo in the third line of the footnote, scientists in general record a unique relation- he says “production” where he should say ship between the antecedent and consequent “control”) ψ excludes the intervention of of a law-like statement. The aim of an ex- other effects, so that a can generate b, so periment is to get a single mechanism going that the sequence a → b is a constant in isolation and record its effects. Outside a conjunction of events. But it is a result closed system these will normally be affected by the operations of other mechanisms, ei- of human action. ther of the same or of different kinds, too, so I think it is not only so in experimental that no unique relationship between the vari- control, but also in building machines, in ables or precise description of the mode of electric wiring: insulating the cables is ψ. operation of the mechanism will be possible. ⇓ says the same thing again, and adds In general, experimental activity requires a that this requires a degree of plasticity of degree of plasticity of the antecedent (stimu- the antecedent and ambient conditions to lus) and circumambient conditions to human manipulation and control. Such plasticity is human manipulation, which is not easily not easily won. ‘Experimental design’ is a won. By “‘plasticity” he means that hu- substantial theoretical labour in itself. man agency must be able to affect those things. ⇓ Again several statements: 53:2 Only if the mechanism is active and (1) experiment is not just putting a the system in which it operates is closed can question to nature, it is a practical ques- 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS65 tion, we put it with our hands so to speaking, that in an experiment we put a speak. The last sentence here is what question to nature. But it has not been said this whole paragraph is all about: the on- that the question we put is a practical one— tological distinction between causal laws with our hands, so to speak. The weakness and patterns of events depends on only of previous analyses of experimental activity two premises: (a) men can interfere with is that they have not appreciated the signifi- cance of the fact that conjunctions of phe- nature, and (b) this “disruption” of na- nomena have to be worked for practically ture is significant for learning about na- (as well as in thought); that conjunctions ture. are not given to, but made by us. In an (2) he calls it somewhat a scandal that important study, von Wright has seen this. empiricists “should not have untertaken But he has not drawn the correct conclusion an analysis of the conditions under which from it: which is that, just because the ex- experience is significant in science.” (By perimenter is a causal agent of the sequence this he apparently does not mean: which of events, there must be an ontological dis- tinction between the sequence he generates experiences are significant, but the ques- and the causal law it enables him to identify. tion: in which kind of world are experi- Any other conclusion renders experimental ences significant in the way they are in activity pointless. (Why generate that se- science.) quence?) The reason for von Wright’s failure 54:1 It has often been said, metaphorically 66 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM to see this stems from his unfortunate initial perimental activity, the planned disruption assumption of (as he puts it) a ‘Tractatus- of the course of nature, is a significant fea- world’, i.e. a world of logically independent ture of science. atomistic states of <40> affairs (which as- ⇓ The question which an experiment tonishingly he seems to regard as a harmless puts to nature is not only a practical simplification);37 which precludes him from seeing laws as anything other than condi- question, but it also needs sophisticated tional statements about atomistic states of measurement instruments. affairs. It is of course something of a scan- 54:2/o In stressing the practical compo- dal that empiricists who invoke experience nent of experimental activity, it is important as the sole ground of knowledge and scien- not to forget the theoretical side. In an ex- tific knowledge as their paradigm should not periment men put a question to nature. But have undertaken an analysis of the condi- they must put it in a language that nature tions under which experience is significant understands, as well as in a form that makes in science. It should be stressed that the re- possible an unambiguous reply. It is difficult sult that there is an ontological distinction to overestimate the importance for modern between causal laws and patterns of events science of the development of instruments depends upon only two premises: (i) that such as clocks and telescopes, which may be men are causal agents capable of interfering seen as devices designed to decipher the vo- with the course of nature and (ii) that ex- cabulary of nature. Both the construction and the interpretation of such instruments 1.5. ONTOLOGY VINDICATED AND THE REAL BASIS OF CAUSAL LAWS67 depended upon theory. Hooke’s law, for ex- constructed and its results interpreted. ample, is literally built into the construction ⇓ Two furthergoing questions: (1) 38 of spring balances. Experimental confirma- role of thought to understand the world tion of Galilean dynamics was delayed for a (Einstein’s Gedankenexperimente). Can long time by the difficulty of measuring ‘the that be understood without reducing the most fundamental magnitude of dynamics’, i.e. time. But when the Huyghens eventually world to a set of axioms? (2) What to succeeded in building such a clock in 1659 it do in sciences which do not allow experi- was only by basing it on the new dynamics ments, i.e., are there surrogates of exper- (the very dynamics it was designed to vin- iments. dicate) and in particular the theory of the 55:1/o Two problems are raised by my isochronous curve of the pendulum.39 Sim- analysis of experimental activity. First, we ilarly it has been convincingly argued that know that much science, of what might be the development of cosmology in the early called a fundamental kind, has proceeded by 17th century was held up by the absence way of ‘thought’ rather than by actual ex- of an adequate theory of telescopic vision.40 periment. As Dijksterhuis has put it: ‘In In short, experimental activity depends cru- general one has to take stories about ex- cially upon the adequacy of the theories periments by Galileo, as well as his oppo- (sometimes referred to as ‘auxiliary’) accord- nents with some reserve. As a rule they were ing to which the experimental equipment is performed mentally, or they are merely de- scribed as possibilities.’41 It seems that Ein- 68 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM stein too was not averse to the occasional at a later stage in this study. ‘Gedankexperimente’.42 This raises the ques- tion of whether, and if so how, pure thought can anticipate a law? And the problem of how, if it can, we then avoid the rational- 1.6 A sketch of a cri- ist conclusion that provided only our anx- tique of empirical iom base is strong enough we could deduce all the laws of nature without recourse to realism experience. Secondly, we know that in many fields, most notably history and the human 56:1/o I have argued that the causal struc- sciences and in the biological sciences in as- tures and generative mechanisms of nature pects of their work, experimental activity is must exist and act independently of the con- <41> impossible. This raises the question of ditions that allow men access to them, so whether there are, or it is possible to devise that they must be assumed to be structured for them, surrogates of the experimental es- and intransitive, i.e. relatively independent tablishment of closed systems in physics and of the patterns of events and the actions chemistry? And here again there lurks an of men alike. Similarly I have argued that unacceptable rationalist implication. Both events must occur independently of the ex- pose prima facie problems for transcendental periences in which they are apprehended. realism, which I hope to be able to resolve Structures and mechanisms then are real and distinct from the patterns of events that they 1.6. A SKETCH OF A CRITIQUE OF EMPIRICAL REALISM 69 generate; just as events are real and distinct into harmony or phase with one another from from the experiences in which they are ap- being described. prehended. Mechanisms, events and experi- ences thus constitute three overlapping do- 57:1 <42> Now these three levels of real- mains of reality, viz. the domains of the real, ity are not naturally or normally in phase. It the actual and the empirical. This is repre- is the social activity of science which makes sented in Table 1.1 below. them so. Experiences, and the facts they ground, are social products; and the con- 56:2 junctions of events, that, when apprehended The crux of my objection to the doctrine of in experience, provide the empirical grounds empirical realism should now be clear. By for causal laws, are, as we have seen, social constituting an ontology based on the cate- products too. It can thus be seen that un- gory of experience, as expressed in the con- derlying and necessary for the implicit ontol- cept of the empirical world and mediated by ogy of empirical realism is an implicit soci- the ideas of the actuality of the causal laws ology in which facts and their conjunctions and the ubiquity of constant conjunctions, are seen as given by nature or spontaneously three domains of reality are collapsed into (voluntaristically) produced by men. In this one. This prevents the question of the con- chapter I have outlined an answer to the ditions under which experience is in fact sig- question ‘what must the world be like for nificant in science from being posed; and the science to be possible?’. In Chapter 3 I will ways in which these three levels are brought ask ‘what must society be like for science to 70 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM be possible?’; i.e. I shall attempt a transcen- sists, paradigmatically, in the transforma- dental deduction of certain basic sociological tion of given products. Scientific change is categories from an investigation of the condi- an integral feature of science, in which what tions for the possibility of science. The an- is transformed is a part of the formally ac- swer to these two questions will constitute credited stock of scientific knowledge. In a the interwoven themes of this work. It is im- scientific training the object transformed is possible to over-emphasize how closely they not knowledge but man himself. But in both are connected. For once, for example, we re- cases what is transformed is itself already ject the doctrine that there are everywhere in a social product. The peculiar significance nature such things as spontaneously occur- of experimental activity is that man qua ring parallel cases and see rather that in gen- material object (rather than simply thinker eral they have to be assiduously worked for or perceiver) exercises his causal powers to and artificially produced in the social activ- transform the natural world itself, of which ity of science, we are forced to constitute an he is also a part. Now corresponding to the ontology of structures distinct from events. dissolution of ontology in philosophy, there has been a parallel denegation of the social 57:2/o For us, for the moment, it is suf- character of science. In Chapter 3 I will set ficient merely to note that the most im- out to vindicate sociology in an attempt to portant feature of science neglected by the render intelligible scientific change. This will doctrine of empirical realism is that it is enable me to reconstitute a transitive dimen- work; and hard work at that. Work con- 1.6. A SKETCH OF A CRITIQUE OF EMPIRICAL REALISM 71 sion, as complementary to the intransitive just because they are a part of the world they one established here. cannot be used to define it. An experience to be significant in science must normally be 58:1 The concept of the empirical world is the result of a social process of production; anthropocentric. The world is what men can in this sense it is the end, not the begin- experience. But the couple of this concept, ning of a journey. But only transcendental and from a realist meta-perspective neces- realism can explain why scientists are cor- sary to sustain it, is the absence of the con- rect in regarding experience as in the last in- cept of the antecedent social activity neces- stance the test of theory. For it is by means sary to make experience significant in sci- of it that, under conditions which are arti- ence. And this has the objectionable ideo- ficially produced and controlled, skilled men logical consequence (from the point of view can come to have access to those enduring of the practice of science) that whatever men and active structures, normally hidden or currently experience is unquestionably the present to men only in distorted form, that world. Now it is central to the argument generate the actual phenomena of our world. of this study that the concepts ‘empirical’ Empirical realism depends upon a reduction and ‘sense-experience’ belong quite unequiv- of the real to the actual and of the actual to ocally to the social world of science. Experi- the empirical. It thus presupposes the spon- ences are a part, and when set in <43> the taneity of conjunctions and of facts. And in context of the social activity of science an doing so presupposes a closed world and a epistemically critical part, of the world. But 72 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM completed science. an interpretation of man. The relationship is not between a real and an imaginary ob- 58:2/o It is important to stress that I am ject, but between two kinds of real object, not saying that experiences are less real than one of which is very small. The relationship events, or events less real than structures. between electrons and tables has to be un- This is the kind of mistake that is encour- derstood in terms of causal connections, not aged by the way in which Eddington for- correspondence rules. Consequents are not mulated his problem of the relationship be- less real, or the statements describing them tween the familiar and the scientific worlds; less true, in virtue of their being effects; any in which he described the situation as one in more than causes, in virtue of being recon- which there were ‘duplicates’ of every object: dite, must be imaginary. In particular, the two tables, two chairs, two pens, etc.43 Since fact that the properties of everyday objects, then the problem has always seemed to be at what has been picturesquely described as that of saying which object is real. For the the zone of the middle dimensions,46 can be ordinary language instrumentalist the scien- explained in terms of the very small (or the tific object is an artificial construct;44 for the very large) does not render them less real scientistic super-realist the familiar object a than the entities that account for them; any- mere illusion.45 For the transcendental real- more than zinc and sulphuric acid cease to ist however this formulation of the problem is react in a certain way when we explain their bogus. For if there is a relationship between the worlds it is one of natural generation, not 1.6. A SKETCH OF A CRITIQUE OF EMPIRICAL REALISM 73 reaction in terms of their atomic structure. ually being extended, to which process there can be no a priori limits. Thus though it 59:1/o For the transcendental realist laws, may be necessary, to the extent that science though not our knowledge of them, are cate- is always incomplete, that at any moment gorically independent of men—as thinkers, of time some laws are unknowable; it is not causal agents and perceivers. Transcen- necessary that any particular laws are. dental realism can thus accommodate both Locke’s view that there are (or may be) laws 60:1 Locke’s mistake in failing to appreci- which are unknowable;47 and Kneale’s sug- ate the possibility that the ‘sad experience’ gestion that there are (or may be) laws whose of chemists who ‘sometimes in vain, search instances are unperceivable.48 But it allows for the same qualities in one parcel of sul- in addition the possibility of known laws, phur, antimony or vitriol, which they have whose instances are perceivable, but which, found in others’49 might come to be trans- when not instanced in closed systems, re- formed in the course of the development of main unmanifest to men. However, my in- science into a knowledge of the ‘constitu- terpretation of these possibilities is different tion of their insensible parts, from which flow from Locke’s <44> (and Kneale’s). For the those sensible qualities, which serve us to dis- transcendental realist, our knowledge, per- tinguish one from another’50 was not a sci- ceptual skills and causal powers are set in the entific mistake. It did not consist in his fail- context of the ongoing social activity of sci- ure to foresee the development of the theory ence; and in the course of it they are contin- of atomic number and valency or to predict 74 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM

Mendeleyeev’s predictions. His scepticism sis about knowledge—and in particular of over the possibility of a scientific knowledge course about the impossibility of a certain of real essences was a philosophical mistake, kind of knowledge, viz. of real essences. rooted in his theory of ideas. For if all our Locke’s case has a general moral. For with- knowledge is acquired in perception and per- out a concept of science as a process-in- ception constitutes the world, there can be motion and of knowledge as possessing (in no place for an antecedent cause of knowl- the sense indicated in Section 1 above) a edge (or of perception). But as only what material cause, it is easy to argue from the is seen as socially produced can be seen as current state of a science to a philosophical putatively socially transformable, this leads thesis about knowledge. Consider, for exam- inevitably to an a-historical view of science. ple, the Copenhagen interpretation of Quan- tum theory. More important perhaps, the 60:2/o Locke’s error was not therefore influence of Newtonian mechanics on 18th based on an inadequate knowledge of chem- century philosophy led to a kind of stasis in istry. But on an inadequate concept of the thought from which the philosophy of science transitive dimension of science, which pre- has still to recover. Action-by-contact as a vented him from seeing the current state paradigm of causality, the celestial closure of chemistry as what it was, viz. the cur- as a model of knowledge, gravity as the tem- rent state of a science; and which thus plate of our ignorance all had a disastrous ef- allowed him to be influenced by it into fect. The underdevelopment of the sciences propounding a general philosophical the- 1.6. A SKETCH OF A CRITIQUE OF EMPIRICAL REALISM 75 of substance in comparison with the science represent things as if it did. In the philos- of motion (of the time), and the form that ophy of science there must be two Coper- the latter took, thus had, at a decisive mo- nican Revolutions. The first establishing a ment in the history of philosophy, through transitive dimension in which our knowledge the generation of a static philosophical con- is seen to be socially produced, and as such ception of knowledge, a permanent effect on neither an epiphenomenon of nature nor a all subsequent ‘philosophy of science’. It is convention of man. The second establishing in this sense that in philosophy we are still an intransitive dimension, based on the re- prisoners of the scientific thought of the past. constitution of a philosophical ontology, in which the world of which, in the social ac- 61:1/o <45> The anthropocentric and tivity of science, knowledge is obtained is epistemic biases of classical philosophy have seen to be in general quite independent of resulted in the dominance, in philosophy, man. These Copernican Revolutions must of what might be styled ‘idols’ of a Baco- be given a Copernican interpretation (for nian kind. These are false conceptions which philosophy has its Osianders too); which is cause men to see, in philosophy, everything why we need the metaphysics of transcen- in relation to themselves (cf. the concept of dental realism, which will be vindicated by the empirical world) and their present knowl- its capacity to render intelligible the under- edge. Six hundred years ago, Copernicus analysed phenomenon of science. argued that the universe does not revolve around man. And yet in philosophy we still 62:1 Corresponding to the two criteria ad- 76 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM vanced on page 24 above two acid tests for a special case our perceptions of them. philosophy of science may be developed: 62:2 (1) is knowledge regarded as socially produced, i.e. as having a material cause of 1.7 Footnotes to Chap- its own kind? or is it read straight onto the natural world or out of the human mind? ter 1 62:3 (2) are the objects of knowledge re- 1 See Aristotle, Metaphysics, 1.3. garded as existing and acting independently 2 See J. R. Ravetz, Scientific Knowledge and of men? or do they depend implicitly or ex- its Social Problems, pp. 116–19. plicitly upon men for their existence and/or 3 Cf. R. Harre, Philosophies of Science, pp. activity? 176–7. 62:4 Scientists try to discover the rea- 4 W. Thomson, Notes of Lectures on Molecu- sons for things and events, patterns and pro- lar Dynamics p. 132. 5 See A. S. Eve, Rutherford. cesses, sequences and structures. To under- 6 Cf. I. Kant, On the Distinctiveness of the stand how they do so one needs both a con- Principles of Natural Theology and Morals. cept of the transitive process of knowledge- 7 J. S. Mill, A System of Logic, Bk. III, Chap. production and a concept of the intransi- 3, Sect. 1. tive objects of the knowledge they produce: 8 Cf. N. R Hanson, Patterns of Discovery, the real mechanisms that generate the ac- Chap. 1. tual phenomena of the world, including as a 9 Cf. J. J. C. smart, Philosophy and Scientific 1.7. FOOTNOTES TO CHAPTER 1 77

Realism, pp. 38–9. entities with that of the foundations of knowl- 10 Cf. J. J. Gibson, The Senses Considered as edge (p. 5). This leads him to argue that ‘if all Perceptual Systems. possible evidence for the existence of theoreti- 11 Cf. M. Hollis, ‘Reason and Reality’, P.A.S. cal entities is provided by common observables Vol. LXVIII (1967–8), p. 279. it follows . . . that the logically indispensable 12 G. E. M. Anscombe, Causality and Deter- evidence, and thus the sense of assertions about mination, p. 22; and G. H. von Wright, Expla- theoretical entities must be capable of being ex- nation and Understanding, pp. 60–4. pressed in terms of those common observables 13 Cf. G. E. M. Anscombe, op. cit., p. 21. and thus that theoretical entities can have only 14 Cf. Ravetz’s ‘4th law of thermo-dynamics’: a derived and dependent existence’ (p. 285). no experiment goes properly the first time. See 21 D. H. Mellor, ‘Physics and Furniture’, J. R. Ravetz, op. cit., p. 76. American Philosophical Quarterly, Studies in 15 Cf. F. Dretske, Seeing and Knowing, Chap. the Philosophy of Science, p. 184. 3. 22 See W. V. O. Quine, ‘Designation and Ex- 16 Cf. G. H. von Wright, op cit., Chap. 2. istence’, Readings in Philosophical Analysis, ed. 17 L. Wittgenstein, Tractatus Logico- H. Feigl and W. Sellars, p. 50; Methods of Logic, Philosophicus, 6.35. p. 224; and From a Logical Point of View, Chap. 18 I. Kant, Critique of Pure Reason, B.724. 1 and passim. 19 See e.g. A. J. Ayer, Language, Truth and 23 D. Hume, Treatise on Human Nature, p. Logic, pp. 31–41. 632. 20 A recent book, A. Quinton’s Nature of 24 D. Hume, op. cit., p. 132. Cf. Newton’s 4th Things, is vitiated by a failure to distinguish rule of reasoning in philosophy: ‘propositions in- these two questions. From the outset Quinton ferred by general induction from phenomena [are tends to identify the problem of fundamental to be regarded as] true . . . till such time as other 78 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM phenomena occur by which they may either be psychology is possible open. This is important made more accurate or liable to exceptions’, I. because for the transcendental realist it is the Newton, Principia Mathematica, Bk. III. nature of the object that determines the possi- 25 D. Hume, op. cit., p. 231. bility of a science. Thus he can allow, without 26 T. S. Kuhn, The Structure of Scientific paradox, that there may be no humanly intelli- Revolutions, Chaps. II–IV. gible pattern to be discovered in the stars or po- 27 See e.g. M. Hesse, In Defence of Objectiv- litically intelligible pattern in voting behaviour. ity, p. 14. So that no science of astrology or psephology is 28 See e.g. G. Buchdahl, op. cit., p. 31. possible, no matter now scrupulously ‘scientific method’ is adhered to. The next footnote gives premonitions 30 K. R. Popper, Conjectures and Refuta- of the arguments in [Bha89]. tions, p. 192. 29 The structure of such a critique would be as 31 Cf. J. J. C. Smart, op. cit., pp. 149–51. follows: If the subject matter of social psychol- 32 In fact men have the capacity to be self- ogy is such that a science of social psychology conscious in two ways: first in being conscious is possible and social psychologists are to have of what they are doing; and secondly, in being knowledge of it, then social psychologists should conscious of their doing it. That these two are do phi, ψ, etc. rather than x, omega, etc. The not equivalent is shown by the fact that in some transcendental realist could thus not accept the contexts a person may know what he has done notorious definition of economics as what econ- but not that he has done it and vice-versa. omists do. For him, whether or not they actu- 33 L. Wittgenstein. Philosophical Investiga- ally do economics is at least in part a contingent tion, 115. question. Notice that the formula I have used 34 A recent antecedent of the view that causal leaves the question of whether a science of social laws should be analysed as tendencies is con- 1.7. FOOTNOTES TO CHAPTER 1 79 tained in P. T. Geach, ‘Aquinas’, Three Philoso- The sequence a.b thus appears as a consequence phers G. E. M. Anscombe and P. T. Geach, pp. of the results of our actions. It is in this sense 101ff. Important works in the recent develop- that a closure is normally a human product. ment of the concept of powers are W. D. Joske, 37 See G. H. von Wright, op. cit., pp. 43–45. Material Objects, Chaps. 4 and 5; M. R. Ayers, 38 Cf. N. R. Hanson, Observation and Expla- The Refutation of Determinism, Chaps. 3–5; and nation, p. 56. R. Harre, Principles of Scientific Thinking, esp. 39 See e.g. A. Koyre, Metaphysics and Mea- Chap. 10. surement, Chap. 4. 35 I owe this term to M. Scriven, ‘Truisms as 40 V. Ronchi, ‘Complexities, advances and the Grounds for Historical Explanation’, The- misconceptions in the development of the science ories of History, ed. P. Gardiner, pp. 464ff. of vision: what is being discovered?’, Scientific Scriven uses it to refer to generalizations ground- Change, ed. A. Crombie, pp. 542–61. ing historical explanations which contain modi- 41 E. J. Dijksterhuis, The Mechanisation of fiers such as ‘normally’, ‘tendency’, ‘usually’, etc. the World Picture, p. 338. My use of the term is substantially different. But 42 See K. R. Popper, The Logic of Scientific it is the nearest thing to an antecedent for the Discovery, App. XI. kind of conditional I am concerned with. 43 A. S. Eddington, The Nature of the Phys- 36 Formally we could say that in experimen- ical World, p. xi. Stebbing substituted the idea tal production by doing φ we change α to a so of ‘counterparts’ for that of ‘duplicates’ in her altering the state that would otherwise have pre- rendering of the problem. See L. S. Stebbing, vailed; and in experimental production by doing Philosophy and The Physicists, p. 60. or being prepared to do ψ we exclude the in- 44 See e.g. L. S. Stebbing, op. cit., p. 66; and tervention of elements β1 . . . βn, so allowing the G. Ryle, Dilemmas, p. 80. mechanism M set in motion by a to generate b. 45 See e.g. W. Sellars, ‘The Language of Theo- 80 CHAPTER 1. PHILOSOPHY AND SCIENTIFIC REALISM ries’, Current Issues in the Philosophy of Science, ed. H. Feigl and G. Maxwell, p. 76; and P. K. Feyerabend, ‘Explanation, Reduction and Em- piricism’, Minnesota Studies in the Philosophy of Science, Vol. III, ed. H. Feigl and A. Maxwell, p. 83. 46 M. Capek, The Philosophical Impact of Contemporary Physics, p. 294. 47 J. Locke, Essay Concerning Human Under- standing, esp. Bk. IV, Chap. III. 48 W. Kneale, Probability and Induction, pp. 97–103. Kneale’s point could be strengthened by an argument to show that in the case of physical theories the basic entities must be unperceivable. For if they were perceivable it would seem possi- ble to ask what caused them to manifest them- selves to us as perceivable; in which case they could not be basic. This is a general argument in favour of a field-theoretic interpretation of ba- sic entities in physics. Cf. Dingle’s comment that if photons could be seen they would get in the way (J. J. C. Smart, op. cit., p. 38). 49 J. Locke, op. cit., Bk. III, Chap. 6.9. 50 J. Locke, op. cit., Bk. IV, Chap. 3.7. 1.7. FOOTNOTES TO CHAPTER 1 81

Table 1.1 Domain of Real Actual Empirical Mechanisms X Events X X Experiences X X X

Note. for transcendental realism dr ≥ da ≥ de . . . (i) where dr, da, and de are the domains of the real, the actual and the empirical respectively. For empirical realism dr = da = de . . . (ii). Comment: (ii) is a special case of (i), which depends in general upon antecedent social activity, and in which (a) for da = de the events are known under epistemically significant descriptions, which depends upon skilled perception (and thus a skilled perceiver); (b) for dr = da an antecedent closure has been obtained, which depends upon skilled experimentation (and thus the planned disruption of nature). Chapter 2

Actualism and the Concept of a Closure

2.1 Introduction: On the Actuality of the Causal Connection

8263:1 (i) ‘We have no knowledge of any- thing but phaenomena; and our knowledge 2.1. INTRODUCTION: ON THE ACTUALITY OF THE CAUSAL CONNECTION83 of phaenomena is relative not absolute. We N [deductive-nomological] explanation of a know not the essence, nor the real mode of particular event the explanans logically im- production, of any fact, but only its rela- plies the explanandum, we may say that tions to other facts in the way of succession the explanatory argument might have been or similitude. These relations are constant; used for the deductive prediction of the that is, always the same in the same circum- explanandum-event if the laws and the par- stances. The constant resemblances which ticular facts adduced in its explanans had link phaenomena together, and the constant been known and taken into account at a suit- sequences which unite them as antecedent able earlier time. In this sense a D-N expla- and consequent, are termed their laws. The nation is a potential D-N prediction.’3 laws of phaenomena are all we know respect- 63:4/o (iv) ‘Criteria of refutation must be ing them. Their essential nature, and their laid down beforehand: it must be agreed ultimate causes, either efficient or final, are 1 which observable situations, if actually ob- unknown and inscrutable to us.’ served, mean the theory is refuted.’4 63:2 (ii) ‘To give a causal explanation of 64:1 (v) ‘Important though other consid- an event means to deduce a statement, us- erations may be, the acid test of a theory is ing as premises of the deduction one or more its predictive power.’5 universal laws, together with certain singular 64:2 It is the argument of this chapter statements, the initial conditions.’2 that there is a distinction between open and 63:3 (iii) ‘Since in a fully-stated D- closed systems, which most existing philos- 84 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE ophy of science ignores; that closed systems fairs must be constant. (This is the assump- are a condition of its most important doc- tion that the concept of a closure is designed trines, such as those expressed in (i)–(v); and to mark.) On the other hand if, as I intend to that once the significance of this distinction argue, they are not in general constant, then is grasped the plausibility of these doctrines atomistic events cannot provide the only ba- collapses. sis of ontology. And the philosophical theo- ries based on the identification of causal laws 64:3 (i)–(v) possess a family connection, in with empirical regularities plus must all be that they all depend upon the Humean the- radically wrong. ory of law. This theory has often been crit- icized on the grounds that a constant con- 64:4/o I shall use the term ‘actualism’ to junction of events cannot be sufficient for a refer to the doctrine of the actuality of causal law. But most of its critics have been con- laws; that is, to the idea that laws are re- tent to allow that it is at least necessary.6 lations between events or states of affairs It is this notion, viz. that laws are constant (which are thought to constitute the objects conjunctions of events (plus some disputed of actual or possible experiences).7 Behind contribution of mind), that I intend to chal- this idea of course lies the notion that only lenge. It arises as follows: If atomistic events the actual (identified as the determinate ob- or states of affairs constitute the world then, ject of the empirical) is real. Given it, the for general knowledge to be possible, the re- constant conjunction plus analysis of laws lations between such events or states of af- must follow. In this chapter I shall not be 2.1. INTRODUCTION: ON THE ACTUALITY OF THE CAUSAL CONNECTION85 concerned with the ‘plus’. Moreover for con- on the other hand, he attempts to avoid venience I shall use the term ‘empiricism’ this dilemma by restricting the application in a generic way so as to cover the entire of laws to closed systems (e.g. by making post-Humean tradition of empirical realism, the satisfaction of a ceteris paribus clause a and in particular both its positivist and neo- condition of their applicability), he is faced Kantian wings. No harm will be done by this with the embarrassing question of what gov- usage as I am here attacking an assumption, erns phenomena in open systems. If he re- viz. that a constant conjunction is necessary fuses the question, he is still left with the for a law, common to both. problem of accounting for experimental ac- tivity, and thus the establishment of ‘laws’, 65:1 The argument of this chapter is both however restricted, in the first place. His simple and, I think, novel. Leaving aside only options here are to deny either that astronomy, it is only under conditions that men are causal agents or that experimen- are experimentally produced and controlled tal activity plays any role in science. For if that a closure, and hence a constant conjunc- laws are sequences of events and men, being tion of events, is possible. The empiricist causal agents, can bring about and prevent is now caught in a terrible dilemma: for to such sequences, there can be no rationale for the extent that the antecedents of law-like according one rather than another sequence statements are instantiated in open systems, the status of law. A sequence of events can he must sacrifice either the universal char- only function as a criterion for a law if the acter or the empirical status of laws. If, 86 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE latter is ontologically irreducible to the for- the identification of causes to our knowledge mer. And so we come back to the argument of laws. A reductio ad absurdum quickly fol- of 1.3, where I showed how the intelligibility lows. For to the extent that we seek to ex- of experimental activity presupposes the on- plain, predict and identify the causes of phe- tological distinctiveness of causal laws from nomena that occur in open systems, these the patterns of events. But it can now be activities become impossible. And to the seen that not only the experimental estab- extent that they are necessary for our so- lishment but the practical application of our cial life, empiricism does. Thus there is no knowledge depends upon this same ontolog- necessity that we should exist. But, given ical distinction. For unless causal laws per- that we do, if our social life is to be possi- sisted and operated outside the context of ble we must ascribe causal responsibility in their closure, i.e. where no constant conjunc- open systems. And given this, the Humean tions of events obtained, science could not theory just cannot apply. Now I want to ar- be used in the explanation, prediction, con- gue both that laws apply in open and closed struction and diagnosis of the phenomena of systems alike; and, in a subsidiary thesis, ordinary life. that these other activities do not necessar- ily depend upon (though they may make use 65:2/o The empiricist makes matters of) a knowledge of laws. From this perspec- worse for himself by the fact that he not only tive the Popper-Hempel theory of explana- ties laws to closed systems, but typically ties tion, for example, may be seen to involve a the activities of explanation, prediction and 2.1. INTRODUCTION: ON THE ACTUALITY OF THE CAUSAL CONNECTION87 double mistake: first, that explanation al- word or even the concept ‘law’. Rather what ways involves laws; and secondly, that laws is essential to the realism developed here is are or depend upon empirical regularities. the idea that the things and mechanisms of nature, that constitute the intransitive ob- 66:1/o My overall aim, it will be remem- jects of scientific theory, both exist and act bered, is to argue that the ultimate objects independently of the conditions, normally of scientific understanding are neither pat- produced by men, that allow men access to terns of events nor models but the things them. For experimental science to be pos- that produce and the mechanisms that gen- sible the world must be at least partially erate the flux of the phenomena of the world. open. But if there is a real distinction be- Scientists attempt to discover the way things tween open systems and closed and our in- act, a knowledge typically expressed in laws; tuitions about the rationality of science are and what things are, a knowledge (to be to be preserved there must be a real distinc- discussed later) typically expressed in real tion between structures and events. In this definitions. Statements of laws, I have sug- respect the differentiation of phenomena still gested, are statements about the tendencies provides the best argument for the stratifi- of things which may not be actualized, and cation of the world. may not be manifest to men; they are not statements about conjunctions of events, or 67:1 In isolating the special conditions un- experiences. But in developing this theory der which a regular sequence or pattern of I do not attach any great importance to the events occurs; that is, in which (to adopt the 88 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE realist mode) there is a correspondence be- alism my primary intention in this chap- tween causal laws and the pattern of events, ter is critical. For it is when confronted I will be leaving it up to the epistemologist with the fact and implications of open sys- whether he wants to sustain the universality tems that the limitations of this approach to of laws (and inter alia the intelligibility of science—with its flat ontology of undifferen- experimental activity) by postulating a cate- tiated experience—become most apparent. gorical ontological distinction between them. But, in dealing with the problems posed by If this is done by the development of a non- the largely unanalysed phenomena of open empiricist ontology and an analysis of laws systems, I will also be compelled to develop as non-empirical and normic along the lines new and more general alternatives to the indicated in 1.5 above, the way is also open theories, such as those expressed in (i)–(v) for an adequate theory of natural necessity above, that are based on the tacit assump- and natural kinds. On the other hand with- tion that a closure is the universal rule in na- out this, I shall argue in Chapter 3, our in- ture; rather than the rare exception I shall tuitions about the lack of sufficiency of the contend it is. Humean criteria for law (and the theories of science based on it) cannot in the last in- 67:3/o Underlying the widespread, if tacit, stance be sustained. acceptance in philosophy of the idea of the ubiquity of constant conjunctions in nature 67:2 In showing how a closure is a con- (an idea which is not confined to the empiri- dition of the intelligibility of empirical re- cist tradition)8 and hence of the doctrine of 2.1. INTRODUCTION: ON THE ACTUALITY OF THE CAUSAL CONNECTION89 the actuality of causal laws is the notion that given that rough-and-ready regularities are the universe is at rock bottom deterministic; everywhere at hand, it seems plausible to that, in the image of Leibniz, the present is suppose that underlying them there must big (in the sense of pregnant) with the fu- be more exact ones. Partly no doubt be- ture; that it, as it were, already contains it cause of an obsession with the consequences now. It is the job of science to discover the and a neglect of the conditions of the exper- iron laws that uniquely determine its motion. imental paradigm, the single case that the Once these laws are discovered, given only a hypothetico-deductive view of science fits. knowledge of any complete state-description, Above all perhaps because of the misconcep- ‘nothing would be uncertain [to science] and tion created by the celestial closure secured the future, as the past, would be present by Newtonian science, and in particular by to its eyes.’9 What accounts for the hold of the idea that this closure embodied both a this fantastic conception on our philosoph- model of phenomena and a model of sci- ical imagination? The philosophical argu- ence. This was a double mistake. For it was ments for it are, taken on their own, as we not the human mind, as Laplace thought,10 shall see, pretty poor. Why then do we feel that gave its special perfection to astronomy. the force of this picture? Partly no doubt be- Rather it was the peculiar conditions of the cause many things are de facto predictable, planets, and in particular the constancy of many processes are effectively isolated and both their intrinsic states and the external many systems more or less closed; so that, forces on them, that made possible the ob- 90 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE served regularities. Moreover for Newtonian, of ‘machines’, including classical mechani- as for any other, mechanics celestial phenom- cal ones, cannot be adequately described, let ena functioned merely as evidence that bod- alone understood, in terms of the ‘whenever ies tend to act in certain ways. The laws of x, then y’ formula of regularity determinism. motion, for example, describe actions which Thirdly, that even if the world were a single are unobservable in principle. But the ten- ‘machine’ this would still provide no grounds dencies of the bodies to which they apply are for the constant conjunction idea, or a for- real; and would account for any disruption in tiori any of the theories of science that de- the established order of our solar system. pend upon it. Regularity determinism is a mistake, which has been disastrous for our 68:1/o But, it might be objected, is not understanding of science. the universe in the end nothing but a giant machine with inexorable laws of motion gov- erning everything that happens within it? I want to say three things: First, that the var- 2.2 Regularity Deter- ious sciences treat the world as a network of minism and the ‘machines’, of various shapes and sizes and degrees of complexity, whose proper princi- Quest for a Closure ples of explanation are not all of the same kind as, let alone reducible to, those of classi- 69:1 So far I have been content merely to cal mechanics. Secondly, that the behaviour identify a closed system as one in which a 2.2. REGULARITY DETERMINISM AND THE QUEST FOR A CLOSURE 91 constant conjunction of events obtains. But every event y there is an event x or set of we must now establish exactly what this en- events x1 . . . xn such that x or x1 . . . xn and tails. It might be thought that the idea of a y are regularly conjoined under some set of closed system could be elucidated quite sim- descriptions.11 That is, the world is so con- ply as a fragment or sector of the world ef- stituted that there are descriptions such that fectively cut off for a period of time from for every event the simple formula, ‘When- non-constant external influences. Although ever this, then that’ applies. this gives one clear sense of ‘a closed system’ such a system would not necessarily satisfy 70:1 Such a thesis stands to the practice of the criterion of invariance implicit in the em- science as a regulative principle. Such princi- piricist analysis of law. For one thing con- ples are, as is well known, neither empirically ditions would have to be placed on the indi- nor theoretically refutable (or confirmable). viduals composing the system and the way But I will contend that they are metaphys- in which the states of the system were to be ically so. My procedure will be to see what specified. But even if this were done there this thesis entails about the nature of the would still be no guarantee that the crite- world and about the nature of science; and to rion of invariance would be satisfied. The assess its adequacy in these respects in rela- assumption that it would be depends upon tion to other possible regulative standpoints. the metaphysical thesis of regularity deter- To do so I will work out critical or test con- minism. This may be defined as follows: For ditions for the thesis of regularity determin- ism; that is, conditions such that if they were 92 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE known to be satisfied and the constant con- word. junction formula was not vindicated the reg- ularity determinist would be bound to admit 70:2/o Regularity determinism must be his thesis refuted. In this way I hope to show straightaway distinguished from two other just how restricted in its ontological presup- forms of determinism: which may be called positions and restrictive in its methodolog- ‘ubiquity’ and ‘intelligibility’ determinism. ical responses regularity determinism is. In Ubiquity determinism asserts that every developing these limit conditions for a clo- event has a real cause; intelligibility deter- sure I will thus be developing the conditions minism that every event has an intelligible under which, on the supposition that regu- cause; regularity determinism that the same larity determinism is true, a constant con- (type of) event has the same (type of) cause. junction of events must obtain. However, I The concepts of ‘cause’ involved in the three will define a ‘closed system’ simply as one determinisms are of course distinct. For the in which a constant conjunction of events ubiquity determinist the cause is that thing, obtains; i.e. in which an event of type a is material or agent which is productive of an invariably accompanied by an event of type effect; for the intelligibility determinist it is b. Clearly the possibility of such a system simply that which renders an event intelligi- does not depend upon the truth of regular- ble to men;12 for the regularity determinist ity determinism. Nor need such a system be it is the total set of conditions that regularly ‘closed’ in any more picturesque sense of the proceeds or accompanies an event.13 Of the three determinisms, regularity determinism 2.2. REGULARITY DETERMINISM AND THE QUEST FOR A CLOSURE 93 is clearly the most restrictive; and ubiquity deduction of traits, computational determin- determinism is more general than intelligibil- ism is a truism. It says merely that given ity determinism, because it licenses no pre- any system it is possible to work out an al- sumption that the real cause of an event will gorithm for the successful computation of its always be intelligible to men. The realist, in- traits; or, in other words, that there is a con- telligibility and regularity concepts of cause sistent way of describing the development of are of course naturally associated with the any system over any finite period of time. transcendental realist, transcendental ideal- Moreover in general there will be an infinite ist and classical empiricist philosophies of number of ways of doing so. It is therefore science. an uninteresting truism—save perhaps to re- mind us that the notion of disorder or chaos 71:1 All three determinisms must be dis- is always relative to a particular type of or- tinguished from the idea of ‘computational’ der or class of functions;14 and that the cri- determinism. This is the supposition that terion of deducibility is too easily satisfied to for each characteristic or trait of any mate- be capable of functioning alone as a decision rial body at any moment of time there exists rule for the selection of ‘law-like’ or ‘theo- at least one set of statements from which, retical’ statements and requires at the very together with the relevant antecedent state- least supplementation by criteria that place descriptions, that trait is deducible. It is some restriction on the number, type or in- important to realize that because no restric- tion is placed on the statements used in the 94 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE terpretation of the statements concerned. of Edge Hill). We need not dwell on these possibilities here. For they have been thor- 71:2/o It is especially important to dis- oughly explored by philosophers concerned tinguish regularity from computational de- to defend the autonomy of historiography terminism. For it is at least part of the in- from (Positivist) science.15 For the regularity tention of the former to assert (a) that the determinist the necessity for such formula- same cause and effect sometimes as a matter tions merely indicates the ignorance of the of fact recurs and (b) that the same cause describer. The anti-regularity-determinist, and effect could always logically recur. For on the other hand, may take it as a sign unless (a) were true instances would not fall of emergence or novelty in nature or even under it and unless (b) were true they could of the self-determination of some agent or not fall under it; so that it would be at best structure. Unlike computational determin- vacuous and at worst false. On the other ism, regularity determinism is not trivially hand computational determinism is consis- satisfied. It does however share with it the tent with ‘law-like’ formulations which are feature that if there is one set of law-like so specific and detailed as to reduce the statements which satisfies it there will be an practical likelihood of the event’s recurrence infinite number of such sets. Hence it too re- towards zero or which mention the spatio- quires supplementation by additional crite- temporal location within which it occurred ria, such as simplicity, intelligibility or real- or which individuate it with a definite de- ism, if it is to be capable of yielding a unique scription or a proper name (e.g. The Battle 2.2. REGULARITY DETERMINISM AND THE QUEST FOR A CLOSURE 95 decision procedure for the selection of laws’. prevail, the effect is uniquely predictable. But it is not consistent with retrodictabil- 72:1/o The total cause (in Mill’s sense) ity. For given the effect, we cannot uniquely of an event will normally be a complex set infer the cause. And this requirement is of conditions x1 . . . xn rather than a single explicit both in the Laplacean ideal, which event x. One could distinguish here be- places the past on a par with the future, tween the individual or component events and the Humean definition of cause (‘in other or states and the total or conjunct event words, where if the first object had not been, or state; and refer to the case where more the second never had existed’), which makes than one factor is at work, following Mill, the cause both necessary and sufficient for as that of ‘multiple causation’. In the same the effect.17 To satisfy this requirement the way the consequent event will also normally idea that every event has one and only one be complex, i.e. y1 . . . yn rather than simply cause (or set of causes) must be incorporated y; and so we could talk of a corresponding into our definition of regularity determinism. ‘multiplicity of effects’. There is a genuine This must now read: the same cause always ‘plurality of causes’ when the same effect has the same effect and the same effect al- arises from different (i.e. alternative) sets of ways has the same cause; so eliminating both conditions.16 Is a plurality of causes consis- the possibility of a disjunctive plurality of tent with regularity determinism? Note it causes and of a disjunctive plurality of ef- is consistent with predictability. For given a knowledge of the conditions which actually 96 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE fects. system’ is defined: if and only if it includes t’ it is open. ‘System’ here carries no indepen- 73:1/o Now suppose we had a system such dent semantic force. Either way the natural that events of type a were invariably followed response of the regularity determinist to this by events of type b. We could then say that situation would be to suppose that we had a closure had been obtained. A closure is of left out of our state-description an individual course always relative to a particular set of or variable that made a difference: that the events and a particular region of space and conjunct events referred to under the same period of time. Now supposing that at some description ‘a’ before and after t’ were not re- time t’ an event of type a was not followed by ally the same in all relevant respects; in short an event of type b we would have to say that that the system had been incompletely de- the system was ‘open’, our criteria of open- scribed (or enumerated). For example if the ness just being the fact that events of type system was a classical mechanical one, where a had not been invariably followed by events the presupposition was that mass, position of type b under their given descriptions; i.e. and velocity were the only relevant variables, the instability, in space or over time, of actu- it would be natural to suppose that a rele- ally recorded empirical relationships. Should vant individual had been omitted from the we say that the system had been closed but specification of the overall state of the sys- was now open or that it was open all along? tem, i.e. the total conjunct event. There is nothing at stake here—it depends entirely on the time period for which ‘the 74:1 To fix the point, imagine a universe 2.2. REGULARITY DETERMINISM AND THE QUEST FOR A CLOSURE 97 composed of a finite number of different ables of an entirely different kind, e.g. the kinds of knives, forks and spoons. Suppose room temperature—to distinguish say be- that we attempted to work out a general tween winter breakfasts when porridge was rule which would enable us to predict that the rule and summer ones when grapefruit when the knives and forks were in a certain was. position (including naturally the datum of whether they were on the dining-room table or in the kitchen drawer) they were invari- 74:2 Subject to an important qualification ably followed by another constellation of po- to be discussed below it can thus be seen sitions at the next meal. We might find this that regularity determinism implies a partic- impossible, unless we took into account the ular kind of response to the phenomenon of positions of the spoons; so that we could say open systems, i.e. to the instability of empir- that our system had been incompletely de- ical relationships, viz. to assume that some scribed in the former case, owing to the omis- causally relevant individual or variable has sion of a causally relevant (in the Humean been left out of the description. On the other sense) variable. Of course in time we might hand, even if this were the case, stable em- find that we could only satisfy the demands pirical relationships might still be possible as of regularity determinism by taking into ac- long as the values of the omitted variables count further variables, e.g. the shapes and remained constant. A closure thus depends number of glasses at the meal; or even vari- upon either the actual isolation of a system from external influences or the constancy of 98 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE those influences. pies a part, will be a variable. So either the absence or the constancy of internal struc- 75:1 Assuming that a system was effec- ture must also be a condition for a closure. tively isolated from non-constant external in- And the regularity determinist now has an- fluences and regularity determinism was true other possible response to the condition of would this then ensure the satisfaction of the open-ness, namely to assume that the indi- ‘whenever this, then that’ formula? Let us viduals of the system have not been given a suppose that we are interested in explain- simple or atomistic enough description. ing (in the sense of Hempel and Hume) the behaviour of some individual N, say an ele- 75:2 It is easy to see that an actual phant. Would a knowledge of the total an- isolation and atomistic individuals will be tecedent state-description enable us to pre- preferred, on epistemic grounds, by the dict its behaviour? No—for if N is character- regularity determinist to constant external ized by internal structure and complexity it and internal conditions. For the regular- may behave differently in the same external ity determinist has no warrant for assuming circumstances in virtue of its different inter- that these conditions will remain constant.18 nal states. Thus what happens when I prod Whether they do or not will depend upon an elephant depends at least in part upon a whole host of factors concerning which ex what state it is in, e.g. whether it is asleep hypothesi he has no knowledge and about or not; and thus to that extent the total state which he is not therefore in a position to of the universe, of which the elephant occu- make any kind of claim. Only with an ac- 2.2. REGULARITY DETERMINISM AND THE QUEST FOR A CLOSURE 99 tual isolation of atomistic individuals will the epistemic grounds. Behind the assumption regularity determinist be able to categori- of additivity lies of course the idea that the cally predict the future; without it, it always behaviour of aggregates and wholes can al- remains on the cards that an unpredicted ways be described in terms of the behaviour change in the external circumstances of the of their component parts. The assumptions system or the internal states of its individu- of atomicity and additivity are closely con- als will occur so as to upset an established nected. For to say of some system that it is regularity and so render inapplicable any hy- irreducible (in this sense) to its component pothetical predictions, formulated subject to parts is presumably to say that it must be two ceteris paribus clauses. viewed as a thing in its own right, at its own level (I am not concerned with the grounds 75:3/o I have been tacitly assuming up till for this now.) Conversely to suppose that now that the overall states of the system can it is always possible to give an atomistic de- be represented as an additive function of the scription of prima facie complex things is to states of the individual components of the suppose that they can always be viewed as system. This represents a third kind of re- systems or parts of systems which can be quirement for a closure. Here again a closure analysed in terms of their component parts, is possible if the principle of organization is conceived as atomistic individuals. non-additive, provided it remains constant; though here again the regularity determinist 76:1/o The critical conditions for a closure will prefer the alternative of additivity on are set out in Table 2.1. The satisfaction of 100 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE one each of the system, individual and orga- things of another type. ‘Isolation’ must also nizational conditions is sufficient, on the sup- be interpreted metaphorically; but there is position that regularity determinism is true, a sense in which ‘atomicity’ must be taken for a closure; but not necessary for it. If literally. a recorded regularity breaks down the reg- 76:2 Table 2.1 Limit Conditions for a Clo- ularity determinist must assume that it is sure, i.e. for the Stability of Empirical Rela- because one of these conditions is not satis- tionships ——————————————— fied. Until now in developing the conditions ——————- Conditions (1) Epistemically for a closure I have been using the categories (2) Epistemically for a Closure Dominant ‘internal’ and ‘external’. But the categories Case Recessive Case —————————— ‘intrinsic’, and ‘extrinsic’ are better in that ———————————- (A) System Isola- they are not explicitly tied to a spatial char- tion Constancy of Extrinsic Conditions (B) acteristic and hence to things of a certain Individuals Atomicity Constancy of Intrinsic type. Thus the category ‘intrinsic’ includes Conditions (C) Principle of Additive Con- some properties of things which lie outside stancy of Non- Organisation Additive Prin- their spatial envelope, e.g. a magnet’s field, ciple ————————————————— and others which cannot be identified spa- ————- tially at all, e.g. a person’s charm. And it excludes others which do lie within their 77:1 Now it is easy to see that once spatial envelope, e.g. properties belonging to an actual isolation and an atomistic de- scription are set up as norms two regresses 2.2. REGULARITY DETERMINISM AND THE QUEST FOR A CLOSURE 101 are initiated, viz. to systems so vast that paradoxes in their wake. For since in the they exclude nothing and to individuals so first case there are at the limit no conditions minute that they include nothing. These re- extrinsic to the system a full causal state- gresses are typically manifest in research pro- ment would seem to entail a complete state- grammes, characteristic of positivistic sci- description (or a complete history) of the ence, which could be dubbed ‘interactionism’ world. Similarly as in the second case there and ‘reductionism’ respectively. It is clear can be no conditions intrinsic to the thing that they can only be halted by constitut- a causal statement entails a complete reduc- ing a level of autonomous being, somewhere tion of things into their presumed atomistic between the universe and atomistic individ- components (or their original conditions). In uals. But for the empiricist committed to the first case we cannot make (or can at regularity determinism to do so involves an best only make in a pragmatic way) the dis- enormous risk. For it means he must be pre- tinction between causes and conditions; in pared to snap the Humean link that ties the the second case that between individuals and justified performance of cognitive acts such variables or between a thing and its circum- as the ascription of causes to a knowledge of stances. In neither case do we have the key empirical invariances; and to say ‘yes I know concept of a causal agent; i.e. the thing that a is not invariably followed by b, yet a caused produced or the mechanism that generated, b here’. in the circumstances that actually prevailed,

77:2 These regresses generate notorious 102 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE the effect in question. determinist and realist turns on the question of whether there are objects not susceptible 77:3/o Open systems situate the possibil- of an atomistic analysis and in what way this ity of two kinds of possibility statements: is significant for science. epistemic and natural. The regularity deter- 78:1 A special and very important case of minist can accommodate the former but not the individual conditions for a closure is thus the latter. For he may allow that an event given by: B1’ the absence of powers, which may be uncertain due to the describer’s is dependent upon the absence of intrinsic ignorance of the complete atomistic state- structure (implied by atomicity); and B2’ description necessary to deduce it. But he the constancy of powers, which is dependent cannot allow that there is a sense to a state- upon the constancy of intrinsic structure. If ment about what an individual can do inde- there are complex things then it becomes im- pendently of whether or not it will do it. For portant to distinguish between the subjects natural possibility statements to be possible and the conditions of action. For conditions condition B1 must be unfulfilled; i.e. there is an epistemic, not an ontological category. must be complex things, possessing intrinsic The conditions change, but they do not have structure, to which the natural possibility is the power to change. Only things and mate- ascribed. Without this, the distinction be- rials and people have ‘powers’. tween a power and its exercise would be in- deed, as Hume supposed, entirely ‘frivolous’. 78:2/o I have argued in effect in Chapter So in part the issue between the regularity 1 that for experimental science to be possi- 2.3. THE CLASSICAL PARADIGM OF ACTION 103 ble the world must be open but susceptible corporate more and more elements into his to regional closures. Now corresponding to descriptions and/or to break down his units the view of the world as open and the view of study into finer and finer constituents in of the world as closed we have two entirely an effort to stabilize his field. And he must different conceptions of science. The tran- see this merely as an attempt to vindicate to scendental realist sees the various sciences as himself the rule, which may be fairly styled attempting to understand things and struc- a dogma, that like follows like or whenever x tures in themselves, at their own level of be- then y. ing, without making reference to the diverse conditions under which they exist and act, and as making causal claims which are spe- cific to the events and individuals concerned. 2.3 The Classical And he sees this not as a tactic or manoeu- Paradigm of Ac- vre or mechanism of knowledge; but as ac- cording with the ways things really are, the tion way things must be if our knowledge of them is to be possible. The regularity determin- 79:1 In developing the critical conditions for ist, on the other hand, will seek in his quest a closure I have argued that epistemological for a closure, if the very special conditions considerations tilt them in favour of the cases specified in Table 2.1 are not satisfied, to in- of atomicity and additivity, i.e. cases B1 and C1 in Table 2.1 above. I now want to bring 104 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE out the concept of action implied by these of action, involving a corpuscularian view of conditions, which will be seen to be as special matter and a mechanical view of causality, as the conditions themselves. in which all causes are regarded as efficient and external to the thing in which the change 79:2 If individuals are atomistic, then all occurs, is implied by the limit conditions for causes must be extrinsic. And if systems a closure. These form the twin pivots of what are lacking in intrinsic structure (i.e. are ex- I am going to call the classical paradigm of hausted by the properties of the individu- action. als composing them), then there can be no action at a distance (nor can distance be a 79:4 Now this paradigm has three varia- variable in action), so all action must be by tions: a physical one; a metaphysical one; contact. But as atomistic individuals cannot and a distinctively positivistic epistemolog- contribute to or be affected by the action, ical one. It is important to keep them dis- it must consist in the communication of re- tinct. For the prestige of the epistemological ceived properties. But as the only property variation, which underpins the Humean the- possessed by such individuals is their posi- ory of laws, has stemmed in large part from tion in space at a moment in time, the only its misidentification, by Locke and Newton property they can communicate is their mo- among others, with the physical one. tion, i.e. their movement through space in 79:5/o On the physical concept, matter time. is viewed as composed of rigid corpuscles 79:3 Thus a particular physical conception whose motion accounts for the aggregative 2.3. THE CLASSICAL PARADIGM OF ACTION 105 and observed behaviour of things. These cor- different, i.e. genuinely emergent ways (as in puscles exchange momenta and redistribute the case of secondary qualities). velocities among themselves by impact; and 80:1 Now this physical concept of action they move through space according to the encourages, though it does not imply, a strict laws of mechanics. Action is seen metaphysical concept, which was especially as consisting in the impression of external prominent in seventeenth century rational- forces upon these corpuscles, which merely ist thought. On it, matter is viewed as es- pass on their received motion by direct im- sentially passive and inert (which is strictly pact. Events are nothing but the displace- speaking irreconcilable with the role played ment of these basic units of matter in space in classical mechanics by the concept of in- and time, they are not the transformation ertia); and causation is viewed as a process of pre-formed substances. Large scale events which is linear and unidirectional, as well as or macroscopic changes are merely the sur- external and inconsistent with real novelty. face effects of such displacements; qualitative In contrast to matter is mind. As matter is variety and change are similarly the effects passive, mind is active. And qualitative vari- of different arrangements and motions of the ety and change, denied a place in the sphere corpuscles and their aggregates. The prop- of matter, are seen as contributions of the erties of aggregates are essentially the same human mind. as those of their component parts, though they can manifest themselves to observers in 80:2 Historically associated with this en- semble of physical and metaphysical ideas 106 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE is another distinctively positivistic epistemo- and reconstruct the world from them, is it at logical concept. On this concept, things are all plausible to suppose that the world con- viewed as ultimately resolvable into simple sists of independent and atomistic events and qualities apprehended in sense-experience, states of affairs, constituting the surrogates rather than as aggregates of elementary units of such sense-experiences, with relations that of matter in motion. The ontology is one are as external to one another as the sense- of atomistic (and independent) events rather experiences that ground them are held to be than one of atomistic (and rigid) corpuscles. distinct. Conversely, if we reverse this pre- And causality is seen as the regular concom- supposition and situate sense-experience as a mitance of such events, rather than the im- natural process occurring in the world then pression of external forces upon such indi- it always makes sense, on being told that viduals. For the positivist, the concept of some event or happening occurred, to ask: action is but a gloss put on our apprehension ‘to what thing?’ (One could then of course of such sequences. It can have no application go on to explain the change, along the lines of in a world of independent events. the corpuscularian/mechanical programme, in terms of the arrangements and motions of 80:3/o It should be noted that the episte- the basic individuals.) mological variation on the classical paradigm only makes full sense if it is given a phenom- 81:1/o On the physical concept it is enalistic interpretation. For if and only if the principles of action (the strict laws of we start from punctiform sense-experiences mechanics) that explain macroscopic be- 2.3. THE CLASSICAL PARADIGM OF ACTION 107 haviour, including observed regularities; i.e. a radical break with the pre-existing Aris- the relationship between individuals and sys- totelian scheme which certainly was then, tems is one of explanation, not analysis. and may still be now, closer to our ordi- Now I have argued in Section 1 that these nary way of thinking.20 Consider, for exam- laws cannot be plausibly construed as em- ple, the radical change embodied in the prin- pirical regularities or constant conjunctions ciple of inertia, that only change in motion of events. Moreover, if it is the behaviour and not motion itself (i.e. change in posi- of the individuals that is required to ac- tion) requires explanation. Thus the notion count for the constant conjunctions of events that ‘the concepts of classical physics are just rather than the other way round the point a refinement of the concepts of daily life’21 of the Humean analysis of cause and the must be rejected, and with it the notion that Hempelian analysis of explanation is lost. If laws are inductive generalizations from ev- the laws cannot be regarded as empirical reg- eryday experiences or at any rate descrip- ularities, the concepts in them cannot be re- tions of them or instruments for predicting garded as given in or abstracted from expe- therm abandoned. For the experiences are rience. Firstly, because we ordinarily experi- literally different—an idea which will only ence motion in terms of transitive verbs such seem paradoxical to someone who is implic- as ‘pushing’ and ‘pulling’ which cannot be itly misusing the category of experience to explicated ostensively.19 Secondly, because define the world. the concepts that figure in them demanded 82:1 The key concept of atomicity is 108 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE given radically different interpretations in properly so-called should be constantly con- the physical and epistemological variations. joined. The requirements of epistemic atom- In the physical variation its primary iden- icity and empirical invariance can only be tification is size; in the epistemological one reconciled if the sequence of events is a lin- simplicity. Physical atoms are (or were) nec- ear process and each event is a member of essarily unobservable, theoretical entities; a homogeneous series of determinations. If epistemological ones were the raw data of ex- this were the case then the apprehension of perience. The former were the ultimate en- the cause event would indeed license the ex- tities of the world, the latter the basic build- pectation of the effect event. But this leaves ing blocks of knowledge. Physical atoms the odd assumption of the linearity of pro- were explainers; epistemological ones justi- cesses to be justified. In the formula ‘when- fiers. Physical atoms were Parmenidean in- ever x then y’, it is only if ‘x’ and ‘y’ are dividuals, epistemological ones Heraclitean taken to refer to the objects of actual or instants. possible experiences that we have an em- piricist (in the broad sense) analysis of the 82:2/o Epistemic atomicity requires that causal modality; and it is only if these ex- events should be apprehended in raw sense- periences are taken to be atomistic that we experience. Hence it restricts the possible have a specifically Humean one. But it is numbers of a cause and an effect set to only if ‘x’ and ‘y’ constitute complete as well one. But the constant conjunction analysis as atomistic state-descriptions that the con- of cause requires that the cause and effect 2.3. THE CLASSICAL PARADIGM OF ACTION 109 junction between x and y will (supposing The essential features of the classical corpus- regularity determinism to be true) be con- cularian/mechanical world-view can be sum- stant. These desiderata can only be recon- marized as follows: – ciled if all causal sequences are linear pro- 83:2 (i) the externality of causation; cesses; an idea which it would seem extraor- 83:3 (ii) the passivity of matter, and the dinarily difficult to defend. Finally, it should immediacy of effects; be noted that whereas the constant conjunc- 83:4 (iii) the atomicity of fundamental en- tions are for the most part unknown (as the tities (whether corpuscles, events or sense- regularity determinist will be bound to con- data); cede), the events they are supposed to con- 83:5 (iv) the absence of internal structure join are regarded on the Humean analysis as and complexity; being intuitively ascertainable! 83:6 (v) the absence of pre-formation, and 83:1 I have dwelt at some length on the of material continuity; differences between the physical and episte- 83:7 (vi) the subjectivity of transfor- mological concepts; as well as hinting in the mation and of apparent variety in nature last paragraph at internal difficulties in the (i.e. metaphysically, qualitative diversity and latter (which will be developed in Chapter change are ‘secondary qualities’). 3). I now want to say something about the 83:8/o It has already been seen that at general character of the paradigm as such the limit conditions for a closure the distinc- and the limitations of its various ingredients. tion between intrinsic and extrinsic condi- 110 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE tions collapses. For atoms possess no intrin- out a drop from the bottle marked ‘ACID’, sic conditions. An atom is distinguishable irrespective of the nature of the substance only with respect to its position (or some or material onto which it is poured; or the higher order derivative of position such as sound of breaking glass every time a cricket motion, acceleration etc.) in space at a mo- ball hits a stationary target at 30 m.p.h., ir- ment in time. Consider now the old mech- respective of whether it hits a greenhouse, a anistic prejudice crystallized in Hobbes’ dic- sand pit or a granite wall. Such ‘explana- tum that ‘nothing taketh a beginning from tions’ border of course on the fatalistic; or itself’.22 Does this mean that nothing taketh rather we could say that they are fatalistic any part of its beginning from itself; i.e. with respect to things. For if there were sub- that none of the total set of necessary and stance or material conditions for the event sufficient conditions for an event is intrin- such that were they not satisfied it would not sic to the thing (whether or not the trig- have occurred, then they must be included as ger that sets it off, so to speak—what we part of the total cause (in Mill’s sense) of the should ordinarily call ‘the cause’—is intrin- event. Once we allow that an event would sic)? If it does, this is tantamount to assum- have occurred, whatever the intrinsic condi- ing that the event occurs, on the satisfaction tions, we are bound to end up denying the of certain antecedent conditions, whatever principle of material continuity. This is the the thing. It encourages, as it were, a pic- principle that events are changes in things, ture of a cloud of smoke every time we pour never replacements of one kind of thing by an 2.3. THE CLASSICAL PARADIGM OF ACTION 111 entirely new kind of thing.23 But if all causes and historically, physicists have tended to are extrinsic there can be no material conti- treat as ‘substantive’ precisely that which nuity through change: either because there tends to be conserved, e.g. matter or energy. is no continuity (positivism) or because there In our ordinary ascriptions of change then a is no change (corpuscularianism). material as well as an efficient cause is nor- mally essential. The classical paradigm di- 84:1/o Ordinarily we think of the world rects exclusive attention to the latter. And as consisting of things which endure through yet in scientific contexts it is often the search some but not other changes. Events we think for hidden entities, leading to the identifica- of primarily as changes in things, i.e. as the tion of novel kinds, set off by what appears transformation of substances, rather than prima facie to be a case of ex nihilo produc- the displacement of physical masses in space tion, that is most important. (Cf. the dis- and time. What is transformed is already covery of the neutrino.) It is the absence given as complex and pre-formed. If it is of the notion of material continuity through partially transformed, material continuity is change, as it is manifest in the epistemologi- preserved through the change. If it is totally cal variation of the classical paradigm, where transformed, we seek for a new kind of sub- it results in the generation of an ontology of stance, or level of ‘thing’, which will allow us atomistic and independent events, that un- to preserve this principle.24 A chemical atom derpins the idea of the contingency of the preserves continuity through chemical reac- causal connection, which we turn to in the tions; a gene-pool through species change; 112 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE next chapter after rejecting the idea of its those of their component parts. But neither actuality here. is there any reason why all things should be conceived as either wholes or parts. I am go- 85:1 Perhaps under the influence of the ing to suggest later that we are radically mis- classical paradigm we tend to view events led by spatial metaphor and imagery here. as happenings to passive things. But events Societies, people and machines are not col- are also the results of actions. Panes of glass lectivities, wholes or aggregates of simpler or do not shatter without being hit by things smaller constituents (just as intentionality is such as cricket balls; nothing could be mag- not an inner urge or push). In the classical netized unless there were magnets. Thus world view it was the function of matter to some things must be agents as well as pa- occupy space;25 so it was natural to assume tients; just as some conditions must be in- that all ‘things’ properly so-called were just trinsic as well as extrinsic. Yet we feel the more or less highly differentiated aggregates source, the stimulus, the trigger is always ex- of matter, and so could be viewed either as trinsic. But this is a pure prejudice. For not wholes or parts (or as both). all efficient causes are extrinsic and not all extrinsic causes are mechanical. Thus the 85:2/o The victories of the corpuscular- structure of a field or the organization of ian/mechanical programme were never as an environment may be the cause of what complete as its propagandists made out. happens within it. There is no reason why Neither gravity positing action at a dis- the properties of wholes should not explain tance nor magnetism in which distance was 2.3. THE CLASSICAL PARADIGM OF ACTION 113 a variable in action could be assimilated chanics had been fully developed and ap- to the paradigm. The fact that Newton plied successfully in domain after domain could not find a contact explanation for grav- there was bound to come a time when, un- ity provided the basis for much early criti- der the stimulus of internal inconsistencies cism of his work; and the theory of a con- and irrefutable counter-instances, the prin- necting medium—the aether—has its post- ciples of mechanics would themselves have relativistic advocates even today.26 Paradox- to be explained non-mechanistically. More- ically the very fact that Newton could not over we know as a matter of fact that the find an explanation for gravity strengthened Newtonian system has been replaced and not the positivist variant of the paradigm—for just subsumed;28 and that in one respect af- it gave a degree of credibility to the belief ter another the classical world view has been that science eschews ‘hypotheses’.27 Never- abandoned by modern physics.29 But given theless the prestige of Newtonian mechan- this, it is still important to ask (in view of ics was such that from Hume’s time on- its perennially attractive features) whether ward scientific explanation came to be iden- it provided a logically coherent schema for tified with mechanical explanation; and a the conceiving of fundamental explanations reduction of all other branches of learning in physics or indeed any other science. to mechanics was loudly proclaimed by Hel- vetius and its other propagandists to be at 86:1/o Now both Boscovitch and Kant hand. Of course we know that after me- argued, anticipating later developments in physics, that contact explanations could not 114 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE be fundamental; but that they themselves location (and its derivatives), the only at- required explanation in terms of forces—of tribute such ‘atoms’ could possess would be attraction and repulsion—acting at a dis- the power of affecting things at a distance. tance.30 For, they argued, since action by In short ‘atoms’ would be playing a purely contact proceeds by compression, the bod- nominal role. Such atoms would not be ies involved cannot be absolutely rigid. If ‘atoms’ at all, but potentials or bare pow- they are, then there cannot be any trans- ers; that is, point-centres of mutual influence fer of motion. For a transfer depends upon distributed in space.31 the individuals being deformed to some ex- tent. But basic individuals cannot be de- 87:1 So far I have not discussed one impor- formed. For to be deformed is to lose a tant aspect of our normal concept of action; property once possessed, but atoms have no and that is the idea that things possess pow- properties to lose. Thus the most funda- ers and liabilities32 to do and suffer things mental interactions cannot be by contact be- that they are not actually doing and suffer- tween corpuscles or atoms. But could there ing and that they may never actually do or be non-mechanical, i.e. non-contact, interac- suffer. It remains true to say of a Boeing tions between atoms? If there could, atoms 727 that it can (has the power to) fly 600 might still be the ultimate ‘things’. Now any m.p.h. even if it is safely locked up in its such action would of necessity have to be at hangar; just as it would remain true to say a distance. But aside from spatio-temporal of a person that he could (would be liable to) get hurt if he happened to be in the way of 2.3. THE CLASSICAL PARADIGM OF ACTION 115 a herd of stampeding buffaloes, even if as a ferent. The potentiality involved remains matter of fact we knew he had no intention purely epistemic: it is still predicated es- of ever going to the Prairies. The elucidation sentially of events and only derivatively of of this concept will occupy us much later on, things. To say that a thing has a power so I will be brief with it here. It is suffi- to do something is, by contrast, to say that cient for our purpose to note that it depends it possesses a structure or is of such a kind not only upon the idea of the complexity and that it would do it, if the appropriate con- pre-formation of things, which we found to ditions obtained. It is to make a claim first be necessary to understand change, but upon and foremost about the thing; and only sub- the further idea of the stratification of their sidiarily, if at all, about events. It is to say properties. something essentially about what the thing is, and only derivatively about what it will 87:2/o Now if things are at any moment do. It is to ascribe a natural possibility to of time complex and pre-formed then it al- the thing, whose actualization will depend ways makes sense to suppose that they might upon the flux of conditions. An old Austin 7 have behaved in ways that they did not. But can go 105 m.p.h. if it is towed by a Jaguar; the ‘might’ here is susceptible of a purely but it does not possess the power to do so. conditional analysis, viz. as meaning merely And yet, if it did, the two events if conceived that they would have behaved in those ways, in themselves (without reference to the in- if the actual conjunction of intrinsic and trinsic structure of the thing) would be the extrinsic conditions had in fact been dif- 116 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE same. corpuscularian/mechanical world view was based, viz. between the manifested qualities 88:1 The ascription of powers differs from of things and the configuration and motion the simple ascription of complexity to things of their parts (by which the former were ex- in that it presupposes a non-conventional plained), as one between two types of qual- distinction between those properties of the ities. For, in the first place, though some thing which are essential to it and those properties of the explanatory stratum, such a which are not. The essence of hydrogen is its solidity and motion, could plausibly be mod- electronic structure because it is by reference elled on the properties of observed things to it that its powers of chemical reaction are (so that one might say their ideas ‘corre- explained; the essence of money is its func- sponded’), others, such as a point mass or a tion as a medium of exchange because it is by frictionless surface, had no such analogues. reference to this that e.g. the demand for it Thus Locke did not sufficiently appreciate is explained. Not all properties of a thing are that their concepts could not be read directly equally important because it is by reference onto experience, but rather had to be pro- to some but not others that its causal pow- duced as a result of the theoretical work of ers are explained. In general it is these that science (when they might subsequently come constitute its identity and allow us to talk of to inform and direct experience). Secondly, the same thing persisting through change. Locke should have articulated the distinction 88:2/o Now Locke was wrong to construe as one between the causal powers that mate- the fundamental distinction on which the 2.3. THE CLASSICAL PARADIGM OF ACTION 117 rial things possess in virtue of their hypoth- ical under-statement, implied by his theory esized internal structure and the manifesta- of ideas, of the conceptual advances made by tion of those powers, including their mani- the corpuscularian/mechanical world-view. festation as the qualities of observed things For unless a philosophy of science acknowl- in sense- experience. If he had done this, it edges the existence of a past and the possi- would have obviated the need to find an im- bility of a future, it cannot pay tribute to the mediate reflection in sense-experience of the achievements of the present. hypothesized basic or ‘primary’ properties of the world—a move which provided the grist 89:1/o Science has travelled far since Ba- for Berkeley’s attack on his ontology and con could take it as axiomatic that ‘nature has been the staple diet of phenomenalists knows only mechanical causation, to the in- ever since. At the same time, it would have vestigation of which all our efforts should both left open the possibility of a qualitative be directed’ and Boyle roundly declare that account of primary properties and removed ‘the phenomena of nature are caused by the necessity to view the latter as ultimate. the local motion of one part of matter hit- This would also have undermined the basis ting against another’. Philosophy unfortu- of his scepticism over a knowledge of real nately has lagged behind. We have seen essences. Locke’s pessimism about the possi- how the two basic interactions of the New- bility of future technological and theoretical tonian system contributed—the one by its advance is really the inverse of his philosoph- success, the other by its failure—to the grip of positivistic epistemology. Paradoxically, 118 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE the very breakdown of the system coincided ically ‘secondary qualities’?) And how are with, and indeed contributed to, a revival we to understand the phenomena of diversity of positivistic-sensationalist and operationist and change? Secondly, what does it mean thought. For one thing the new ‘fundamen- to say that the subject matter of one science tal entities’ seemed both event-like and sta- can be ‘reduced to’ (in the sense of ‘explained tistical in character. And the apprehension in terms of’) that of another? How does such of scientific change seemed merely to under- a reduction proceed? And how, if at all, does line the necessity for a subjectivist ontol- it affect the ontological status of the entities ogy. (Later I shall show how, on the con- of the reduced science? trary, scientific change provides in fact the best possible argument for the objectivity 90:1 Finally, something should be said of things.) As for the programme of reduc- about the significance of the psychological tion, historically associated with the corpus- studies of Piaget, Michotte and others for cularian/mechanical world-view, two ques- the classical paradigm.33 These studies show tions raised by it survive the eclipse of that that we ordinarily experience mechanical world-view: First, whatever the nature of causation, i.e. the displacement of physical the basic entities postulated by physics at masses in space and time, in terms of tran- any time, how are we to understand the sta- sitive verbs such as ‘pushing’ and ‘pulling’ tus of apparently emergent things and prop- which cannot be explicated ostensively; but erties? (Are people, for example, metaphys- rather embody an intensional relationship between cause and effect. Such verbs take 2.3. THE CLASSICAL PARADIGM OF ACTION 119 an objective complement: We understand experience (a) because our ordinary concepts ‘pushing’ as ‘pushing away’, ‘pulling’ as cannot be explicated ostensively; (b) be- ‘pulling towards’. Jack does not just fall, cause Newtonian ones cannot either; and (c) he falls down; and the ‘down’ is an essential because the concepts are radically different part of what the ‘falling’ means. Ink bottles anyway. Thirdly, it shows the poverty of the do not only get knocked, they get knocked Humean account of our understanding of the over; doors do not only get slammed, they basic interactions of mechanics. Because we get slammed shut. And this is how we come have to ask why for three hundred years sci- to understand the meaning of these verbs.34 entists and philosophers found this paradigm Now it seems to me that this raises points so compelling. And if today we must resist against Hume, though not against Newton. this compulsion, it is as well to be aware of In the first place, it undermines Hume’s psy- the source of its power. chological account of the genesis of our idea of ‘connection’. For it is now not regu- lar comcommitances, but completed move- ments, transitively understood, that provide the source of the latter. Secondly, it shows that the concepts in terms of which scien- tists come to understand motion could not have been given in or abstracted from sense- 120 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE

2.4 Actualism and naturally or spontaneously, i.e. without the Transcendental Re- active intervention of men? alism: the Inter- 91:4/o One could usefully distinguish here between ‘universal’ and ‘restricted’ closures; pretation of Normic and between ‘artificial’ and ‘spontaneous’ ones. Only a universal closure is consistent Statements with the empiricist concept of a law as a uni- versal empirical regularity. For to say that 91:1 In Section 2 the critical conditions for the antecedent is instantiated in an open sys- a closure have been developed and in Sec- tem is just to say, according to the criterion tion 3 the concept of action implied by them of Section 2 above, that given the antecedent has been brought out. Now two questions of the consequent fails to materialize on at least great significance may be asked of any clo- one occasion in the space-time region for sure: which the system is defined. In general if 91:2 (i) are the conditions for the closure a closure has been artificially established it universally satisfied or is the antecedent of cannot also be universal.35 It is of course pre- the law-like statement for which the closure cisely the ubiquity of open systems in nature is defined instantiated in some open system? that makes necessary an experimental rather 91:3 (ii) were the conditions for the clo- than a merely empirical science. Once this is sure artificially produced or did they occur accepted, the idea of invariance over space- 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS121 time must give way to the idea of invariance tems, i.e. in systems where their consequents under experiment as a criterion of the em- are not invariably realized, the empiricist pirical basis of a science. Moreover, strictly must abandon either the laws or his concept speaking, the invariance is that of a result, of them, viz. as universal empirical general- not a regularity. In general the result will izations. For whatever is empirical must be be invariant to space and time, but not over actual. And in open systems laws if they are them. On the other hand, it is clear that to be actual cannot be universal; and if they if the notion of laws as universal empirical are to be universal cannot be actual. So he regularities is retained, then the same logic must say either that they are not laws; or that led to the regress of interactionism will that laws are not universal; or that laws are lead to the demand for a closure of all in- not empirical. The first position, which may teracting systems until—if everything is as- be characterized as ‘strong actualism, was in sumed to be in interaction—we have what effect adopted by Mill in his doctrine of laws might be called a ‘global’ or ‘Laplacean’ clo- as ‘unconditional sequences’.36 The trouble sure. (Such a slide can only be avoided if it with it is that there are no unconditional se- is supposed that a non-interacting eternally quences known to science. The second po- closed system can be found—without this af- sition, which may be characterized as ‘weak fecting anything in the system.) actualism’, involves restricting the applica- tion of laws to closed systems. This may 92:1 Now confronted with the instantia- be done by making the satisfaction of a ce- tion of the antecedents of laws in open sys- 122 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE teris paribus clause a condition of the law’s cause the things to which laws are ascribed applicability. The trouble with it is that it go on acting in their normal way indepen- leaves unanswered the question of what gov- dently of whether or not a closure obtains erns phenomena in open systems. Moreover that the scientific investigation of nature is it cannot provide a rationale for either the possible. experimental establishment or the practical 93:1 The empiricist, when confronted with application of our knowledge (ironically in the phenomena of open systems, i.e. the view of its sponsorship by self-styled ‘em- non-availability of universal closures, is faced piricists’ and ‘pragmatists’). The third po- with the trilemma of choosing one of the sition is that of transcendental realism. It forms of actualism (which involves either rejects the idea, common to both forms of ac- preserving his philosophical integrity at the tualism, that laws are empirical statements expense of science or abandoning his in- or statements about events. Instead, it re- tegrity to justify science) or succumbing to gards them as normic or transfactual state- transcendental realism. My strategy will be ments that apply in open and closed sys- to argue that weak actualism is not a gen- tems alike. On this view, closures are impor- uine alternative and if pushed must collapse tant in the experimental establishment of our into one of the other two. knowledge. But they do not affect the onto- logical status of laws. On the contrary, the 93:2 One way of describing these options transcendental realist asserts, it is just be- is in terms of their different responses to the identifying mark of an open system, viz. 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS123 the non-realisation of the consequent, given true, namely if the statement has been in- the instantiation of the antecedent of a law- dependently verified (e.g. under experimen- like statement. For the strong actualist this tally closed conditions) and there is no rea- means that the statement must be false, for son to suppose that the nature of the thing the weak actualist it may be inapplicable, possessing the tendency whose operation is for the transcendental realist it can be both described in the law has changed. applicable and true. It must be false for the strong actualist because a law-like statement 93:3/o The weak actualist is immediately asserts the invariance of the conjunction be- faced with a problem here. For although the tween antecedent and consequent. It may be law-like statement may be inapplicable, viz. inapplicable rather than false for the weak if the CP condition was not satisfied, rather actualist, if the ceteris paribus clause, sub- than false, viz. if it was satisfied, there is ject to which it is regarded as being formu- no way on actualist lines that he can de- lated is not satisfied. It can be both appli- cide between these alternatives. The way cable and true for the transcendental realist, in which this is normally settled is to see if it correctly describes the working of a gen- if the consequent is realized; if the conse- erative mechanism and the mechanism was quent is not realized this means that the CP really at work in that instance. Moreover clause was not satisfied. But this involves for the transcendental realist the statement using the law (thus presupposing both its can be known to be both applicable and truth and applicability—the latter in virtue of the satisfaction of the explicitly mentioned 124 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE antecedent conditions) as a criterion of the supposing a law were experimentally verified stability of the circumambient conditions. its use in open systems would presuppose a Hence any attempt to use the stability of general principle sanctioning the applicabil- the circumambient conditions as a criterion ity of laws when their consequents were unre- of the applicability of the law is viciously alized. But this is precisely what is in ques- circular—as the law’s applicability would be tion here and what both forms of actualism already presupposed in the test for the sta- deny. bility of the conditions. The situation in which the weak actualist finds himself has 94:3 Now the strong actualist can only been expressed as follows: – justify the retention of a law-like statement whose antecedent is instantiated in an open 94:1 When a prediction turns out to be system as a temporary proxy or stand-in for false, the situation as regards the general the yet to be discovered unconditionally uni- laws used in making it is indeterminate: it versal statement. But can the weak actu- cannot be known with certainty whether one alist do any better? For we are bound to or all the general laws have been discon- ask him: if laws are restricted to closed sys- firmed or whether the ceteris paribus con- tems what governs or accounts for phenom- dition has not been fulfilled.37 ena in open ones? His options here are lim- 94:2 It might be thought that the situation ited: either nothing does or something does. would be improved if an independent means The former entails complete indeterminism. of verification for the law was available. But But the latter sets the weak actualist on 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS125 the road to strong actualism. For to sup- our ascription of laws we are referring to a pose that something accounts for the phe- way of acting or a level of structure that is nomena and to hold that current laws are in- not confined to closed systems. The neces- applicable only makes sense on the assump- sity to view the satisfaction of the CP clause tion that open systems may be eventually as a condition of a law’s applicability van- closed. So it seems that the incomplete or ishes once we realize that it is precisely a non-atomistic descriptions that we currently key function of the concept of law to apply call ‘laws’ must be replaceable in time by transfactually, in open and closed systems complete atomistic ones which (given only alike. The satisfaction of the CP clause is, that regularity determinism is true) will after on the other hand, a condition for a deci- all be both strictly universal and still empir- sive test situation (its verification depending ical. So that the weak actualist too comes to necessarily upon the applicability of ‘auxil- regard present ‘laws’ as temporary stand-ins iary’ or bridge laws). But the truth of any for the Laplacean hour. normic statement is in general determined quite independently of, and antecedently to, 94:4/o The trouble with weak actualism is its explanatory and other uses in open sys- that it is prepared to acknowledge the fact of tems. open systems without generating the means for science to cope with it; that is, it is pre- 95:1 Given only a knowledge that the an- pared to differentiate but not to stratify re- tecedent is instantiated and the absence of ality, thus removing the possibility that in specific reasons for supposing that the ten- 126 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE dency is no longer possessed by the thing and initiates interactionist and reductionist we can then be justifiably sure that the ten- regresses in an attempt to overcome it. Weak dency is being exercised or as it were in play; actualism acknowledges the de facto exis- although only if we have grounds for suppos- tence of open systems but then proceeds to ing the system closed does that certainty li- fence them off from science. For strong and cense the prediction of its fulfilment. The ci- weak actualism alike, open systems fall out- tation of a law presupposes a claim about the side the pale of science. In this way empiri- activity of some mechanism but not about cism understates its potential scope of appli- the conditions under which the mechanism cation. Lacking from both forms of actual- operates and hence not about the results of ism is the concept of generative mechanisms its activity, i.e. the actual outcome on any which endure, so that the laws they ground particular occasion. This will in general be continue to prevail, in open and closed sys- co-determined by the activity of other mech- tems; so making possible the scientific under- anisms too. Indeed it is precisely because it standing of things and structures which ex- is non-committal about the nature of the cir- ist and act quite independently both of our cumambient conditions that a statement of descriptions and the exercise of our causal law does not in general justify a claim about powers. events, let alone experiences. 96:1 Braithwaite falls into the same trap 95:2 Strong actualism regards the appear- as weak actualism by arguing that a ten- ance of open systems as a mark of ignorance dency statement is a conditional with an un- 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS127 specified antecedent.38 For if it is unspecified applicability of the law is circular. Thirdly, we cannot know when to apply it. It is in fact as the satisfaction of the CP clause is time- vital to distinguish the explicit conditions in dependent (being trivially satisfied instanta- the protasis of the law-like statement from neously), acceptance of it as a condition for the unknown conditions that the CP clause the law’s applicability generates absurd and may be required to cover. The satisfaction totally counter-intuitive results. For exam- of the former is a condition for the appli- ple, on it a law may be applicable for every cability of a law. But neither a knowledge five-minute interval in a day, but not for the (strong actualism) nor the stability (weak ac- day overall. The proper place of the phrase tualism) of the latter can be a condition for ‘other things being equal’ is not as part of the applicability of a law. There are three the protasis but at the tail-end of the state- reasons for this. First, it is in principle im- ment as a reminder that, because the sys- possible to specify all the conditions that the tem in which the thing’s behaviour occurs CP clause may be required to cover. Indeed, may not be closed, the tendency postulated if one could do so, there would be no need in the statement may not be actualized. for the CP clause in the first place. Sec- ond, as has been seen, the satisfaction of the 96:2/o Satisfaction of the CP clause is not CP clause cannot normally be verified inde- a condition for the applicability of a law. It pendently of the actualization of the conse- is, however, a condition for the actualiza- quent; hence to make it a condition for the tion of the tendency designated in the state- ment (for which it is sufficient, although not 128 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE strictly necessary). And from this are de- tions (as suggested in Section 2 above) and rived its two main roles: first and foremost, between the constancy of more and less im- as a signal of the normic nature of the propo- portant intrinsic ones (as suggested in Sec- sition being expressed, as a reminder that tion 3) then the constancy of intrinsic struc- the tendency designated may not be actu- ture is a condition for the applicability of alized; and secondly and derivatively, as a a law. Tendencies are only possessed, and warning to historicists and pseudo-falsifiers, hence can only be exercised, as long as the cautioning the former that the prediction of nature of their possessor remains unchanged. the tendency is not deductively justified and But this does not vitiate my account. For the latter that if the tendency is unfulfilled law-like behaviour is predicated essentially of the statement should not – on that ground things, which are typically referred to in the alone – be held to have been falsified. Thus protasis. There is a real asymmetry, which the CP clause does not place a condition on is reflected in the structure of law-like state- explanation, for one can explain an event in ments, between the intrinsic structure or es- terms of tendencies when the latter are never sential nature of a thing (which in general realized. Rather it places a condition on pre- constitutes its identity or fixes it in its kind) diction and falsification. and the conditions under which it acts in that a change in the former but not the latter 97:1 This account needs qualifying in two leads to a change in the thing’s tendencies, ways. First, if we distinguish between the constancy of intrinsic and extrinsic condi- 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS129 liabilities and powers. more detail the character of normic state- ments. A full analysis of the logic of ten- 97:2 Secondly, I have said that the CP dency statements must, however, be post- clause functions as a reminder and a warn- poned until Chapter 3. ing. But such reminders and warnings are only necessary as long as law-like statements 97:4/o On the view of science advanced continue to be formulated and thought of in here, power and tendency statements are the actualist mode. If there were no histori- categorical rather than, as maintained by cists or pseudo-falsifiers there would be no Hume and Ryle, hypothetical. Hypotheti- need for reminders to them. Hence a fully cals provide the empirical grounds for our as- realist philosophy of science could in princi- criptions of powers and tendencies, but they ple dispense entirely with the CP clause (at do not capture their meaning. Tendencies least in this aspect of its work).39 For what- are roughly powers which may be exercised ever is conveyed by ‘This happens CP’ can be unfulfilled. They are thus well adjusted to equally well conveyed by ‘This tends to hap- cope with open systems. If a system is closed pen’. (To add CP to this statement would be then a tendency once set in motion must be to qualify the tendency, not its fulfilment.) fulfilled. If the system is open this may not This is not a shallow, equivocal, sloppy or happen due to the presence of ‘offsetting fac- mean formulation; but the logical form of all tors’ or ‘countervailing causes’. But there the laws of nature known to science. must be a reason why, once a tendency is 97:3 I want to turn now to consider in set in motion, it is not fulfilled; in a sense in 130 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE which it would be dogmatic to postulate that systems; what does differ is the inference there must be a reason why the tendency is that can be drawn from our knowledge of set in motion. Once a tendency is set in mo- the applicability of the statements in the two tion it is fulfilled unless it is prevented. cases. Notice that although the application of a normic statement warrants a subjunc- 98:1 The following is my interpretation tive conditional about what would have hap- of the mode of application of lawlike state- pened if the system were to have been closed, ments. Such statements, when their ini- the full force of its meaning cannot be un- tial conditions are satisfied, make a claim derstood or captured in this way. It has to about the activity of a tendency, i.e. about be interpreted categorically and indicatively the operation of the generative mechanism to the effect that a generative mechanism that would, if undisturbed, result in the ten- was really at work; which helps to account dency’s manifestation; but not about the for, though it does not completely determine, conditions in which the tendency is exer- whatever actually happened. cised and hence not about whether it will be realized or prevented. Because the op- 98:2/o The ‘thing’ which possesses the eration of the generative mechanism does tendency is not necessarily the same ‘thing’ not depend upon the closure or otherwise as that whose behaviour is recorded in the of the system in which the mechanism op- law-like statement. Indeed it is character- erates, the mode of application of law-like istic of science to postulate novel entities statements is the same in open and closed as the bearers of the tendencies and powers 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS131 manifest in the behaviour of observed things. unfulfilled two things are not in doubt: (a) The class of ‘things’ is far wider than that of that something actually happens, towards ‘material objects’: it includes fluids, gases, explaining which the exercise of the tendency electronic structures, fields of potentials, ge- goes some way; and (b) that something is re- netic codes, etc.; so we must try to divest the ally going on, i.e. there is a real generative concept of its normal material object conno- mechanism at work, which accounts for the tations. The idea of a tendency exercised influence of the factor the tendency repre- unfulfilled seems strange if we think of or- sents in the generation of the event. In the dinary material objects such as tables and case of (a) there are two conceptual traps. chairs. People provide in this respect a bet- The first is to think of the exercise of the ter model for the entities discovered and in- tendency unfulfilled as an action without re- vestigated by science. There is nothing mys- sults, rather than as an action with modified terious about tendency ascriptions to people. results. Something does happen; and the We know what it is like to be in a situa- tendency, as one of the influences at work, tion where we tend to lose our patience or helps to explain what. The second is to temper and we know what it is like keeping think of it as if it were an action fulfilled, it. Tendencies exercised unfulfilled; shown, i.e. in terms of its fulfilment. It is a mis- perhaps, but unrealized in virtue of our self- take to think of the exercise of a tendency control. in terms of the imagery, metaphors or de- scriptions appropriate to its fulfilment. Yet 99:1/o Now when a tendency is exercised 132 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE

Mill in his unofficial doctrine of tendencies an exact balance of buying and selling; when in effect does this when he argues that ‘al- the beam finally collapses it is due to the though two or more laws interfere with one real cumulative effect of the woodrot. Mill’s another, and apparently frustrate or modify mistake is to think of the exercise of the ten- one another’s operations, yet in reality all dency under the description of its fulfilment, are fulfilled, the collective effect being the as if Balaam’s ass, in order to be pulled two exact sum of the causes taken separately’.40 ways, had actually to go in both directions. Mill’s mistake here is to suppose that when- Geach’s mistake is to suppose that because ever a tendency is set in motion the effect neither tendency is fulfilled neither tendency must be in some sense (or in some realm) oc- can be in play. In other words, they both curring (as if every time we ran fast we had make the mistake of seeing the fulfillment of to be in some way winning). But Geach (and a tendency a condition of its exercise. following him Ryan) in ridiculing this posi- tion make the converse mistake of supposing 100:1 Let me stress that the scientist’s sit- that whenever no effect (of a given type) oc- uation is such that he is never in any doubt curs, nothing can be in motion or really go- that the given an effect something is produc- ing on.41 But here Mill is right and Geach is ing it; his doubts is only over what is. Now wrong. Balaam’s ass is pulled in two ways; clearly this does not mean that he is commit- we do just manage to keep our tempers; the ted to a realist interpretation of every theory; market equilibrium is explained in terms of what it does mean is that as a theorist his task remains essentially incomplete until he 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS133 has produced a theory which correctly de- it cannot be assumed that all theoretical scribes the mechanisms by means of which statements are idealizations in this sense. A the effect in question is produced. It is in model of the intrinsic structure of an atom this light that other possible interpretations or a DNA molecule or the solar system is of normic statements must be considered. not necessarily more perfect than the intrin- sic structure of a real atom, DNA molecule 100:2/o It is misleading to think of normic or solar system. The standard of perfection statements as ‘idealizations’ or ‘abstrac- is not set by men. Of course if one takes ‘the- tions’. For both concepts conceal a crucial oretical’ as a synonym for ‘unreal’ (or at any ambiguity as to the object idealized or ab- rate ‘less real’) normic statements will ap- stracted from, in which the superior reality pear as ‘ideal’ in that the tendency they des- of events or experiences is tacitly assumed. ignate or mechanism they describe is rarely The conception of the generative mechanism if ever manifest in unmodified form;42 and or structure that backs a normic statement as ‘abstract’ in that they select from what need not be ‘idealized’ or ‘abstract’ in rela- is in open systems a mesh of influences and tion to really existing or the reality of exist- cross-influences just one as the focus of at- ing structures. Once the necessity for a re- tention.43 But to think like this is to fall definition of the objects of a science as struc- into the error of supposing that events are tures rather than events is accepted then the more real than the structures and mecha- concept of an idealization must be used in relation to the reality of that object. And 134 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE nisms that generate them. if one cannot be sure of the realization of the consequent then one can only assert the law- 101:1 Scriven makes a similar mistake like statement ‘guardedly’. But of course in in contending that normic statements are asserting a normic statement one is not as- ‘guarded generalizations’.44 The only thing serting the realization of the consequent; but one need be ‘guarded’ about in using a the operation of a mechanism irrespective of normic statement is the assumption that the its results (which it is precisely the function tendency whose activity is designated in the of the normic statement to be non-committal normic statement will be realized. If such about). statements have been independently and well confirmed (under experimentally closed 101:2/o Both these ideas depend upon an conditions) then we may be completely and implicit recognition that reality is differen- rationally confident in using them. Such con- tiated in a way that classical empiricism ig- fidence is expressed in, rather than weak- nores and so requires something more of sci- ened by, our willingness to use the CP clause ence than it provides. But the possibility against naive actualist objections on their opened up by this recognition is constrained behalf. It is only if one tacitly views law-like by a continuing commitment to empirical re- statements as in the final analysis empirical alism. It is this which prevents the acknowl- generalizations that one will feel that (be- edgement that reality is not only differenti- cause in asserting a law-like statement one is ated but stratified too. Once the stratifica- asserting the realization of the consequent), tion of the world is grasped it is possible to 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS135 see how our knowledge can be both univer- normic statements must be guarded against. sally applicable and rarely (empirically) in- Normic (or transfactual) statements are not stantiated; and so to resolve Poncare’s prob- counterfactual statements. They legitimate lem that ‘on the one hand, [laws] are truths the latter; and, like them, are only validat- founded on experiment and approximately able in relation to an antecedently and inde- verified so far as concerns isolated systems. pendently established body of theory. But On the other hand, they are postulates ap- whereas to say that a statement is a counter- plicable to the totality of the universe and factual is just to say that the conditions spec- regarded as rigorously true’.45 Normic state- ified in the antecedent do not obtain; in the ments speak of structures not events, the case of a normic statement these conditions generator not the generated. In asserting may obtain, and if they do (and the state- a normic statement one is not making a ment has been independently verified) it can guarded or idealized statement about an em- then be interpreted quite straightforwardly pirical reality. Rather one is making a state- as a statement about what is really going ment, which may be ‘guarded’ or ‘idealized’ on though in a perhaps unmanifest way. (In in its own right, about a different level of re- the case of counterfactuals antecedents are ality. Normic statements are not second best by definition unsatisfied; in the case of trans- kind of empirical generalizations. They are factuals it is contingent whether consequents not empirical statements at all. are realized.) It is only if the CP clause is re- garded as a component of the protasis that 102:1 Two further misinterpretations of 136 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE it is plausible to interpret a normic state- plicable (a mechanism is at work) if we know ment as making, in its open systemic uses, that its antecedent has been instantiated and a counter-factual claim. This is a postion, it has been in dependently verified. But both most naturally associated with weak actu- antecedents and consequents are events in alism, that has been argued against above. open systems. Is there not an asymmetry Normic statements have also sometimes been here? Am I not placing a higher demand justified as ‘averages’ or ‘rough approxima- on antecedents than consequents? Ontologi- tions’; or alternatively as elliptical probabil- cally no; but epistemically yes. For a mech- ity statements. Both ideas involve a con- anism may be set in motion and because of fusion of epistemic and natural possibility. the complexity or opacity of the conditions For, on the one hand, I may be quite cer- under which this happens the describer may tain about the activity of a natural mecha- not know that it has been set in motion; so nism on a particular occasion but incapable that a fortiori he cannot know that the law it of any judgement about the outcome; and, grounds is applicable. To explain an event by on the other, I may be sure that some rule invoking a law I must have grounds for sup- of thumb will hold though quite uncertain posing that a mechanism is at work; but the about the reasons why. mechanism may be at work, given that its stimulus and other conditions are satisfied, 102:2/o I have argued that in open sys- without my knowing it. Some fields may be tems consequents may be unrealized but that despite this we may know that a law is ap- 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS137 incapable of detection. is active. But it is only if I have grounds for supposing that the system in which the 103:1 In section 2 the critical conditions mechanism acts is closed that the prediction for a closure were developed and in section of the consequent event is deductively justi- 3 the concept of action implied by them was fied. With this in mind let us return to the brought out. In both cases their restricted- theories expressed in statements (i)-(v) on ness was noted. In this section a realist ac- pages 634 above. It is only under conditions count of laws has been counterposed to the of a closure that given the antecedent, the actualist account and its superiority clearly deduction of the consequent event is possi- demonstrated. Once we are persuaded of the ble, so that the conditions for the Popper- very special conditions presupposed by actu- Hempel theory of explanation are satisfied alism and the possibility of an alternative, (ii) or those for the symmetry between ‘ex- what havoc must we make of the doctrines planation’ and p‘rediction’ obtain (iii). It of orthodox philosophy of science? is only then that ex ante criteria of refu- 103:2 In nature, constant conjunctions are tation can be laid down for a theory (iv) the rare exception; not, as supposed by ac- or that it makes sense to judge a theory tualism, the universal rule. And in general by its predictive success (v). For it is only it requires human activity to generate them. then that the resemblances and sequences To invoke a law I must have grounds for between phenomena, that Mill identified and supposing that the antecedent conditions are satisfied, so that the mechanism designated 138 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE so confused with laws, are constant (i). epistemic significance. We will also be in a better position to understand not just this 103:3/o It is contingent whether some en- failure, but their absurdity, when they seek during thing or mechanism is activated. And to apply these same criteria to fields such as though, given this, it is necessary that a cer- history and the human sciences, where the tain tendency should be ‘in play’, it is con- conditions for even a restricted closure (of a tingent, upon the occurrence of a closure, non-trivial kind) are not naturally and can- whether the consequent of the law-like state- not be experimentally satisfied, and where ment is realized. In short, to know that law the concept of action implied by these crite- is effective I do not need to be in a position ria is patently inapplicable. to predict any event (and, it might be added, vice versa). 104:2 For a closure one each of the sys- 104:1 Now once we have grasped the ubiq- tem, individual and organizational condi- uity of open systems in nature we will be in a tions must be satisfied. Reflection on the better position to understand the embarrass- conditions set out on page 76 above and the ment with which textbooks in the philosophy concept of action implied by them (see (i)- of science gloss over their failure to produce (vi) on page 83) shows the patent absurdity a single law or explanation which satisfy the of trying to apply the constant conjunction criteria they so laboriously develop and de- formula to the domain of social life. Con- fend; a fact which bears eloquent witness to sider the conditions for a closure as applied the non-ability of universal closures of any to e.g. the category of persons. Remem- 2.4. ACTUALISM AND TRANSCENDENTAL REALISM: THE INTERPRETATION OF NORMIC STATEMENTS139 ber that people are individuals, which means based on it, just cannot apply. that they are complexly structured and pre- formed in different ways, so that they will 104:3/o Conversely, it is just because the respond differently in the same external cir- very special conditions for a closure are cumstances (i.e. to the same stimulus). Re- sometimes satisfied in physics and chemistry member too that they are subject to a con- (though they are not normally possible in tinuing flow of contingencies, none of which the other natural sciences—from cosmology can be predicted with deductive certainty. to biology) that accounts for the prima fa- And, without calling into question the ap- cie plausibility of these theories there. But plicability of the classical paradigm (with its the transcendental analysis of experience al- assumption that the stimulus conditions for low us to turn the tables on actualism and action are always extrinsic), that they are en- empiricism here. For it is not given conjunc- gaged in activities such as writing and cook- tions of events (or experiences) but struc- ing, bar billiards and chess, which cannot be tures which are normally out of phase with plausibly analysed in terms of atomistic com- the patterns of events (and experiences) that ponents. In short, where the subjects, condi- emerge from it as the true objects of scien- tions or forms of action are characterized by tific understanding. This raises the question structure, diversity or change, the Humean of whether there are analogous structures at theory of the actuality of causal laws, and work in fields other than physics and chem- ipso facto the theories of science that are istry. If there are, we must bear in mind that it would not even be plausible to mis- 140 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE construe them as empirical generalizations. an incompletely described world of agents.A On the other hand, if we continue to confuse world of winds and seas, in which ink bottles laws and empirical generalizations we shall get knocked over and doors pushed open, in never be able to identify them. which dogs bark and children play; a criss- cross world of zebras and zebra-crossings, cricket matches and games of chess, mete- 2.5 Autonomy and Re- orites and logic classes, assembly lines and deep sea turtles, soil erosion and river banks duction bursting. Now none of this is described by any laws of nature. More shockingly perhaps 105:1 Laws we already know do not describe none of it seems even governed by them. It the patterns of events. But how do they is true that the path of my pen does not vio- stand to the world of our everyday action late any laws of physics. But it is not deter- and of perceived things? mined by any either. Laws do not describe 105:2 Reflect for a moment on the world as the patterns or legitimate the predictions of we know it. It seems to be a world in which kinds of events. Rather it seems they must all manner of things happen and are done, be conceived, at least as regards the ordinary which we are capable of explaining in vari- things of the world, as situating limits and ous ways, and yet for which a deductively- imposing constraints on the types of action justified prediction is seldom, if ever, possi- ble. It seems, on the face of it at least, to be 2.5. AUTONOMY AND REDUCTION 141 possible for a given kind of thing. inconsistent with the thesis of regularity de- terminism which underpins the doctrine of 106:1 Laws then not only predicate ten- the actuality of causal laws, and to which I dencies (which when exercised constitute the must now return. normic behaviour) of novel kinds (or of famil- iar things in novel or limit situations); they 106:2/o So far I have discussed regular- impose (more or less absolute) constraints ity determinism merely as an epistemologi- on familiar things. In this section I want cal thesis to the effect that our knowledge of to reconcile these aspects of laws by argu- the world can be cast in a certain form. But ing that familiar things are comprehensive of course this presupposes that the world is entities which may be controlled by (or sub- such that our knowledge of it can be cast ject to the control of) several different prin- in that form. To deal with regularity deter- ciples at once; and that they may be said to minism I must thus draw out this ontologi- be agents. Laws ascribe possibilities which cal presupposition; i.e. cast the thesis itself may not be realized and impose necessities in ontological form. The main work for this which constrain but do not determine; they has already been done. For I have already ascribe the former to novel kinds and impose shown in sections 2 and 3 that regularity de- the latter on familiar things. These features terminism makes a claim about what would cannot be explained away as an imperfection happen (and the way it would happen) if cer- of knowledge; but must be seen as rooted in tain highly restrictive conditions were sat- the nature of our world. They are therefore isfied. These were, it will be remembered, 142 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE conditions such that if we knew they were not satisfied and/or the constant conjunction satisfied and the constant conjunction for- formula is abrogated. In short, the ontolog- mula was not vindicated, the regularity de- ical untruth of regularity determinism is a terminist would be bound to admit his thesis condition of the possibility of science. refuted. Now regularity determinism’s on- tological claim is simply that the world is 107:1 Close to the appeal of determinism such that these conditions are satisfied and lies the following error: to think that be- his thesis is not refuted. Now of course be- cause something happened and because it cause we can never know that these condi- was caused to happen, it had to happen tions are satisfied we can never refute regu- before it was caused. Now if we take de- larity determinism in this way. But I have terminism to assert that all events are de- also asserted that regularity determinism is termined before they happen and conceive metaphysically refutable. How can this be their determination as lying in the satisfac- done? In the only way open to transcenden- tion of antecedent sufficient conditions for tal realism: that is by showing that if the them then we have a picture of a chain of world were as claimed by regularity deter- antecedent sufficient conditions for events minism science would be impossible. But as stretching back infinitely into the past (as- science is possible (which we know, because suming that conditions can be analysed as as a matter of fact it occurs) the world must events or vice-versa). So if we ask how long be such that either the critical conditions are is an event determined before it actually hap- pens the answer must be at any (i.e. at ev- 2.5. AUTONOMY AND REDUCTION 143 ery) time before it happens. And so if we tured and pre-formed wholes; which may be now take cause in the ordinary sense, we simultaneously constituted at different lev- have the result that every event is deter- els and simultaneously controlled by differ- mined before it was caused (or made) to hap- ent principles. It is because things cannot be pen. At play here are of course two concepts reduced to the conditions of their formation of cause: qua causal agent (cause1) and qua that events are not determined before they antecedent condition (cause2). I am going are caused to happen. This fact accounts to argue that the former is irreducible to for both the temporal asymmetry of causes the latter and essential to science. To say and effects and the irreversibility of causal that something is determined before it has processes in time. And it is because things been caused to happen is either to say that cannot be reduced to atomistic components it can be known before it has been caused1 that when events are caused to happen it is to happen (epistemic determinism) or that it by the thing which acts (i.e. the agent), the has been caused2 before it has been caused1 event being produced in the circumstances to happen (ontological determinism). The that prevail. former depends upon a closure, the latter depends upon the critical conditions for it 107:2/o Now I want to argue that deter- being satisfied. Now I want to argue that minism is ontologically false (it is not true at any (and every) time the world consists that events are determined before they are of things which are already complexly struc- caused to happen, whether in a regularly re- curring or non-recurring way) and epistem- 144 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE

ically vacuous (there are no significant de- same as cause2). Now any refutation of reg- scriptions that satisfy the formula of regu- ularity determinism as an ontological thesis larity determinism). This has the method- must depend upon establishing the auton- ological corollary that the search for such de- omy of things, in the sense of the impossi- scriptions is likely to be unrewarding. (And bility of carrying out the reductions implicit here once again it is necessary to counter- in the vital conditions B1 and C1 of Table pose the investigation of complex preformed 2.1 on page 76 (their being a clear asymme- things to the search for the complete atom- try, for the realist, between the subjects and istic state-descriptions that it is supposed the condition of action, and the constancy would enable us to predict their behaviour.) alternative being recessive). It is here that The only sense in which science presupposes I will pitch my attack. Thus I am not go- ‘determinism’ is the sense in which it pre- ing to argue that if the critical conditions supposes the ubiquity of causes1 and hence were satisfied the constant conjunction for- the possibility of explanations. And the only mula would not be vindicated. Rather, I am sense in which it presupposes ‘regularity de- going to argue that the critical conditions terminism’ is the sense in which it presup- could not be satisfied in any world contain- poses the ubiquity of causes1 for differences ing science. The question of whether or not and hence the possibility of their explana- history would repeat itself is one that need tion. But it is probably better not to use ‘de- not detain us here. A nagging doubt may terminism’ in this way (nb. cause1 is not the remain: surely, it might be felt, in the (very) 2.5. AUTONOMY AND REDUCTION 145 last instance regularity determinism must be of dual (and multiple) control. true. But this is not so. For once we have established an ontology of structures there 109:3 I will argue that for these features is no earthly reason why events should [have to be possible the world must be composed to] be constantly conjoined. There are in- of agents. Agents are particulars which are deed principles of indifference (as we shall the centres of powers. In an incompletely de- see in Chapter 3). But they do not apply, scribed world of other agents powers must be nor is there any reason why they should, to analysed as tendencies. And laws are noth- events, states-of-affairs and the like. ing but the tendencies or ways of acting of 108:1/o In establishing the autonomy of kinds of thing. By an agent 1 mean sim- things I will follow the normal procedure of ply anything which is capable of bringing transcendental realism; that is, I will first about a change in something (including it- analyse some more or less underanalysed fea- self). A hydrogen atom is, in virtue of its ture of science and then ask what the world electronic structure, an agent. For it pos- must be like for this feature to be possible. sesses the power to combine with an atom The feature I am concerned with are two as- of chlorine to produce, under suitable con- pects of scientific laws, viz: – ditions, a molecule of hydrochloric acid. It should perhaps be said at the outset that I 109:1 (i) their normic and non-empirical am not going to refer to quantum mechan- character; and ics in my argument. It seems to me to be 109:2 (ii) their consistency with situations always a mistake, in philosophy, to argue 146 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE from the current state of a science (and espe- unattached to any theory, could be defeated cially physics). In general, I have refrained by a single counter-instance, the truth of the from scoring points against determinism and former is consistent with almost anything actualism which turn on the inaccuracy (or that might happen in the world of material imprecision) of our descriptions or the inde- objects and human beings. For they do not terminacy of our measures. This is because attempt to describe this world; i.e. they can- they do not in general raise important on- not be interpreted as undifferentiated em- tological questions. It is debatable whether pirical generalizations. Rather they must quantum mechanics does—but if it in fact be interpreted as principles of theories—of requires a reinterpretation of the category of physics and chemistry—which tell us some- causality in fundamental physics it will not thing about the way things act and interact be in the Humean direction and can only in the world. As such they specify condi- strengthen the anti-determinist’s hand. tions which we presume are not contravened but rather continually satisfied in the count- 109:4/o I have already discussed (i) at less different actions and interactions of the some length so I will be brief with it here. world, including those of which we have di- Contrast the law of conservation of energy or rect experience. And they are manifest in of mass action with a simple empirical gener- certain impossibilities, e.g. that of building alization like ‘all pillar-boxes are red’ or ‘all a perpetual motion machine. Nevertheless blue-eyed white tom cats are deaf’. Whereas they are principles for which any test would the latter, at least so long as they remain 2.5. AUTONOMY AND REDUCTION 147 require not only fine measurement but closed established laws of mechanics and electric- conditions. As such they are not normally ity, and such aspects are real. Clockwork empirically manifest to us or actually sat- soldiers and robots do not more nearly ob- isfied. (For the scientist this feature ap- serve the laws of mechanics than real peo- pears as a difference between the real or cor- ple. Rather their peculiarity stems from the rected and the actual or measured values of fact that if wound up and left alone their in- the variables he is concerned with.) Thus trinsic structure ensures that for each set of we could say that relative to these vantage antecedent conditions only one result is pos- points, viz. of experience and actuality, these sible. But outside the domain of a closure principles specify levels of deep structure or the laws of mechanics are, as Anscombe has (metaphorically) place conditions on the in- put it, rather like the rules of chess; the play ner workings of the world. is seldom determined, though nobody breaks the rules’.46 110:1 Now it might be said that laws, such as those of mechanics or electricity, do not 110:2/o Closely connected with this fea- describe the world as such, but only those as- ture of laws is their consistency with situ- pects or parts of it which fall within their do- ations of ‘dual control’. A game of cricket main, i.e. the mechanical or electrical aspects is only partially controlled by the rules of of it. But this concedes my point. For one cricket, language-using by those of gram- can only say which aspects are mechanical mar. Chemical reactions are only partially or electrical by reference to the antecedently controlled by Dulong and Petit’s law, black 148 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE bodies behave in all kinds of ways that are the investigation of this distance that I am not specified by the Stefan-Boltzmann law. here concerned. Coulomb’s law does not completely describe 111:1 To say that laws situate limits but the action of charged particles, or Fara- do not dictate what happens within them day’s law all that happens to an electrode. does not mean that it is not possible to com- Similarly the ‘boundary conditions’ for the pletely explain what happens within them. laws of mechanics, the domain within which The question ‘how is constraint without de- they apply, are controlled by the operating termination possible?’ is equivalent to the principles defining a machine.47 Laws leave question how ‘can a thing, event or process the field of the ordinary phenomena of life be controlled by several different kinds of at least partially open. They impose con- principle at once?’ To completely account straints on the type of action possible for a for an event would be to describe all the given kind of thing. But they do not say different principles involved in its genera- which out of the possible actions will actu- tion. A complete explanation in this sense ally be performed. They situate limits but is clearly a limit concept. In an historical do not dictate what happens within them. In explanation of an event, for example, we are short, there is a distance between the laws of not normally interested in (or capable of giv- science and the ordinary phenomena of the ing an account of) its physical structure. world, including the phenomena of our ac- tual and possible experience. And it is with 111:2 In deciding to write ‘!’ on this piece of paper I select the conditions under which 2.5. AUTONOMY AND REDUCTION 149 the laws of physiology and physics are to ap- ing in accordance with a plan or in the light ply. So that it is absurd to hold that the lat- of reasons. ter might account for my ‘!’; or that it might have been predicted in the basis of a knowl- 112:1 Human freedom, on this view, if it edge of a physical state-description prior to exists, would not be something that some- my writing it. On the other hand my neuro- how cheats science (as it is normally con- physiological state and the physical condi- ceived) or, on the other hand, something tions must be such that I can write it; they that belongs in a realm apart from science; could prevent it (e.g. if I were suddenly to but something whose basis would have to be fall asleep or be propelled into orbit around scientifically understood. As freedom would the moon). There is a space between the be ana1ysed as a power of men and science laws of physics and physiology and what I is, for us, non-predictive there is nothing in- do within which deliberation, choice and vol- consistent or absurd about such an asser- untary behaviour have room to apply. The tion; any more than to say that purposeful- theory of complex determination, in situat- ness in animals, which is no doubt not the ing persons as comprehensive entities whose same as intentionality in men, has (still) to behaviour is subject to the control of several be scientifically understood. I suggest that different principles at once, allows the possi- only the theory of complex determination is bility of genuine self-determination (subject compatible with agency; and that there are to constraints) and the special power of act- no grounds for assimilating intentional ac- tion to the classical paradigm or supposing 150 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE that intentionality is not a real attribute of that the higher-order level is open with re- men. However, this is peripheral to my main spect to, in the special sense of irreducible to, concerns here. Dogs cannot fly or turn into the principles and descriptions of the lower- stones, but they can move about the world order level. It is easy to see why this must and bark in all kinds of ways. To deny the be so. For it is the operations of the higher- latter possibility is as absurd as to deny the order level that control the boundary con- former necessity. But the reasons why they ditions of the lower-order level, and so de- behave in canine ways is an open question termine the conditions under which the laws for a putative science of animal ethology to of that level apply. It is the state of the answer. weather that determines, in England, when and where the rules of cricket can apply; the 112:2 The difference between laws of na- state of the conversation that determines the ture and empirical generalizations is anal- ways in which we can express ourselves in ogous to the difference between the rules speech; the state of the market that deter- of cricket and a television recording of the mines the use of machines, the use of ma- actual play on some particular occasion. chines that determines the conditions under Whether or not Boycott scores a century is which certain physical laws apply. The use not determined by the rules of cricket; but of machines is thus subject to dual control: by how he bats and how the opposition play. by the laws of mechanics and those of eco- Now it is clearly necessary for the intelligi- nomics. But it is the latter that determine bility of the idea of dual (or multiple) control 2.5. AUTONOMY AND REDUCTION 151 the boundary conditions of the former. formed, biology would not be otiose. For a knowledge of biological structures and prin- 113:1 It follows from this that the opera- ciples would still be necessary to account for tions of the higher level cannot be accounted any determinate state of the physical world. for solely by the laws governing the lower- Whatever is capable of producing a physical order level in which we might say the higher- effect is real and a proper object of scien- order level is ‘rooted’ and from which we tific study. It would be the task of biolo- might say it was ‘emergent’. Now an his- gists to investigate the causal powers of liv- torical explanation of how a new level came ing things in virtue of the exercise of which to be formed would not, it is important to inter alia they brought about various deter- see, undermine this principle. Let us suppose minate states of the physical world. Living that we could explain the emergence of or- creatures qua causal agents determine the ganic life in terms of the physical and chemi- conditions under which physical laws apply; cal elements out of which organic things were they cannot therefore already be manifest in formed and perhaps even reproduce this pro- the latter. Sentience determines the condi- cess in the laboratory. Now would biologists tions of applicability of physical laws, but lose their object of inquiry? Would living it is also subject to them. If the elements things cease to be real? Our apprehension of of the lower-order are real then so must be them unmasked as an illusion? No, for in as the causes that determine the conditions of much as living things were capable of acting their operation, i.e. the comprehensive enti- back on the materials out of which they were 152 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE ties formed out of them. If black bodies are not change the laws that governed and so real then so are physicists, if charged parti- explained its mode of operation. And I can cles are real then so are thunderstorms. In come, in science, to have a knowledge of such short, emergence is an irreducible feature of normic and non-empirical statements; and our world, i.e. it has an irreducibly ontolog- perhaps in time begin to recognize analogous ical character. principles at work controlling my own be- haviour (marking the site of a possible psy- 113:2/o Reflect once more on the distinc- chology). tiveness of laws of nature and empirical gen- eralizations. The laws of nature leave the 114:1 I have argued that complex objects conditions under which they operate open, are real (and that the complexity of ob- so the field of phenomena is not closed: it is jects is real); and that the concept of their subject to the possibility of dual and mul- agency is irreducible. Complex objects are tiple control, including control by human real because they are causal agents capable agents. What I can do is constrained by the of acting back on the materials out of which operation of natural laws. But I can hack they are formed. Thus the behaviour of e.g. my way all over the physical world, defeat- animate things is not determined by phys- ing empirical generalizations. I can interrupt ical laws alone. But that does not mean the operations or break the mechanism of a that their behaviour is not completely de- machine and so falsify any prediction made termined: only that an area of autonomy is on the basis of its past behaviour. But I can- marked out which is the site of a putatively 2.5. AUTONOMY AND REDUCTION 153 independent science. And because the forms world as we currently know it. And that of determination need not fall under the clas- I have not shown that if we were in fact sical paradigm this in turn situates the pos- able to reduce (apparently) complex things sibility of various kinds of self-determination to complete atomistic state-descriptions that (including the possibility that the behaviour we would be unable to predict future phys- of men may be governed by rational princi- ical states of the world without referring to ples of action). comprehensive entities and principles of be- haviour special to them. The final stage of 114:2 From the normic and non-empirical my argument against actualism must thus nature of laws and their consistency with sit- constitute a critique of strong actualism in uations of dual control I conclude that the which the incoherence of the programme of world is a world of agents incompletely de- reduction it envisages for science is demon- scribed. Laws neither undifferentially de- strated. scribe nor uniquely govern the phenomena of our world. And this is accounted for 114:3/o It is important to be clear about by the fact that it is an incompletely de- the different senses of ‘reduction’. There scribed world of agents which are consti- are three distinct ways in which a science tuted at different levels of complexity and or- might be said to be ‘reducible’ to a more ganization.48 However it might be objected basic one, which ought not to be confused. here that all I have shown is that the laws There is first the idea of some lower-order that we currently possess do not describe the or microscopic domain providing a basis for 154 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE the existence of some higher-order property to see that it is to this claim that the strong or power; as for example, the neurophysi- actualist is committed, if he is to eliminate ological organization of human beings may complex behaviour in favour of its atomistic be said to provide a basis for their power of surrogates. It depends not only upon the speech. There is secondly the idea that one establishment of a complete parallelism be- might be able to explain the principles of the tween the two domains, but upon a closure, higher-order science in terms of those of the i.e. the attainment of a complete atomistic lower-order one. This depends upon being state-description of all the systems within able to undertake at least a partial transla- which the events covered by the descriptions tion of the terms of the two domains; though of the higher-order science occur. it is conceivable that they may retain sub- stantially independent meanings and overlap 115:1/o Now it is especially important to only in some of their reference states. Such keep the second and third senses distinct. a ‘reduction’ may of course result in modi- For though it is clear that we can explain fications of the laws of the higher-order do- the principles and laws of chemistry in terms main.49 There is finally the sense in which of those of physics or of classical mechanics it is suggested that from a knowledge of the in terms of quantum mechanics, we cannot states and principles of the lower-order sci- predict physical and chemical events such as ence we might be able to predict behaviour the next eruption of Vesuvius on the ba- in the higher-order domain. It is important sis of that knowledge alone. For that we would need an antecedent complete atom- 2.5. AUTONOMY AND REDUCTION 155 istic state-description, i.e. a closure, as well. this must be so: for without the specifica- Now the strong actualist, claiming that the tion of some already more or less clearly de- world is in the end closed, must, unless he marcated and well charted domain no pro- is to limit himself merely to a dogmatic re- gramme of reduction could possibly get to assertion of this claim, presumably map out work. But this means that as a means of dis- a strategy for the sciences to attain such a covery, i.e. of achieving such a body of knowl- closure. The fact that a successful reduction edge reductionism must fail. For it presup- in science does nothing in itself to achieve poses precisely what is to be discovered. empirical invariances is something of a blow to the programme (as distinct from dogma) 116:1/o I still have not refuted strong ac- of strong actualism. But even if it did there tualism as a possible account of the world. is an even more damaging objection at hand This I shall now do by arguing that it is in- (which carries a more general moral for all consistent with any world containing science, those who see in ‘reduction’ the hope of the and thus in any world in which science is ‘less developed’ sciences). For every histor- possible. The only way of reconciling exper- ically successful reduction of one science to imental activity with the empiricist notion another has depended upon the prior exis- of law is to regard it as an illusion; that is, tence of an established corpus of scientific to regard the actions performed in it as sub- principles and laws in the domain of the re- sumable in principle under a complete atom- duced science. It is easy to appreciate why istic state-description. In principle this ap- plies not only to experimental activity but 156 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE to all scientific activity (including theory- be of normic form; and this is quite indepen- construction) in as much as it involves phys- dent of our knowledge of them. In short, the ical effects. Now this has the absurd conse- complexity of agents and the normic char- quence that the apparent discovery of nat- acter of laws are irreducible ontological fea- ural laws depends upon the prior reduction tures of the world; that is, they are necessary of social to natural science. Or to put it an- features of our world established as such by other way, in an actualist world there would philosophical be no way of discovering laws which did not already presuppose a knowledge of them. So 117:1 It is relatively easy to show that all a closed world entails either a completed or (and not just scientific) action depends upon no science. But as ‘completion’ is a process our capacity to identify causes in open sys- in time the former possibility is ruled out: tems. For all action depends upon our ca- so a closed world entails the impossibility pacity to bring about changes in our phys- of science. But as science occurs the world ical environment. Hence we must belong must be open. This is not the reason why to the same system of objects (nature) on the world is open (though it is the reason for which we act. But we not only act on it, my justified belief that it is). Rather it is be- in the sense of bringing about changes that cause the world is open that science, whether would not otherwise have occurred; we act or not (and for how long) it actually occurs, on it purposefully and intentionally, i.e. so is possible. In an open world all laws must as to bring about these changes (as the re- sults and consequences of our actions) and 2.5. AUTONOMY AND REDUCTION 157 knowing that we are acting in that way. This the experimental testing of a scientific hy- depends upon our being able to identify fea- pothesis. Human freedom is not something tures of our environment as the objects of our that stands opposed to or apart from science; causal attention and as part of the system to but rather something that is presupposed by which causality applies. Thus we must be ca- it. The idea that freedom is opposed to or pable of identifying and ascribing causes in apart from science stems from the empiricist our environment, and knowing ourselves as a conception of scientific experience as consist- causal agent among others. Unless we could ing in the passive observation of repeated se- do this, we could not act intentionally at all. quences rather than in the active interven- Thus all human action depends upon our ca- tion of men in the world of things in an en- pacity to identify causes in open systems (to deavour to grasp the principles of their be- which of course the Humean theory cannot haviour. Men are not passive spectators of a apply). given world, but active agents in a complex one. 117:2 I suggested earlier that human free- dom is not only compatible with science, but 117:3/o The view of the world as open and had to be scientifically understood. This is the view of the world as closed lead to to- important because it is inter alia a precon- tally different conceptions of science. The dition for science. For science to be possible laws of nature, which are painstakingly un- men must be free in the specific sense of be- covered by the theoretical work of science ing able to act according to a plan e.g. in supplemented wherever possible by experi- 158 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE mental investigation, do not seek to describe applied phases of scientific activity or, per- the myriad phenomena of the world, the con- haps better, between science and its uses. It tents of a biscuit tin or the junk in the is with the second that I will be concerned builder’s yard. They do not seek to trace here. Now consistency with our conception the path of a squirrel, predict which rafter a of the objects of science as the mechanisms sparrow will light on or how many buns the that produce phenomena, not the phenom- vicar will have for tea.50 They can indeed ena they produce (which must now be seen come to explain such things in a certain way, as both complex and differentiated), means but only on the condition that they are not that we must carefully distinguish between interpreted as describing them. two moments of the scientific enterprize (in- terpreted broadly): the moment of theory, in which closed systems are artificially es- 2.6 Explanation in tablished as a means of access to the en- during and continually active causal struc- Open Systems tures of the world; and the moment of its open-systemic applications, where the results 118:1/o The fact that closed systems are a of theory are used to explain, predict, con- presupposition of the actualist account of sci- struct and diagnose the phenomena of the ence is reflected (a) in the absence of a the- world. Actualism cannot sustain this dis- ory of their establishment and (b) in the ab- tinction; or, if we confront it with it, show sence of a clear contrast between pure and 2.6. EXPLANATION IN OPEN SYSTEMS 159 how the practical application of our knowl- phenomena with which they have to deal. edge is possible in open systems. This de- But we can place a great deal of rational pends upon precisely the same ontological confidence in their ex post explanations. For distinction as is necessary to sustain the in- the law-like statements they use to retrodict telligibility of experimental activity, namely the antecedent events and states by means that between causal laws and the patterns of of which they both explain what actually phenomena, the mechanisms of nature and happened and excuse their forecasts of it the events they generate, the domains of the are not meteorological laws. So that me- real and the actual. In this way actualism’s teorology is in this sense not a theoretical assumption of an undifferentiated reality is science. Rather, mentioning general physi- mirrored in the assumption of an undifferen- cal variables, they are physical laws which tiated science. have been confirmed quite independently of their use to explain and predict the weather. 119:1 It is because of this ontological dis- Thus meteorology, like engineering, stands tinction that theory is never disconfirmed by to physics and chemistry as an applied to the contrary behaviour of the uncontrolled a pure science, using the experimentally- world, where all our predictions may be de- established results of the latter. (I am not feated. Meteorology provides an instructive ruling out the possibility that there may be example here. We can have very little confi- irreducibly meteorological principles.) dence in the ex ante predictions of weather forecasters, because of the instability of the 119:2 Now it is characteristic of open sys- 160 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE tems that two or more mechanisms, perhaps distinct kinds of events, but as differentiat- of radically different kinds, combine to pro- ing distinct kinds of mechanisms. For in the duce effects; so that because we do not know generation of an open-systemic event several ex ante which mechanisms will actually be of these predicates may be simultaneously at work (and perhaps have no knowledge of applicable. their mode of articulation) events are not de- ductively predictable. Most events in open 120:1 The skills of an applied and a pure systems must thus be regarded as ‘conjunc- scientist are characteristically different. The tures’. It is only because of this that it applied scientist must be adept at analysing makes sense to talk of a stray bullet or an a situation as a whole, of thinking at several unhappy childhood affecting ‘the course of different levels at once, recognizing clues, history’. And it is only in virtue of this piecing together diverse bits of information that laboratory closures can come to be es- and assessing the likely outcomes of various tablished. The importance of experimental courses of action. The pure scientist, on the activity in natural science, conceived as a other hand, deliberately excludes, whereas specific kind of conjunctural occurrence, al- the applied scientist seeks always to accom- lows us to stress that the predicates ‘natu- modate, the effects of intervening levels of ral’, ‘social’, ‘human’, ‘physical’, ‘chemical’, reality. Though he is unafraid of flights of ‘aerodynamical’, ‘biological’, ‘economic’, etc. daring (always risky for the practical man), ought not to be regarded as differentiating he holds fast to his chosen objects of inquiry. The applied scientist is an instrumentalist 2.6. EXPLANATION IN OPEN SYSTEMS 161 and a conservative, the pure scientist a real- needed to penetrate it. Needless to say, how- ist and (at the highest level) a revolutionary. ever, that stratification cannot justify any Keynes had the rare gift among economists particular institutionalized form or any so- of knowing both how to make money and cial division e.g. between scientists and non- how money is made.51 scientists (the educator and the educated) arising from the latter. Secondly what pri- 120:2/o I said in section 1 that the ac- marily distinguishes scientific from lay expla- tivities of explanation, prediction and the nations of events is not their structure but identification of causes not only do not pre- the concepts that figure in them. Thus the suppose a closure, but they do not neces- role played by laws in the scientific expla- sarily involve, though they may make use nation of events, a role which is played via of laws. There are two points here. First, the invocation of the concept of the mecha- there is a difference in general between sci- nism at work in the generation of the event, entific and lay explanations. That this is so which is the function of the citation of the is entailed by one of the most obvious fea- law (and which will be discussed further in tures of science, namely the prolonged period the next chapter), is paralleled in lay expla- of scientific education and training a novice nations by other kinds of normic statements must normally undergo before he is consid- such as platitudes, truisms, assumptions of ered capable of ‘scientific explanation’. This rationality or more crudely or vaguely formu- has a rationale in the real stratification of lated law-like statements. Moreover, there the world and a real effort, which is science 162 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE is a case, which I am now going to exam- explain why the door is open but it does so ine in some detail, in which both scientific only contingently. On the other hand ‘Tania and lay explanations have exactly the same observed the door open’ cannot explain why form and in which they do not involve normic the door is open because there is no conceiv- statements at all. This is the transitive verb able way in which observing can bring about model to which I have already alluded in sec- a change in the object concerned (viz. the tion 3 above. state of the door). Now the role of the verb ‘push’ in ‘Tania pushed the door open’ is to 121:1 ‘Why is the door open?’—‘Because link the A-sequence and the B-sequence in Tania pushed it open’. The door is open be- Diagram 2.1 by supplying an interpretation cause Tania pushed it [open]’ is a paradigm of the latter, so that the door’s movement causal explanation, accomplished without can be seen as the result of a continuous reference to laws, by the redescription of the action sequence. Note that though ‘Tania explanandum event in terms of its cause. It moved up to the door and then the door is informative—there are other reasons why moved away’ is a true description it does not the door might be open. But it is also mean the same as ‘Tania pushed the door logically necessary; i.e. the explanation is open.’ deductive—if Tania pushed the door open, it must be open. In this ‘Tania pushed the door open’ differs from ‘Tania pushed the door 121:2 A Sequence T —–¿–+ — Door — B hard’. ‘Tania pushed the door hard’ may Sequence +———¿—- 2.6. EXPLANATION IN OPEN SYSTEMS 163

Diagram 2.1 by-contact was not itself felt to be in need of explanation. In neither case is there refer- 121:3/o Now transitive verbs such as ence to laws or any other general statements. ‘pushing’, ‘pulling’, ‘knocking’, ‘twisting’, ‘Juanita made Xara push the door open’, ‘binding’, ‘squeezing’, ‘holding’, ‘forcing’, ‘The mixture made him sick’, ‘He drove his ‘driving’, ‘turning’, ‘stimulating’, ‘produc- wife to despair’, ‘The sergeant forced him ing’, ‘generating’, ‘bringing about’, ‘making’, to pull the trigger’, ‘The elephant crashed etc. lie at the root of our notion of cause.52 into the juggernaut’, ‘The first billiard ball When something is cited as a cause it is being smacked into the second’, ‘The irate posi- viewed as that element, paradigmatically an tivist knocked his ink bottle over’, ‘The psy- agent, in the total situation then prevailing choanalyst suggested he open the window’ which, from the point of view of the cause- these are the primaeval explanation forms. ascriber, ‘so tipped the balance of events as It has been suggested that it is the fact that to produce the known outcome’.53 Now the something is subject to human manipulation importance of the transitive verb model is or control that accounts for our identifica- that it accounts for both the large number tion of it as the cause.54 But apart from ob- of ordinary causal explanations which are de- vious counter-examples, it is clear that we ductive (or become so with the addition of a could only know ourselves as causal agents suitable objective complement, perhaps tac- in a world of other causal agents and that itly understood) and the basic interactions of our notion of cause takes in the possibility classical mechanics; i.e. the fact that action- 164 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE of a world without men. It is because men ‘condensation’ or ‘distillation’ of its compo- are agents, not because ‘other agents’ have nent causes. affinities with men, that the concept of cause 122:2/o I now want to illustrate this by would still find application in such a world. looking at a fairly typical piece of histori- cal narrative. This will also enable me to 122:1 Now if most events in open sys- identify some more general characteristics of tems are conjunctures, i.e. are to be ex- explanation in open systems. In the piece of plained as the results of a multiplicity of narrative that follows I underline obviously causes, to the extent that basic causal ex- causal notions. planations are involved, one would expect a modification of the transitive verb model 123:1 This pressure from the Labour to be necessary, corresponding and similar Party, with its great influence on the in- to that which required a restatement of the dustrial workers, combined with the attitude nomological model in normic form. This is of President Wilson himself, slowly propelled so. For if a single influence was responsi- Lloyd George in the direction of the formu- ble for the outcome the event could be seen, lation of war aims. Hindered as he was by as in Diagram 2.1, as the simple pure linear the obligation of earlier agreements with the displacement of its cause (and deducibility European allies, he ensured that his decla- would be preserved). To the extent how- ration, made on the 5th January 1918, was ever that more than one factor is at work only in the vaguest terms. It was, how- the event will have to be seen as a kind of ever, not incompatible with the much more 2.6. EXPLANATION IN OPEN SYSTEMS 165 specific Fourteen Points enunciated indepen- brought about the event and in rendering dently by the American President a few days intelligible their efficacy can be seen. Why later, and appealed by the German Govern- did Lloyd George formulate his war aims? ment as a basis for peace negotiations at the Because of pressure from the Labour Party time of the armistice in November.’55 and from President Wilson. Here we imagine the event as if it were a simple displacement. 123:2/o The first thing to notice about But now the simple displacement is modi- this piece of historical narrative is its de- fied by the effect of another factor, viz, his centralized focus, allowing the emergence, in previous obligations, and so Lloyd George a series of redescriptions of the event con- formulates his war aims vaguely. The event cerned, of a picture of the conjuncture or becomes a condensation of the different ex- balance of forces in which it occurred and planatory linkages. Thirdly, each of these in terms of which it is explained. The event individual linkages could in principle be lo- is in fact known under three different de- cated within some interpretative schema or scriptions: Ea, Lloyd George’s formulation theoretical structure. But it is simple dis- of his war aims; Eb, his vague formulation placements, transitively understood, and the of these war aims; and Ec, his vague yet role that causal notions play in them that ex- compatible (with the Fourteen Points) for- plains the peculiar efficacy of what Dray has mulation of these war aims. Secondly, the called ‘continuous series’.56 Finally, the non- indispensable role that causal notions play unified ontology of the explanation should be in both indicating the key variables which 166 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE noted. The industrial proletariat and Pres- get pushed open and it is perfectly legitimate ident Wilson’s attitude co-exist within the to talk in this way. What is illegitimate is to same explanation. The pattern of the expla- regard corpuscles as acting like doors. (For nation is illustrated in Diagram 2.2. if the door was a corpuscle it could not re- tain its shape—it would have to be bent to 124:1 A possible misunderstanding must be ‘opened’.) A puzzle may arise about pre- be avoided and a possible puzzle allayed. cisely what event is being explained in our The physical action causal notions used simple historical explanation, when the same in the explanation of such an event are event is referred to under three different de- of course employed metaphorically. Lloyd scriptions. But the puzzle dissolves when it George is not literally propelled. In this is realized that the phrase ‘the event which way they stand in for what some would say occurred (in si at tj )’ is essentially syncate- are trivial, though I would prefer to say are gorematic; that is to say that it refers only (as yet) inadequately understood, processes. on the basis of some prior description of the Now this differs from the kind of criticism event concerned. And it is precisely the func- that I directed against the action-by-contact tion of the notion of an event to generate re- paradigm when I argued in section 3 that descriptions of events as specified under their though it may provide the source of our con- original descriptions in their explanation. In cept of causality, it cannot provide an ad- this way it also acts as a possible signpost equate model for the understanding of ul- timate physical actions. For doors do really 2.6. EXPLANATION IN OPEN SYSTEMS 167 into the language of theory. tion from redescribed component events or 124:2 states to the antecedent events or states of affairs that could have produced them. To graph commented out the extent that for each determinate effect Diagram 2.2 there is a plurality of possible causes retro- 125:1 I have taken a simple historical ex- diction alone cannot be decisive. And so it planation because it illustrates some more will need to be supplemented by indepen- general features of explanation in open sys- dent evidence for the antecedents until we tems. The pattern of explanation, even have eliminated from the total set of possible where well-developed scientific theory can be causes all but the one which, together with brought to bear on an event, is substan- the other factors at work, actually produced tially the same. In general as a complex the effect on the occasion in question. The event it will require a degree of what might four stages in the explanation of an open- be called ‘causal analysis’, i.e. the resolution systemic event may therefore be summarized of the event into its components (as in the as follows: (i) causal analysis (or resolution) case above). These components will then re- of the event; (ii) theoretical redescription of quire theoretical redescription, so that the the component causes; (iii) retrodiction via theories of the various kinds of mechanism normic statements to possible causes of the at work in the generation of the event can components; (iv) elimination of alternative be brought to bear on the event’s explana- tion. The next step will consist in retrodic- 168 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE causes. event entails a potential prediction of it de- pends upon the possibility of the reduction 125:2/o Now it is particularly important of the various sciences to a single level and to beware of the supposition that if we have a complete description of all the individuals achieved such a complete explanation of an at that level; i.e. it depends upon the idea event (normally of course we will only be of an antecedent closure. Now it is not that interested in one or two of the influences this represents an unreasonable ideal for sci- at work) this would put us in a position ence; but rather that it constitutes a conjec- whereby we could have predicted it. For the ture about the nature of the world which is different levels that mesh together in the gen- in fact false and which, if acted upon, could eration of an event need not, and will not have the most deleterious effects on science. normally, be typologically locatable within If science is to be possible the world must be the structures of a single theory. In general open; it is men that experimentally close it. the normic statements of several distinct sci- And they do so to find out about structures, ences, speaking perhaps of radically different not to record patterns of events. kinds of generative mechanism, may be in- volved in the explanation of the event. This does not reflect any failure of science, but the complexity of things and the multiplicity of forms of determination found in the world. The idea that a complete explanation of an 2.7. FOOTNOTES TO CHAPTER 2 169

2.7 Footnotes to Chap- stant conjunctions of atomistic events. 9 P. S. de Laplace, A Philosophical Essay on ter 2 Probabilities, p. 4. Positivists still pay obeisance to this concept of knowledge. See e.g. Brodbeck’s 1 J. S. Mill, Auguste Comte and Positivism, p. characterisation of it, in an echo of Laplace, as 6. ‘perfect knowledge’ (M. Brodbeck, ‘Methodolog- 2 K. R. Popper, The Logic of Scientific Dis- ical Individualism: Definition and Reduction’, covery, p. 59. Readings in the Philosophy of the Social Sci- 3 C. G. Hempel, Aspects of Scientific Expla- ences, p. 289); and Hempel’s wry admission that nation, p. 366. classical deterministic (i.e. Laplacean) systems 4 K. R. Popper, Conjectures and Refutations, conform ‘best’ to his model of explanation as p. 39, n. 3. deductive subsumption under universal laws (C. G. Hempel, Aspects of Scientific Explanation, p. 5 J. Gibbs and W. Martin, Status, Integration 351). It is but a short step to identifying such and Suicide, p. 197. systems with theories. (See e.g. M. Brodbeck, 6 See e.g. N. R. Hanson, Observation and Ex- op. cit., p. 288; C. G. Hempel, ibid; and R. Rud- planation, p. 45. ner, The Philosophy of Social Science, p. 91). 7 M R. Ayers, The Refutation of Determin- 10 P. S. Laplace, ibid. ism, p. 6 and passim uses the term ‘actualism’ to refer to the doctrine that only the actual is 11 The concept ‘event’ functions here syncate- possible. The connection between the two con- gorematically. Its purpose is, in context, to gen- cepts will become clear in due course. erate the appropriate redescriptions of the events 8 Leibniz’s pre-established harmony of mon- concerned. ads may be usefully compared with Hume’s con- 12 See e.g. W. Kneale, Probability and Induc- 170 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE tion, p. 60. losophy: to the same natural effects we must, as 13 See e.g. J. S. Mill, A System of Logic, Vol. far as possible, assign the same natural causes’, I, Bk. III, Chap. 3, Sect. 3, or A. J. Ayer, Foun- ibid. And Hume: ‘the same cause always pro- dations of Empirical Knowledge, Chap. 4, Sect. duces the same effect, and the same effect never 17. As has been frequently pointed out, by Mill arises but from the same cause’, A Treatise on and Ayer among others, this concept does not Human Nature, p. 173. The rationale for this re- accord well with our normal usage; so in prac- quirement lies in the counter- intuitive nature of tice the Humean tends to modify it in the direc- the implication that the future be better known tion of the intelligibility concept by making the than the past. Moreover given the logical re- cause an individually critical factor in a jointly versibility of the connective and the classical sufficient set. On the other hand, to the extent concept of time a disjunctive plurality of causes that the intelligibility theorist is committed to could be transformed into a disjunctive plurality the doctrine of empirical realism he must rely of effects so as to produce a radical indetermin- on a background of empirical generalizations to ism in nature, i.e given x then y1 or y2 . . . or justify his citation of the cause. yn! 14 Cf. e.g. E. Nagel, The Structure of Science, p. 334. 18 Regularity determinism does not make a 15 See e.g. P. Gardiner, The Nature of His- claim about the constancy of conditions. Its torical Explanation; W. Dray, Laws and expla- claim concerns the constancy of the conjunction nation in History; C. B. Joynt and N. Rescher, between conditions, whether conditions should ‘The problem of Uniqueness in History’, History happen to be changing or not. and Theory, Vol. 4; and M. Scriven, op. cit. 16 Cf. M. Bunge, Causality, p. 122. 19 Cf. J. R. Aronson, ‘Explanation with- 17 Cf. Newton’s 2nd Rule of Reasoning in Phi- out Laws’, The Journal of Philosophy, Vol. 66, 2.7. FOOTNOTES TO CHAPTER 2 171

(1969), pp. 541–57. and of our sea.’ I. Newton, op. cit, Book III, 20 See e.g. A. Koyre, op. cit., Chaps 1–2 General Scholium. 21 W. Heisenberg, Physics and Philosophy. 28 See e.g. P. K. Feyerabend, ‘Problems of 22 T. Hobbes, On Human Nature, reprinted Empiricism’, Beyond the Edge of Certainty, ed. in Body, Man and Citizen, ed. R. S. Peters. R. G. Colodny, esp. pp. 168–70, and T. S. Kuhn 23 Cf. M. Bunge’s genetic principle or prin- op. cit., esp. pp. 99–103. ciple of productivity: ‘nothing can arise out of 29 See e.g. M. Capek, op. cit., Pt. II. nothing or pass into nothing’, op. Cit., pp. 25– 30 R. J. Boscovitch, A Theory of Natural Phi- 6. Cf. also Kant’s 1st Analogy of Experience, losophy, pp. 10–13 and 19–68 and I. Kant, The op. cit.; and W. H. Walsh, ‘categories’, Kant- Metaphysical Foundations of Natural Science. studien Band 45 (1954), reprinted in Kant, ed. 31 Cf. R. Harre, op. cit., p. 308. R. P. Wolff, pp. 54–70. 32 By a ‘liability’ I mean simply what Hobbes 24 Cf. S. K¨orner, ‘Substance’, P.A.S. Supp. called a ‘passive power’. Vol. 38 (1964), pp. 79–90. 33 See J. Piaget, The Child’s Perception of 25 Cf. M. Capek, op. cit., p. 54. Physical Causality, and A. Michotte, The Per- 26 see e.g. I. Dirac, Is there an aether?’, Na- ception of Causality. ture 168, pp. 906–7 34 Cf. J. R. Aronson, op. cit., pp. 551–5. 27 Cf: ‘Hitherto we have explained the phe- 35 However a closure might be both univer- nomena of the heavens and our sea by the power sal and artificial if a generative mechanism had of gravity, but we have not yet assigned the cause endured as a latent potentiality of nature until of this power.... To us it is enough that gravity awakened by science under experimentally con- does really exist, and act according to the laws trolled conditions or if it had never been acti- we have explained, and abundantly serves to ac- vated in its experimental range. But to the ex- count for all the motions of the celestial bodies, tent that the sciences are concerned with struc- 172 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE tures that not only exist but act independently of art Mill, pp. 65–6. them (and so explain what goes on in the world 42 See e.g. E. Nagel, op. cit., p. 493. outside the laboratory) the first possibility will 43 Cf. Weber’s concept of an ‘ideal type’ as be exceptional; and to the extent that they are a one-sided exaggeration of an aspect of ‘con- concerned with the conditions under which these crete’, i.e. empirical, reality. See e.g. M. Weber, structures act the second possibility will. Methodology of the Social Sciences. 36 J. S. Mill, A System of Logic, Vol. I, p. 378. 44 M. Scriven, op. cit., p. 466. 45 H. Poincar´e,Science and Hypothesis, p. 98. 37 E. Grunberg, ‘The Meaning of Scope and 46 G. E. M. Anscombe, op. cit., p. 21. External Boundaries of Economics’, The Struc- 47 Cf. M. Polanyi, ‘The Structure of Con- ture of Economic Science, ed. S. R. Krupp, p. sciousness’, The Anatomy of Knowledge, ed. M. 151. Grene, p. 321. 38 R. B. Braithwaite, Scientific Explanation, 48 Cf. M. Bunge, The Myth of Simplicity, p. 302. Chap. 3; and M. Polanyi, The Tacit Dimension, 39 There is another possible use for a ceteris Chap. 2. paribus clause, viz. as a protective device in the 49 Cf. P. K. Feyerabend, ‘Explanation. Re- early stages of a science’s development. This will duction and Empiricism’, Minnesota Studies in be considered in Chapter 3. the Philosophy of Science Vol. III, eds. H. Feigl and G. Maxwell, pp. 28–95. 40 J. S, Mill, op. cit., Bk. III, Chap. 10, Sect. 50 A caricature of such an empiricism exists in 5. some of the early experiments conducted under 41 P, T. Geach, ‘Aquinas’, Three Philoso- the august auspices of the Royal Society. The phers, G. E. M. Anscombe and P. T. Geach, p. following is an example: ‘1661, July 24: a cir- 103; and A. Ryan, The Philosophy of John Stu- cle was made with a powder of unicorn’s horn, 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS173 and a spider set in the middle of it, but it im- 2.8 Appendix: Ortho- mediately ran out several times repeated. The spider once made some stay upon the powder’, dox Philosophy of C. R. Weld, History of the Royal Society, Vol. I, p. 113. Among the items of allegedly scien- Science and the Im- tific interest collected by the Society were ‘the skin of a moor, tanned with the beard and hair plications of Open white’ and ‘an herb which grew in the stomach of a thrush’, ibid, p. 219. Quoted in P. K. Fey- Systems erabend, ‘Problems of Empiricism’, op. cit., p. 156. 127:1 It may be felt that I have dealt rather 51 See R. F. Harrod, The Life of John May- summarily in Chapter 2 with some of the nard Keynes. most hallowed doctrines of received philos- 52 Cf. H. L. A. Hart and A. M. Honore, Cau- sation in the Law, Chap. 2, Sect. 2. ophy of science; so I want here to turn to 53 M. Scriven, ‘Causes, Connections and Con- a more detailed examination of them in the ditions in History’, Philosophical Analysis and light of the phenomenon of open systems. History, ed. W. H. Dray, p. 248. 127:2 The structure of orthodox philos- 54 D. Gasking, ‘Causation and Recipes’, Mind ophy of science is based squarely on the 1955, pp. 479–87. Humean theory of the actuality of causal 55 H. Pelling, Modern Britain 1885–1955, p. 77. laws. But it is convenient to give Nicod’s 1 56 W. H. Dray, Laws and Explanation in His- criterion, which presupposes and implies it tory, pp. 66ff. co-equal status. I shall formulate them as 174 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE two principles: those who regard them as merely necessary. Once more no harm will be done by this con- 127:3 P1, the principle of empirical- invariance, viz. that laws are or depend upon flation as all my objections here turn on the empirical regularities; and lack of necessity of these principles and ipso facto of the theories they sustain. It is for 127:4 P2, the principle of instance- the sake of explanatory convenience and to confirmation (or falsification), viz. that laws avoid repetition that I formulate and discuss are confirmed (or falsified) by their in- them in their ‘necessary and sufficient’ form. stances.

127:5/o Post-Humean philosophy of sci- 128:1 Both P1 and P2 depend upon a clo- ence has called into question only the suffi- sure, and hence upon the assumption of a ciency, not the necessity of these principles; simple undifferentiated reality. P1 gives rise i.e. it has left the ontology implicit in them to the truth-functional concept of natural intact. Thus in the most advanced recent necessity. This is the idea that the logical positions theory is regarded as irreducible; status of laws can be explicated, at least in and as supplying at least part of the grounds part, by the formula “for all x, fx inplies gx,” for laws. The significance of such modifica- where the predicates ‘f’ and ‘g’ are defined tions will be considered in the next chapter. extensionally or are at least given some defi- Here, as in the body of Chapter 2, I will not nite empirical interpretation e.g. by means of distinguish between philosophers who regard correspondence rules. The definition may be 2 P1 and/or P2 as necessary and sufficient and ostensive or operational; and if ostensive, 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS175

3 4 either sensationalist or physicalist. P1 is mulated lawlike statement. (The exception susceptible of descriptivist (Mach)5 and in- is provided by the pure Humean case where strumentalist (Ryle)6 interpretations; and of the events are intuitively ascertainable atom- classical empiricist and transcendental ide- istic instants, and each event is a member of alist ones. P2 is susceptible of inductivist a linear series.) For one sequence to give sup- (Carnap)7 and falsificationist (Popper)8 in- port to another the antecedent events must terpretations; and of positivist and conven- be alike in relevant respects. But to talk of tionalist ones. the relevance of the ‘respects’ already pre- supposes a tentative (conjectured) or con- 128:2/o I am going to use the term ‘de- firmed law. Hence the inductivist in theory ductivism’ to refer to the ensemble of theo- must be a deductivist in practice. Similarly ries erected on the basis of P1 and P2. My a rule of inference can always be recast as choice of the term ‘deductivism’ may not the major premise of a syllogism. That is seem an altogether happy one in view of the to say, an inference ticket (a ‘season ticket’, fact that it is meant to cover philosophers as Ryle revealingly calls it)9 remains valid who have regarded themselves, as ‘induc- only as long as some empirical generalization tivists’ and ‘instrumentalists’, as opposed to is true. Thus the instrumentalist questions ‘deductivism’. However, I do not think that only the descriptivist interpretation of the this is a serious difficulty. For there is no way use of laws, not their logical form. in general of getting an inductive policy go- ing without appealing to an antecedently for- 129:1 I will first set out the overall struc- 176 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE ture of deductivism before examining its specified under their descriptions X and Y components, individually and collectively, in are regularly conjoined; and the light of open systems. 129:5 (2)” as a theory of how we must be 129:2 Underpinning deductivism is the ac- prepared to justify the claim that X causes tualist thesis that laws are relations between Y, viz. by showing that the event as spec- events or states of affairs. If the world con- ified under these descriptions are regularly sists only of atomistic events or states-of- conjoined. affairs then for a general knowledge of it to 129:6 In the case of both theories it is be possible their relations must be constant. possible to substitute the weaker require- And so we have P1 the principle of empirical- ment that the events be specified under invariance these or some other set of descriptions. The 129:3 (1) laws are or depend upon con- weaker variants could be indicated by the stant conjunctions of events or states of af- subscript1. fairs (which constitute the objects of actual 129:7 We then have a theory of explana- or possible experiences). As such it gener- tion, that ates the familiar Humean theory of causal- 129:8 (3) events are explained by subsum- ity. This theory is susceptible of different ing them under one or more universal laws; interpretations, viz. i.e. by deducing them from a set of one or 129:4 (2)’ as a theory of what we mean by more universal laws, together with a state- saying ‘X causes Y’, viz. that the events as ment of their initial conditions. This has be- 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS177 come known as the Popper-Hempel theory of symmetrical (in the sense of (iii) on page 63 explanation. Popper was the first to restate above). 10 it in modern times and Hempel has been 130:6 On this view their difference lies 11 its most systemic advocate and defender. merely in the fact that in ‘explanation’ the It is convenient to divide the theory into two explanandum event lies in the past and in requirements: ‘prediction’ in the future. 130:1 (3a) a deducibility requirement, viz. 130:7 Then we have a theory of the expla- that the explanandum be deducible from the nation of laws, namely explanans; and 130:8 (6) laws are explained by subsum- 130:2 (3b) a covering-law requirement, ing them under or deducing them from more viz. that the explanans contain at least one general, abstract or inclusive statements. universal law. Such statements may be regarded as theo- 130:3 Next, a theory of prediction to the retical principles or hypotheses. They may effect that be interpreted descriptively, instrumentally 130:4 (4) events are predicted by deducing or as fictions. I leave aside consideration of them from a set of universal laws together the various theories of theories until Chapter with a statement of their initial conditions. 3. (3) and (4) together give rise to the theory 130:9 Next a theory of the explanation of that theories and even sciences, viz. 130:5 (5) explanation and prediction are 130:10 (7) theories and sciences are ex- 178 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE plained by deductively subsuming them un- theory.13 der more basic or general ones. The ex- 131:2 These conditions tend inevitably to plained theory or science is then said to have 12 have a conservative effect. And they may been ‘reduced’ to the explaining one. Thus be regarded as rationalizing the practice of we have a theory of the development of sci- what Kuhn has called ‘normal science’.14 ence, viz. 131:3 I do not intend to discuss theories 130:11 (7)* science develops monistically (7)* and (7)** in any detail here. That will or in a linear fashion so as to leave meaning be done in Chapter 3. But their connec- and truth-value unchanged. tion through thesis (7) with the doctrine of actualism and their consequent dependence 130:12/o Theories (7) and (7)* are re- upon the presupposition of a closure should jected by many philosophers, most notably be clear. Popper, committed to other components of the deductivist structure. As Feyerabend 131:4 According to theses (3), (6), (7) and has pointed out these theories generate their (7)*, the explanation of events, laws, the- own restrictive methodology, embodied in ories and sciences all partake of essentially the conditions that the same ideal ‘deductive-nomological’ form. And they share this also (through theories 131:1 (7)** theoretical innovations should (2), (4) and (5)) with the activities of pre- be consistent and meaning-invariant with re- diction and the identification of causes. Sci- spect to established, i.e. currently accepted, entific knowledge then must consist (in part 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS179 or in whole) of a deductive structure. But and (9) have sometimes been interpreted in which one? Which out of all possible deduc- a conventionalist way. But as what is re- tive structures is the best (or in Popperian garded as ‘conventional’ is only the decision terms, ‘least worst’)? to accept a report as being genuinely ‘ob- servational’ (and so capable of furnishing a 131:5 Here P2 gets to work. Thus we have the theories that: genuine instance or counter-instance of a pu- tative law), a conventionalist interpretation 131:6 (8) laws, theories and sciences are does not affect the status of the principle it- directly or indirectly confirmed or corrob- self.15 orated by their instances (which constitute the objects of actual or possible experiences); 132:1 Finally, following on from (9), we and have a maxim of scientific practice to the ef- fect that: 131:7 (9) laws, theories and sciences are directly or indirectly falsified by their 132:2 (10) scientists should, in formulat- counter-instances (which constitute the ob- ing their theories, state quite unambiguously jects of actual or possible experiences). the empirical conditions under which they 131:8 Unless the meaning of theoretical are prepared to reject them (cf (iv) on pp. terms is reduced to ostensively defined in- 63–4 above). This may also be taken as a stances, as in Machian descriptivism, such criterion of what it is to be ‘scientific’ and criteria cannot be sufficient, but they are ‘unscientific’, viz. normally posited as at least necessary. (8) 132:3 (10)* to be ‘unscientific’ is not to be 180 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE prepared to state such conditions or having ularities; nor theories, because they are con- done so to revise them ex post facto.16 tinually being falsified; so that neither ex- planation nor prediction can be given a ra- 132:4 Most received philosophy of science tional basis. I have examined two actual- is based on a core extracted from theories ist responses to this predicament (viz. weak (1)-(10).* These theories all share one great and strong actualism) and showed how nei- weakness: they all presuppose a closure. If, ther can sustain the concept of laws applying as I have argued, the world is in fact open transfactually, viz. in open and closed sys- then they must all be more or less drasti- tems alike, that we need to render intelligi- cally revised and in some cases completely ble both the experimental establishment and rejected. I have already argued against (1) the practical application of our knowledge. in Chapter 2 (especially section 4) above so I will not discuss it separately here. It will be 132:5/o It is important to keep the de- remembered that the criterion of open-ness is ducibility and covering-law requirements, as the non-invariance of empirical relationships. expressed in (3a) and (3b) distinct. For ei- Now clearly if the law-like statements whose ther can be non-trivially satisfied without antecedents are instantiated in open systems the other. I have already shown in 2.6 how are interpreted as invariant empirical regu- one can have explanation in terms of a net- larities they must be regarded as false. But work of normic statements (which may be this means that there can be neither laws, strictly universal in the sense of space-time- because there are no invariant empirical reg- invariant) without the event being deducible. 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS181

Of course the sense in which the covering-law time restriction cannot itself be derived from ‘covers’ in this case is different. In the same a strictly universal law. way but from another aspect, the deducibil- ity requirement may be violated in the de- 133:1/o Thesis (9) implies that all law-like velopment of science although all the state- statements whose antecedents are instanti- ments involved are universal. An example ated in open systems are false. It is there- of this is given by the way in which New- fore as stated quite useless as a decision rule ton’s theory both explained and corrected for choosing between different law-like state- Kepler’s and Galileo’s laws.17 In these cases ments (or as an ‘organon of criticism’).19 (3b) is satisfied but not (3a). On the other One way of dealing with this would be to hand our atavistic causal explanation ‘Tania allow theory a role in grounding laws. But pushed the door open’ is deductive, though with an unchanged ontology this is bound to no laws are involved. It is also possible for be ultimately unsatisfactory. Of course if we the deducibility requirement to be satisfied possess a good theory it is irrational to re- by statements mentioning named individu- linquish it in the face of recalcitrant facts— als or specific space-time regions.18 It should without a better one. But our justification be noted that one can have deductive ‘ex- for holding on to the theory must be that it planations’ of events under transient empir- might eventually be able to explain them (by ical regularities but not deductive explana- suitable modifications, refinements or devel- tions of their explanations: for the space- opments) or show that they are not after all facts (i.e. that the statements used to state 182 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE them are untrue). If it could never explain rated by their occurrence in an undifferen- and/or correct them this justification would tiated way, i.e. independently of the context collapse. Of course the sufficiency of Nicod’s in which the putatively falsifying/confirming criterion must be disputed: the grounds for instance occurs. Theses (9) and (8) must a law or theory cannot be exclusively em- therefore be restated so as to place a restric- pirical. Theory must supply some idea of a tion on the system in which a genuinely fal- ‘connection’, without which it would be im- sifying/confirming instance occurs, viz. that possible to tell necessary from accidental se- it be closed. quences. But in an open world Nicod’s crite- 134:2 Not all evidence is equal; or rather rion cannot be necessary either: the grounds not all evidence is evidence for or against for a law or a theory cannot be undifferenti- a law. In general it takes a closed system atedly empirical. For the conditions must to furnish evidence capable of falsifying or normally be carefully controlled so that a confirming a law. And within the class of hypothesis about the connecting mechanism closed systems, experimentally closed ones can be put to a fair test. are preferable. 134:1 If it is wrong to regard law-like 134:3 For experimentally we can test and statements and theories as being falsified re-test a greater number and variety of sub- by the non-occurrence of their consequents junctive conditionals of the form ‘if x were to in open systems, it is equally wrong to re- take on a certain value, then yj would occur’ gard them as being confirmed or corrobo- by instantiating their antecedents. Whereas 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS183 outside the laboratory we are restricted to tion and falsification in that, taken together, observing whatever sequential performances they imply a correct prediction explains and nature is obliging enough to put on. an incorrect prediction falsifies. They de- pend upon the assumptions that it is possible 134:4 Similar considerations apply to the- to give a purely syntactical account of scien- sis (10). One cannot lay down hard and tific activities and that these activities al- fast criteria spelling out beforehand which ways occur in the context of an antecedently observable situations ‘if actually observed given closure. Recent philosophy of science mean the theory is refuted’ (see (iv) on pp. has clearly demonstrated the poverty of the 63–4 above). For one can never know be- former; it is with the incorrectness of the lat- forehand whether the system will be actu- ter assumption that I am here concerned. ally closed. On the other hand the closure of the system is not a part of the observ- 135:1 Now explanation in open systems, able situation; so that it cannot be incorpo- failing the attainment of an antecedent clo- rated into the criterion of scientificity with- sure, normally requires, as I have pointed out out destroying it. The judgement that the in 2.6 above, retrodiction; that is the infer- system is closed can only be made ex post ence from present effects to prior (perhaps after we have observed (and theoretically as- hidden, perhaps just unrecorded) causes, via sessed) the observable situation. the application of normic statements. Now 134:5 Theses (3) and (9) postulate a syn- the significance of this activity is that it pre- tactical identity between explantion, predic- supposes a non-conventional division of the 184 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE class of law-like statements into those which or putative law. are and those which are not capable of func- tioning in this way, i.e. into those which are 135:2/o It is easy to see why the expla- accepted (for transfactual application) and nation/falsification symmetry collapses once those which are still under test. Now once we allow the legitimacy of retrodiction. For we allow this the postulated symmetries be- the activity of retrodiction presupposes the tween explanation and prediction and expla- truth and applicability of the law used; the nation and falsification break down. And it possibility that it is false is ruled out a pri- becomes important to distinguish between ori. Now the intelligibility of falsification de- two kinds of prediction conflated in deduc- pends upon the idea that the would-be fal- tivism’s syntactical account of science: prac- sifier has independent grounds for the oc- tical predictions of categorical form which currence of the initial conditions. If the are rarely made in science but which are legitimacy of retrodiction is denied a vi- important in some of its practical applica- cious regress back to sense-experience en- tions in open systems and about which the sues. Thus suppose we have a law-like state- applied scientist can never be deductively ment of the form S1 ‘whenever events of type certain; and test predictions of hypotheti- E0 occur events of type E1 occur’. For S1 to cal form made under effectively closed condi- be used to explain the occurrence of E1 in tions in order to test a theoretical hypothesis a way which is consistent with the idea of its being subject to falsification independent grounds for E0 are required, say G0. But 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS185

the connection between E0 and G0 is itself a event could have been predicted. The root of contingent causal one, which may be repre- the trouble here is that the causal relation- sented by the hypothesis S, ‘whenever events ship is taking too much strain: it is required of type G0 occur events of type E0 occur’. both to be contingent (and as such to be Hence we need independent grounds for the subject to falsification) and to explain; func- occurrence of G0, say G’0, if our use of S2, is tions that it cannot combine without vicious to be consistent with the idea of its falsifiabil- regress to sense-experience. Once we distin- ity. But as G’0 stands in a contingent causal guish between open and closed systems, how- relationship we need independent grounds ever, this regress can be avoided. For we may for it too, and so on.... There is of course allow that events may be explained in open only one connection with E0 which, being and closed systems alike, but that law-like non-causal and non-contingent, does not re- statements may only be falsified under ef- quire independent grounds, namely immedi- fectively closed conditions (where deductive ate sense-experience. Thus insistence on in- test predictions are possible). dependent grounds for the initial conditions of an explanation, which is an inevitable con- 136:1/137 Now the point of the explana- sequence of the idea of its susceptibility to tion/prediction symmetry thesis is vitiated falsification, inevitably leads to the require- in open systems. For we can give excellent ment that the initial conditions be appre- explanations, in virtue of the transfactual hended in sense-experience; in which case the applicability of our independently validated knowledge, where we are incapable of any 186 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE predictions (save perhaps of the most imme- of much practical science of an exploratory diate or the most tentative sort). Moreover kind, such as prospecting for oil, depends the kinds of statements involved in the two upon the existence of a radical asymmetry activities are radically different: explanation between explanation and prediction. Be- proceeding by way of normic, and prediction cause of such difficulties defenders of the by way of empirical, statements. An empir- symmetry thesis have been forced to mod- ical generalization typically merely general- ify it so that it requires only that were we to izes the problem to be explained, whereas a be in possession of all the information avail- normic statement locates it in the context of able at the time of the explanation then we an explanatory theory. On the other hand, could have predicted it.21 Now I have argued normic statements may be inferior predictors in 2.6 that it is possible to give a complete to the most crude generalizations or rules of explanation of an event without thereby be- thumb. Further, the occurrence of the event ing in a position to deduce it, namely if the itself may be a practically necessary condi- different generative mechanisms at work are tion of our knowledge of the former state of of radically different kinds; so that the refor- the system, as in the case of the collapse of mulated symmetry thesis is either false or, a bridge or an airplane crash, or even of the if deducibility is built into the definition of kind of system with which one is concerned, ‘explanation’, uninterestingly tautologous.22 as in the case of the sudden onset of uncon- One further point on prediction. We are only trollable hysteria.20 Again, the intelligibility deductively justified in predicting an event if 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS187 the system is closed. But there is no way of our concept of a complex and differentiated knowing in advance (at the only time when reality we can see what is wrong with it. For, a prediction is relevant) whether the condi- on the one hand, there will always be more tions for a closure will in fact be satisfied. than one hypothesis capable of saving any Hence the probability of an event’s occur- given set of facts, so independent tests for rence can never be 1. them will always be necessary; and, on the other, it is only under closed conditions that 138:1 The undifferentiated ontology of re- such tests can be decisive. Consistency with ceived philosophy of science results in the the facts is neither necessary nor sufficient very damaging view expressed in (v) on page for a theory. 64 above, viz. that the acid test of a theory is its predictive power. On this view the more 138:2/o Popper does not seem to see the accurate a theory’s predictions—no matter connection between the criteria of explana- of what or where—the more worthwhile it is tion and rationality he expouses and the his- retaining. Coupled with permissiveness over toricist view of social science he condemns. the use of the CP clause, such a position pro- That there is a connection is clear. For if we vides a powerful rationale for scientific con- know that power corrupts and regard this servatism of any school. It has been used as a true, if trivial, empirical generalization as such by Heisenberg in physics and Skin- then given only a knowledge of the initial ner in psychology, by Friedman in economics condition that N is powerful we can predict and Osiander in astronomy. But armed with with deductive certainty that N will be cor- 188 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE rupted. On the other hand if we cannot know ence, but against the predictability of cer- when the initial condition is satisfied the law tain kinds of social events, viz. large-scale cannot be applied and so is quite useless for social changes and their consequences. It is either explanatory or social engineering pur- an argument against certain theories of so- poses. If Popper is committed to thesis (3) cial becoming which he interprets historis- in social science he is committed to thesis tically. Indeed one could almost say it is (4) and so to the historicist view of science an argument for an a-historical form of his- as the prediction of events (savoir, pour pre- toricism (in which the laws involved are re- voir). Popper nowhere denies the applicabil- garded as strictly universal) against an his- ity of thesis (8) to social science.23 But he torical one (in which they are regarded as equivocates between a conception of histori- spatio-temporally restricted). Historicism is cism as the view that the aim of the social in general invalid in all its forms. And it sciences is the prediction of future events24 is invalid in both natural and social science. and the view that its aim is to make un- And for exactly the same reasons. Inciden- conditional historical prophecies.25 But the tally, this is true for the refutation of his- authors he attacks did not make uncondi- toricism that Popper derives from the logical tional historical prophecies: Hegel in fact impossibility of predicting the precise effects made no predictions, and Marx only condi- of future knowledge in as much as any such tional ones. Popper’s real argument is not prediction depends upon a knowledge of that against predictability as such in social sci- knowledge.26 For its effects on nature are no 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS189 less predictable than its effects on men.27 mean to imply and in the case of (2)” we have grounds for supposing that a closure 139:1/o I turn now to the central unify- has been obtained of the system in which the ing theory of explanation, viz thesis (3) and events occur. Without this we cannot pos- the theories of causality, viz. (2)’ and (2)”, sibly mean nor can we reasonably be com- that underlie it. Consider once more the mitted to showing that the events which we paradigm of the kind of context in which claim are causally connected are so in virtue a causal claim is made. ‘Why is the door of being constantly conjoined. In short the opening?’ ‘Because Tania’s pushing it’. Now Humean theory cannot be a general theory it is certainly not the case that in saying of causality but at best a theory of what is ‘Tania’s pushing it caused the door to open’ involved in the making of causal claims in we mean that every time Tania pushes it closed systems and where the events are sep- the door opens. For there are times when arately identifiable atomistic instants. it is locked and times when she must turn the door knob too. (Nor equally is it the 140:1/o Davidson has proposed a sophis- case that every time the door opens is it be- ticated reformulation of the Humean theory. cause Tania pushes it.) But neither could ‘It is an error’, he says, ‘to think that no we produce any universal law which would explanation has been given until a law has show why the door opened in this particu- been produced. Linked with [it] is the idea lar case. Generally, theories (2)’ and (2)” that singular causal statements necessarily are only plausible if in the case of (2)’ we indicate, by the concepts they employ, the 190 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE concepts that will occur in the entailed law. unnecessary. For, on it, ‘A’ and ‘B’ stand in, Suppose a hurricane which is reported on in the original causal claim, for the complete page 5 of Tuesday’s Times causes the event atomistic state-descriptions that would form reported on page 13 of Wednesday’s Tribune. the antecedents and consequents of such a Should we look for a law relating events of causal law. But it is presumably only be- these kinds? It is only slightly less ridicu- cause (at least from a Humean viewpoint) lous to look for a law relating hurricanes we have not got and perhaps cannot get such and catastrophies. The laws needed to pre- state-descriptions that we make the original dict the catastrophy with precision would, of causal claim in the first case. Davidson holds course, have no use for concepts like hurri- that such claims are defended by ‘producing canes and catastrophies’.28 Davidson’s anal- a relevant law or giving reasons for believ- ysis of ‘singular causal statements’ is as fol- ing such exists’.30 How? then would one set lows: ‘“A caused B” is true if and only if about defending a claim about the causes of there are descriptions of A and B such that the French Revolution? By giving grounds the sentences obtained by putting these de- for believing that a true ‘neurological, chem- scriptions for “A” and “B” in “A caused B” ical or physical law’31 exists? But such laws, follows from a true causal law’.29 The objec- were they to be known, would cover any set tion to this suggestion is that it places a re- of historical causes of the French Revolution; quirement on the verification of causal claims in short, they would not enable us to discrim- which is (a) impossible, (b) useless, and (c) inate between true and false causal claims at 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS191 the level we are concerned with; that is, at terms of general statements at all, by an ap- the level at which the initial causal claim is peal to the normic statements of the level at made. Neurological laws are consistent with which the original causal claim was made. any social event and so cannot be possibly used to defend specific causal claims involv- 141:1 Hempel makes a similar mistake in ing people. Worse still, in the end the only arguing that singular causal statements of way of defending the belief that a Humean the ‘q because p’ (man bites dog) type ‘claim causal law exists covering some particular by implication’, ‘tacitly presuppose’, or ‘as- case will be by appealing to the truth of regu- sert by implication’ the existence of covering larity determinism. Hence the defense of the laws.32 Now it is not clear from his account most specific causal claim becomes an ex- whether he regards this as part of the mean- pression of faith in the philosophical dogma ing of the original causal claim (cf. thesis that the world is so constituted that the sim- (2)’ above) or merely as indicating the way ple formula ‘same cause, same effect’ every- in which it is to be justified (cf. thesis (2)’ where applies. On the other hand, once we above). If the former we have the absurdity deny the premise that a causal claim entails a that ‘man’ and ‘dog’ change their meaning Humean causal law, and hence a closed sys- when they are used in the making of a causal tem, these absurdities can be avoided; and claim. If the latter a confusion between what we can then allow that causal claims are de- it is necessary to know to ascribe causes and fended, where they require justification in what is necessary for our knowledge of the ascription of causes to be possible, viz. that 192 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE certain underlying laws hold. The latter is other categories: all explanations, contextu- neither necessary nor sufficient for the for- ally speaking, are elliptical; and the deduc- mer. tivist must show how partial explanations can be universally applicable. 141:2/o The non-availability of Humean causal laws is undoubtedly an embarrass- 142:1 At the methodological level one of ment to the modern Humean. Inevitably he the most unfortunate consequences of the falls back on the idea that our explanations spell of actualism is the blurring of the real are sketches to be filled out in the fullness of differences that exist between the various sci- time (cf. strong actualism) or that they are ences: both in their subject matters and the subject to an implicit ceteris paribus clause degrees to which they have achieved knowl- (cf. weak actualism). In addition he may re- edge of them. The experimental sciences lax the stringencies of the deductive model have been able as a result of theoretical en- by allowing other categories of explanation: deavour and technical ingenuity, to carve such as statistical, elliptical and partial for- out a chunk of the uncontrolled world and mulations.33 Statistical explanations, how- use it as an object of inquiry. The non- ever, spoil the point of the deductive model: experimental scientist has no such easy ac- for, as Scriven has put it, ‘they abandon cess. Now the view of the world as closed the hold on the individual case’.34 Moreover sets him off in the wrong direction—for it there is no a priori reason to suppose that the sets him looking for a complete description world is not statistically open.35 As for the of a given field. A view of the world as open 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS193 can, on the other hand, concentrate his en- ence. This is all that natural laws are’, E. Mach, deavours enormously. For it means that all Popular Scientific Lectures, p. 192. he has to do is to identify and describe (in 6 G. Ryle, The Concept of Mind, esp. pp. 116– ways to be considered later) some interest- 22. Cf. also C. S. Pierce, Collected Papers, Vol. 2, p. 354; F. P. Ramsey, The Foundations of ing and significant object of inquiry, without Mathematics, pp. 194–7; and S. Toulmin, The supposing that this will enable him to make Philosophy of Science, Chap. 3. deductively successful predictions. 7 R. Carnap, The Logical Foundations of 1 After J. Nicod, Foundations of Geometry Probability or ‘Testability and Meaning’, Read- and Induction, p. 219. See C. G. Hempel, As- ings in the Philosophy of Science, eds. H. Feigl pects of Scientific Explanation, p. 11. and M. Brodbeck, pp. 47–92. — 2 P. W. Bridgman, The Logic of Modern 8 K. R. Popper, The Logic of Scientific Dis- Physics, esp. pp. 1–25. covery, esp. Chap. 4 or Conjectures and Refuta- 3 Physics cannot be regarded as validly based tions, Chap. 10. upon empirical data until [light] waves have been 9 G. Ryle, op. cit., p. 117. expressed as functions of the colours and other 10 K. R. Popper, The Logic of Scientific Dis- sense-data’, B. Russel, Mysticism and Logic, p. covery, p. 59. 109. 11 See esp. C. G. Hempel, op. cit., Chap. 12. 4 See e.g. O. Neurath, ‘Universal Jargon and 12 See E. Nagel, op. cit., Chap. 11. Terminology’, P.A.S. Vol. 41, pp. 127–48. 13 p. K. Feyerabend, op. cit., p. 164. 5 ‘The communication of scientific knowledge 14 T. S. Kuhn, op. cit. Chap. II. involves description: that is, the mimetic repro- 15 Poincare is widely regarded as the founder duction of facts in thought, the object of which of conventionalism in the philosophy of science. is to replace and save the trouble of new experi- For a conventionalist interpretation of Popper 194 CHAPTER 2. ACTUALISM AND THE CONCEPT OF A CLOSURE see I. Lakatos, ‘Criticism and the Methodology of tions must be ‘deductive’. Most of the early Scientific Research Programmes’, Criticism and objections to the deductive model turned on the Growth of Knowledge, eds. I. Lakatos and the non-availability of generalisations connect- A. Musgrave, pp. 104ff. ing events like the cracking of radiators or mis- 16 See e.g. K. R. Popper, Conjectures and sile failures. Although this was no doubt encour- Refutations, Chap. I and passim. aged by the way in which its advocates presented 17 See e.g. P. Duhem, The Aim and Structure it, it was somewhat disingenuous of its critics of Physical Theory, Chaps. 9–10. not to realise the possibility of sophisticated re- 18 Cf. A. Donagan, ‘The Popper-Hempel The- formulations of it. Mandelbaum, for example, ory Reconsidered’, Philosophical Analysis and has argued correctly that such events must be History, ed. W. H. Dray, pp. 136–8. regarded as complex and analysed into compo- 19 K. R. Popper, Objective Knowledge, p. 21 nents (M. Mandelbaum, ‘Historical Explanation: and passim. The Problem of Covering Laws’, History of The- 20 Cf. M. Scriven, ‘Explanation, Prediction ory, Vol. I, pp. 229–42). However he still sees and Laws’, Minnesota Studies in the Philosophy explanation as depending upon a knowledge of of Science, Vol. III, eds. H. Feigl and G. Maxwell, laws (which he interprets in the Humean way) esp. pp. 176–90. covering the component events (ibid., p. 241); 21 See e.g. A. Grunbaum, ‘Temporally Asym- and given this, the complex event itself still re- metric Principles, Parity between Explanation mains deductively predictable. I have argued, by and Prediction and Mechanism versus Teleol- contrast, that the laws covering the components ogy’, Philosophy of Science Delaware Seminar, are normic and that they may involve reference Vol. I, ed. B. Baumrin, p. 73. to radically different kinds (so that they cannot 22 The interesting question is then of course be incorporated within a single theory). Hence shifted to that of whether complete explana- the complex event, even though completely ex- 2.8. APPENDIX: ORTHODOX PHILOSOPHY OF SCIENCE AND THE IMPLICATIONS OF OPEN SYSTEMS195 plained, may not be deducible. without a nature. 23 On the contrary he repeatedly emphasises 28 D. Davidson, ‘Actions, Reasons and the essential similarity in the logical form of the Causes’, Journal of Philosophy 1963, reprinted natural and the social sciences. According to in ed. A. White, Philosophy of Action, p. 92 him what is peculiar to the latter is its subject 29 Ibid, p. 92. See also D. Davidson, ‘Causal matter, and to history its interest in the partic- Relations’, Journal of Philosophy 1969, p. 701. ular (see K. R. Popper, The Poverty of Histori- 30 D. Davidson, ‘Actions, Reasons and cism, p. 143). But the pattern of explanation is Causes’, p. 92. the same. 31 D. Davidson, ibid., p. 93. 32 C. G. Hempel, op. cit., pp. 360, 362 and 24 See e.g. ibid., p 3. 363. 25 See e.g. K. R. Popper, Conjectures and 33 C. G. Hempel, op. cit., pp. 376ff and pp. Refutations, p. 339. 415ff. 26 K. R. Popper, Poverty of Historicism, pp. 34 M. Scriven, ‘Truisms as the Grounds for vi-vii. Historical Explanation’, p. 465. 27 Indeed one could even go so far as to say 35 The question of the closure of systems is that the latter depends upon the former; that distinct from that of their statistical or non- is, that it is only in so far as the effects of fu- statistical properties. The latter turns on the de- ture knowledge are unpredictable on nature than ducibility of single as distinct from mass events; they are unpredictable on men, in as much as whereas the former turns on the stability of em- all human action has a physical aspect but the pirical relationships. Clearly statistical systems converse is not the case. A social closure pre- can be open or closed. supposes a natural one but not vice versa; there could be a nature without men, but not men Chapter 3

The Logic of Scientific Discovery

3.1 Introduction: On the Contingency of the Causal Connec- 196 tion

143:1 In Chapter 2 I assumed the existence of a body of knowledge and asked how it 3.1. INTRODUCTION: ON THE CONTINGENCY OF THE CAUSAL CONNECTION197 could be applicable to the world. My partic- of atomistic events that implies it; and hence ular concern was to establish its universality from a critique of the idea of the actuality (transfactuality). I now want to turn to the of the causal connection to a critique of the question of how such knowledge, given that idea of its contingency. In Chapter 4 I will it is transfactually applicable to the world, ask what accounts for the assumption of the comes to be produced; and in particular to atomicity of the events conjoined that entails the question of how law-like statements come a closed system and generates, in its wake, a to be established as necessary. My concern host of philosophical problems. shifts here then from the synchronic to the diachronic aspects of science, and in partic- 143:3/o The connection between my con- ular to the question of how, in the social ac- cerns in this and the preceding chapter is tivity of science, natural necessity comes to clear. For once an ontology of atomistic be ascribed. In the course of this chapter I events is constituted, it follows that, for gen- will consider to what universality and neces- eral knowledge to be possible, events must be sity is properly ascribed, and what must be always conjoined (under appropriate descrip- the case for these ascriptions to be possible. tions) and never connected.1 That is, order in the world must consist of an unfailing or 143:2 In order to show how the concept invariant order of the co-existence of events of natural necessity is possible I will need in space and their succession in time. Con- to turn from a critique of the ontology of versely once it is appreciated that events, closed systems to a critique of the ontology though caused (and consisting in transfor- 198 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY mations), are very rarely conjoined, it can der to describe this movement I will need be seen why order in the world must be to reconstitute the other dimension of the pitched at a level categorically distinct from Copernican Revolution in the philosophy of events. Now I have argued in effect that science, viz. the transitive (or sociological) we produce conjunctions to discover con- dimension in which men come, in their social nections and apply connections in a world activity, to acquire knowledge of the endur- of non-conjunctions; so that events, though ing and transfactually acting mechanisms of rarely conjoined, are sometimes connected. nature, in virtue of which some but not other In this chapter I want to consider the na- sequences of events are necessarily connected ture of the connection that holds between and some but not other statements are uni- events (when it does) and the nature of the versally applicable. The idea that there are necessity implicit in the concept of law. I no necessary connections between matters of will thus be shifting my attention from the fact occupies an analogous position in under- differentiation of the world as such to the pinning the doctrine of the contingency of nature of the stratification that, if we are the causal connection, as the idea that there to render intelligible the experimental es- are always descriptions for events such that tablishment and practical application of our the formula ‘whenever this, then that’ ap- knowledge, it implies. Science attempts, I plies does in underpinning the doctrine of will argue, in its essential movement, to cap- its actuality. And I will argue that just as ture the stratification of the world. In or- for science to be possible the world must be 3.1. INTRODUCTION: ON THE CONTINGENCY OF THE CAUSAL CONNECTION199 open; so there must be necessary connections teristic kind of dialectic in which a regular- between matters of fact, if science is to be ity is identified, a plausible explanation for possible. it s invented and the reality of the entities and processes postulated in the explanation 144:1/o In Chapters 1 and 2 I have shown is then checked. This is the logic of scientific how the intelligibility of the activities of the discovery, illustrated in Diagram 3–1 below. experimental establishment and the practi- If the classical empiricist tradition stops at cal application of our knowledge presupposes the first step, the neo-Kantian tradition sees the categorical independence of causal laws the need for the second. But it either denies from the patterns of events, and how causal or does not draw the full (transcendental re- laws must be given an ontological basis in the alist) implications of the third step. If and enduring and transfactually active mecha- only if the third step is taken can there be an nisms of nature. Modern transcendental ide- adequate rationale for the use of laws to ex- alist philosophies of science, which are per- plain phenomena in open systems (where no haps more influenced by Wittgenstein than constant conjunctions prevail) or for the ex- Kant, stop at what is in effect the second perimental establishment of that knowledge stage of a dialectic or process of discovery in in the first place. science, by refusing to allow (or inadequately interpreting) the possibility of a realist inter- pretation of theory. Question 26 Describe in detail the 145:1 Thus there is in science a charac- three steps in scientific discovery. 200 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY

145:2 Diagram 3.1. The Logic of Scien- one. Now what is imagined at t1 may come tific Discovery at t2 to be known to be real. And for tran- scendental realism the move from (2) to (3) 145:3/o Just as transcendental realism dif- involves experimental production and con- ferentiates itself from empiricism by inter- trol, in which the reality of the mechanisms preting the first stage of the dialectic as the postulated in the model are subjected to invariance of a result rather than that of a empirical scrutiny. For transcendental re- regularity, so it differentiates itself from tran- alism that some real things and generative scendental idealism in its interpretation of mechanisms must exist can be established the second stage. Both transcendental real- by philosophical argument (their existence, ism and idealism see the move from (1) to and transfactual activity, is a condition of (2) as involving creative model-building, in the possibility of science). But it is contin- which plausible generative mechanisms are gent and the job of substantive science to imagined to produce the phenomena in ques- discover which ones actually do. That is, it tion. But whereas for transcendental ideal- is the task of science to discover which hy- ism the imagined mechanism is imaginary, pothetical or imagined mechanisms are not for realism it may be real, and come to be es- imaginary but real; or, to put it the other tablished as such. What is imagined may be way round, to discover what the real mech- real; but what is imaginary cannot. ‘Imag- anisms are i.e. to produce an adequate ac- inary/real’ marks an ontological watershed; ‘imagined/ known to be real’ an epistemic 3.1. INTRODUCTION: ON THE CONTINGENCY OF THE CAUSAL CONNECTION201 count of them. than in the ordering of present ones or the memory of past ones that our rational and 146:1 Science is a process-in-motion. It in- empirical ‘faculties’, ‘whose unkind and ill- volves three distinct stages, which cannot be starred divorce’ Bacon saw as responsible for omitted or collapsed into one another with- all the confusion in ‘the affairs of the human out doing tremendous violence to our under- family’,2 are most productively combined. standing of science. But these stages cannot be identified with moments of chronological 146:2/o It is only, I shall argue, if we al- time; they are phases of science. It should low the possibility of the move from (2) to be noted that the move from (1) to (2) just (3) that we can, in the end, uphold the legit- because it involves the postulation of novel imacy of the move from (1) to (2). Moreover entities and processes cannot be given a de- it is only if we begin to see science in terms ductive interpretation. But given this it can of moves and are not mesmerized by termi- only be justified in a non-pragmatic way if nals that we can give an adequate account of we hold out the possibility of a realist inter- science. In this respect much philosophy is pretation of some of the hypothetical entities still in the same position as a Martian trying etc. invoked to explain the behaviour. Such to discover what trams are but able only to an interpretation can in turn only be justified observe them in open-air museums with chil- empirically if it is set in the context of the dren scrambling over them. It is the task of ongoing social activity of science. Thus it is the philosophy of science to capture science’s in the planning of future experiences rather essential movement, not to guess its eventual 202 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY destination. ality, of the world known by science. And it is in this unacknowledged ontological legacy 147:1 Recent work in the philosophy of sci- that the weaknesses of both developments ence has established (i) the fact of scientific lie. My aim in this chapter and the next is change and (ii) the poverty of a purely de- to pinpoint these weaknesses. And to show ductivist analysis of explanation. In this way in particular why and how an adequate non- it has done much towards the establishment empiricist account of science, capable of ac- of a conception of science as a critical social commodating the facts of scientific change activity. The case for transcendental real- and structure, requires an ontology of the ism can, however, be strengthened by con- kind outlined in Chapter 1 and elaborated sidering the limitations of this work. For in Chapter 2. Indeed, recent philosophy of unless these two insights are taken together science illustrates very well the kind of ‘on- and a new ingredient is added to the existing tological tension’ that can occur when a fun- philosophical mix they are, I think, vulner- damental objection is made to a philosophi- able to positivist counter-attack. This new cal theory without simultaneously question- ingredient must be in the field of ontology. ing that theory’s ontology. The general dif- The argument of Chapter 1 enables us to see ference between recent philosophy of science why this is so. For the logical empiricism and transcendental realism could be summed against which recent philosophy of science up by saying that whereas recent philosophy has reacted contained not only an account has asked merely what are the conditions of of science, but (implicitly) an account of re- 3.1. INTRODUCTION: ON THE CONTINGENCY OF THE CAUSAL CONNECTION203 the possibility of individual experience and a social transmit;3 which they must them- found an answer in the intersubjective world selves reproduce or partially transform. The of science, transcendental realism asks in ad- Copernican Revolution in the transitive di- dition for the conditions of the possibility of mension of the philosophy of science thus has the social activity of science, finding an an- the profound implication that man never cre- swer in the intransitive world of things. ates, but only changes, his knowledge, with the cognitive tools at his disposal. Secondly, 148:1 I will need in this chapter not only what is to be changed, has first to be ac- to show the necessity for the philosophical quired. And what is acquired consists always ontology of transcendental realism, but also of an ensemble of theoretical and empirical to begin the development of the philosophi- ideas, so that knowledge can never be anal- cal sociology that I argued in 1.6 is presup- ysed out as a function of individual sense- posed by any theory of science. Scientific experience. Once this is grasped the grounds development, I have argued so far, consists for the atomistic ontology that generates the in the transformation of social products, an- idea of the contingency of the causal connec- tecedently established items of knowledge, tion collapse. which may be regarded as Aristotelian ma- terial causes. Certain implications flow from 148:2 Science then is an ongoing social ac- this conception. First, that men never con- tivity which pre-exists any particular genera- struct their knowledge from scratch. It tion of scientists and any particular moment stands to them always as a given product, of consciousness. Its aim is the production 204 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY of the knowledge of the independently exist- 3.2 The Surplus- ing and transfactually active mechanisms of nature. Corresponding to the criterion de- element in the veloped in the intransitive dimension of the Analysis of Law- philosophy of science, viz. the conceivabil- ity of a world without men, we thus have a like Statements: criterion in the transitive dimension, namely the inconceivability of knowledge without A Critique of the antecedents. Theory of Models

148:3/o It has often been held that a con- stant conjunction of events is not a sufficient condition for a causal law. This may be be- cause it is regarded as incapable of sustain- ing the intuitively obvious and important difference between necessary and acciden- tal sequences or in Johnson’s time-honoured terminology between universals of law’ and ‘universals of fast’.4 Or it may be because it is regarded as incapable of licensing what 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS205 it is intuitively felt causal laws do licence, philosophers then is to provide an alterna- namely counter-factual conditionals.5 It is tive account of the ‘surplus-element’7 in the never seriously denied that we feel, and sci- analysis of law-like statements; that is, that entists act as if, some but not other se- element over and above the (presumed) con- quences of events are ‘necessarily connected’; stant conjunction that explains our ascrip- so that we must possess the concept. What tions of necessity; and which will show how, the radical empiricist, in the form of Hume, and the conditions under which, a distinction denies is: (a) that there is any objective basis between necessary and accidental sequences for this distinction, i.e. that it corresponds and the assertion of counter-factuals can to any real difference between the two se- be rationally justified. The usual response quences of events; and (b) that there is any to this challenge consists in the attempt justification, apart from habit or custom, for to locate the surplus-element in the state- our ascriptions of natural necessity and ac- ment’s ‘explanation’, and more particularly cident.6 Most philosophers since Hume have in the ‘theory’ which explains it. However attempted to show how he was wrong in (b) the terms ‘explanation’ and ‘theory’ cover a without objecting to (a). I want to argue gamut of philosophical positions, which must that Hume was wrong in (a); and that it is now be considered. only if we can establish this that we can show why he was wrong in (b) also. 149:2/o The deducibility of a law-like statement from a set of higher order state- 149:1 The radical empiricist challenge to ments is often regarded as a criterion of 206 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY

‘explanation’.8 However if deducibility is than another, but not why one statement the only criterion for explanation and the rather than another is regarded as law-like.10 source of the surplus-element is its expla- nation there will be an infinite number of 150:1/o Of course it might be objected surplus-elements for any statement. Hence that when everything is explained all factual any statement can be said to be law-like on statements will be law-like. But what would an infinite number of grounds!9 Deducibil- count as an explanation then? Could it be ity alone cannot explicate the distinction be- anything other than an inexplicable constant tween necessary and accidental or nomic and conjunction of events, as in the case of Mill’s non-nomic universals. Moreover additional unconditional laws?11 If it could not, we are criteria such as simplicity can only reduce back with Hume, and have done nothing the number of possible explanations for a to allay the sting of the radical empiricist statement which has already been identified challenge. If it could, some alternative non- as law-like. But they cannot be used to say Humean analysis of the ultimate or highest- which statements are law-like and so possess order laws must be given which will show the surplus-element. For even if there were how they, as uniquely qualified ‘explainers’, a simplest explanation for every statement, do possess a genuine surplus-element. We there are no absolutely simple explanations. are thus faced with the following dilemma: Thus such criteria can at best be used to ex- either explanation is achieved by subsump- plain why we choose one explanation rather tion under higher-order laws in which case the problem is merely shifted, for a surplus- 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS207 element must be found for them if they are to accidental generalization is capable of yield- qualify as ‘laws’; or an alternative analysis of ing correct predictions, viz. as long as the ‘explanation’ must be given, which does not conditions that account for it persist. This identify the explanans with a further set of suggests that, even if we were to possess laws, and so provides room for the location some general inductive warrant, predictive of a surplus-element in the analysis of laws, success alone could not differentiate neces- within the context of their explanation, at sary from accidental sequences or license the any one level. assertion of counterfactuals.

151:1 It might be thought that it is in 151:2/o It seems clear that if we are to the capacity of the law-like statement to get any further in our search for the surplus- yield successful predictions that the source element the idea of purely formal differen- of the surplus-element lies. But this will not tiae must be abandoned. Inductive consid- do without an analysis of the ‘capacity’ or erations prove no better than deductive ones. ‘power’. For the Humean it is the past and For accidental generalizations may be induc- actual successes of the statement that count, tively confirmed, just as they may be deduc- not its potential ones. And these can at best tively explained. In practice then the non- explain, not justify, the surplus-element. It radical empiricist, if he is not to concede the is the surplus-element that must provide our game, is forced to re-examine the account inductive warrant, if we have one: rather of science that seems to render any non- than the other way round. Moreover even an Humean conclusion impossible. The funda- 208 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY mental fact about science that has been miss- Hume’s notion of custom and can no more ing from the discussion so far is the existence justify our attributions of necessity than the at any moment of time of an antecedently latter could. established body of theory. And it is here that the non-radical empiricist attempts to 152:1 In short, unless theory contains el- locate the surplus-element. But can ‘theory’ ements irreducible to experience and truth- do what experience and deducibility fail to functional operations on it there is no basis do, i.e. provide a rational ground for our as- for a non-Humean theory of natural neces- criptions of natural necessity? The answer sity.15 Thus the possibility of the latter de- clearly depends upon the extent to which pends upon some terms of the theory not be- the former contains components irreducible ing explicitly defined in terms of experience to the latter. And the onus is on the philoso- and/or some statements of the theory not pher who attempts to locate the surplus- being deductively connected and/or some element in the systematic organization of ideas of the theory being non-propositional our knowledge or the capacity of a theory in logical (or non-sentential in linguistic) to explain many different laws12 or to pre- form. These establish the possibilities of dict novel kinds of facts13 to show how their intensional relationships between predicates, concept of theory escapes Humean analysis. non-deductive (e.g. analogical) relationships Goodman’s notion of entrenchment,14 for ex- between ideas and non-propositional (e.g. ample, functions in exactly the same way as iconic) ideas respectively as potential sources of necessity. It is the second of these that has 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS209 been most thoroughly explored; and it is to other and exchanging their momentum by Campbell’s initial formulation of the theory impact. And it is in our prior understanding of models that I now turn. of this that the necessity of the gas laws ul- timately lies. Notice that for Campbell it is 152:2/o On Campbell’s view a theory not the mere availability of a theory or even must contain not only a ‘dictionary’ corre- the organization that the theory makes pos- lating some, but not all, of the theoretical sible (e.g. the fact that Boyle’s law, Charles’ concepts with empirical terms but a ‘model’ law and Graham’s law are all deductive con- for the hypotheses or theoretical statements sequences of the kinetic theory) but the in- of a theory, by means of which its hypothet- terpretation theory explains.17 ical subject matter may be imagined to be like in some, but not all, respects the real 153:1 As a critique of the deductivist view empirical subject matter of some field which of the structure of scientific theories, as typ- is already known.16 On this view the surplus- ified by Mill, Duhem and Hempel, Camp- element just is the model. Thus what distin- bell’s case is a strong one. The deductivist, guishes Boyle’s law from a merely accidental he says, merely exhibits ‘the dry bones of generalization is, according to Campbell, the science from which all the spirit has de- corpuscularian model informing the kinetic parted”.18 His project is to revitalize it. He theory of gases. By means of this model gas sees the driving force of science as the ex- molecules are imagined to be, in certain re- ploitation of analogies in the conquests of spects, like billiard balls bouncing off each new fields, without which neither theory nor 210 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY the range of facts could grow or the language deduction; and it is by means of it that we 19 in which to state them develop. But is his render Ta intelligible to ourselves. See Dia- case unanswerable? How does it fare when gram 3.2 below. According to Campbell the faced with the challenge of radical empiri- entities and processes postulated at Ta are cism? Is it capable of providing an adequate unknowable; i.e. they do not constitute part account of the surplus-element in the analy- of the phenomenal world described by sci- sis of law-like statements? To answer these ence. Although we cannot know what pro- questions we must look more closely at the duces Ea we can imagine it to be like some- terms of the modelling relationship which thing we do know. Such an act of imagina- is intended to provide the basis for a non- tive daring need not be totally arbitrary. For Humean theory of natural necessity. it is possible to conceive of principles of ana- logical, just as there are principles of deduc- 153:2/o Now essential to Campbell’s cor- tive or inductive reasoning. Only when we rection of the deductivist view of explana- have constructed a model can we be said to tion is the idea that for the explanation of have achieved scientific understanding. That a range of phenomena say Ea to have oc- is, not just saved the facts, preferably with curred the relationship between the theory elegance and economy, but explained them. Ta which explains the phenomena and from Using the analogy provided by Eb a real or which the latter is deducible must be sup- empirical phenomena can thus ‘enliven’ the plemented and informed by another relation- abstract theoretical relationships from which ship. This is a relationship of analogy not 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS211

Ea is deduced. And Eb does this by standing shave to Occam’s razor. Tyndall formulated in for or representing (in the sense of the Ger- the criterion for the selection of explanations man ‘darstellung’) the unknown causes of implicit in Campbell’s theory as follows: ‘ask Ea. Explanation thus involves, centrally, the yourself whether your imagination will ac- substitution in our imagination of a real or cept it’.21 Now such a criterion is clearly ca- empirical relationship for an unreal or theo- pable of selecting a theory within a given retical one.20 This is Campbell’s debt to em- metaphysical schema, such as that provided piricist ontology: a debt that it summed up by the classical mechanical world-view. But by his tacit acceptance of the concept of the it is not capable of judging between differ- empirical world. For on his theory Ta cannot ent schemas, when it is precisely the nature be, or come to be known as, real; though it is or the limits of the imagination that is in at any moment of time, and perhaps forever, question. To take an obvious example: Aris- unperceivable to us. For him theoretical en- totelian and Galilean dynamics are in con- tities, such as molecules, can only be said to flict over whether when a stone falls to the be ‘real’ by analogy with material objects. earth, the earth should be conceived as fixed (Aristotle, Ptolemy and Tycho Brahe) or as 154:1 Diagram 3.2 Campbell’s ‘Theory’ moving (Copernicus, Giordano Bruno, Ke- 154:2 Campbell does not deny that pler and Galileo). Now, try as you may, there deducibility is necessary for explanation, is no neutral way of conceiving the falling merely that it is sufficient. His theory may of the stone.22 Our imagination, although thus be regarded as providing an alternative 212 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY not fixed, is either Aristotelian or Galilean. for the scientist to stand Tyndall’s criterion Tyndall’s criterion cannot help us to decide on its head, and dizzily ask himself whether between the competing frameworks, because he can continue to accept his imagination. what it in question is the nature of the con- cept in terms of which any motion has to be 155:2/o Although its inadequacy to deal understood. with fundamental scientific change is most evident, Tyndall’s criterion is no less inade- 155:1 There is a similar break involved in quate to deal with the continuing processes the transition from Newtonian to Einsteinian of conceptual micro-adjustment, in which dynamics. Part of the trouble with current our imagination is continually modified and micro-physics is that our imagination can- extended, that are a part and parcel of the not accept it, and yet we have every reason process of ‘normal science’. More gener- to believe it to be true. If Tyndall’s criterion ally, it is always legitimate for scientists to were acted upon it could have effects on sci- ask and sometimes possible for them to an- entific practice as conservative and dogmatic swer, questions about whether gases are re- as the consistency and meaning-invariance ally composed of molecules or whether the conditions of classical empiricism. A new sci- earth really moves. Such questions cannot entific ontology or a fundamental change in be rephrased as questions about the plausi- scientific concepts may transform our con- bility of our conceptions. This would be, in ception of what is plausible. At such times terms of Diagram 3.1, to reduce phase (1) to in the history of science it becomes necessary phase (2). Rather the normal procedure in 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS213 science is if we have a plausible conception but as essentially pragmatic devices, servic- to go on to ask whether it is true, which is ing the needs of the understanding. Theory to ask whether the entities and processes it involves a journey from one set of experi- postulates are real, or only fictional. Plausi- ences Eb to another Ea. Because of this it is bility is a prima facie criterion for a theoret- always possible for the radical empiricist to ical explanation. But is is neither sufficient, ask whether the journey is really necessary. nor in the last instance necessary. Moreover, even if a way could be found of showing that some model is necessary, there 156:1 How does Campbell’s theory fare as would seem to be no way of justifying the a response to the challenge of radical empiri- choice of any particular one (given that the cism? According to it, the surplus-element idea that its necessity could be demonstrated in the analysis of law-like statements is the a priori is rejected as being inconsistent with model at the heart of the theory that ex- the fact of scientific change). plains it. But for Campbell the model can- not prompt questions about the reality of the 156:2/o To this it may be contended that abstract entities and processes postulated in models are necessary not only as concep- the theory. For theoretical entities are by tual crutches for the tender-minded and as definition unperceivable and hence, given the heuristic devices for the young (which the fundamental equation of empiricist ontology, radical empiricist may graciously concede) viz. real = empirical, cannot exist. Mod- but for a theory’s growth and development, els function then not as knowledge-extending and in particular (so as not to beg the is- 214 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY sue by positing non-empiricist criteria of de- ‘world’ that Campbell’s account of science velopment) for the generation of facts em- contains: the world of Mach and Hume. In pirically relevant for the theory but which such a world causality is bare and invariant would not have been forthcoming without conjunction; and scientific knowledge con- it.23 But this only pushes the argument back sists, for its part, in ‘description, that is the a stage further. In a completed science mod- mimetic reproduction of facts in thought, the els would be dispensable. For, as Duhem has object of which is to replace and save the put it, ‘to explain is to strip reality of the ap- trouble of new experience’.26 pearances in which it is wrapped as in veils in order to see this reality naked and face 157:2 Suppose now that arguments are ad- to face’.24 When we have done this, what vanced to show that no science can ever be more can there be to do? The objection that complete in the requisite sense. Science still ‘explanations are practical context-bound af- remains, on the Campbellian conception, a fairs’25 either is covered by the heuristic role purely internal process, locked in a dosed allowed to models or depends upon the in- circle of thought. Science is still a creature completeness of science, in which case their of custom and habit, the only difference be- nemesis is merely (if perhaps indefinitely) ing that the habit is now one of the imag- postponed. ination, rather than sensation. In virtue of their shared ontology Campbell is closer to 157:1 We are thus forced inexorably back Mach and Tyndall to Occam than one might to a particular conception of reality, the only think.27 In neither case can the possibility 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS215 of major conceptual revisions be accommo- surplus-element distinguishing a law-like dated or the mechanism of scientific discov- from a non law-like statement is the concept ery be displayed. of the generative mechanism at work produc- ing the effect in question. Such mechanisms 157:3/o Let us apply to Campbell’s the- exist and act independently of men; so that ory the litmus test for the adequacy of an ac- the necessity can be properly ascribed to the count of science developed in Chapter 1. Can sequence. Moreover as the world is open not it sustain the idea of the applicability of the all events will be connected by a generative concept in question, viz. that of necessary mechanism; so that the transcendental real- connection, in a world without men? The ist can sustain a concept of natural accident. answer is obvious. In the case of Campbell, as of Hume, there is still no difference, inde- 158:2 Only a real difference between nec- pendent of men, between a necessary and an essary and accidental sequences can justify accidental sequence of events. The Camp- our distinguishing law-like from non-law-like bellian can at best talk of a nomically neces- statements. Hence one cannot deny Humean sary statement; he cannot talk of a nomically conclusion (b) (on page 149 above) with- necessary sequence. The attempt to locate out objecting to Humean conclusion (a), and the surplus-element in the analysis of law- thus to the ontology that implies it. like statements in the imagination of men is 158:3/o Nowhere is the anthropocentric- a failure. ity of post-Humean philosophy more evident 158:1 For transcendental realism the than in the notion that natural necessity 216 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY must be sought in the behaviour or nature of the world, which experience can (in ways of men. And nowhere is the displacement to be explored in paragrpah 3 below) confirm of rational intuitions more obvious than the or falsify. We are not locked in a dosed circle attempt to locate structure in the imagina- of thought; because there are activities, viz. tion of men. ‘Connection’ is, as Chisholm perception and experimentations by means has remarked, an ‘ontological category and of which under conditions which are delib- a source of embarrassment to empiricism’.28 erately generated and carefully controlled, But it is not an irreducible one. For its basis relatively independent cross-bearings on the lies in the generative mechanisms of nature intransitive objects of thought can be ob- which connect events as cause and effect and tained. Such activities are not independent which exist as the powers of things. Thus of thought, but their results are not implied to assert a counterfactual is not to make a by them either. meta-statement29 (which would be to make a statement about its grounds), but to make 159:1 Campbell’s achievement is to have a statement about the way some thing would seen that scientific theory cannot be identi- have behaved (exercised its tendencies, lia- fied with a deductive system erected on the bilities or powers) had the conditions in fact basis of a single set of experiences. But he been different. Theory is not an elliptical made two mistakes. He too, like the em- way of referring to experience,30 but a way piricists, missed the essential point that sci- of referring to hypothesized inner structures ence is essentially developing; so that the hy- pothetical mechanisms of yesterday may be- 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS217 come today’s candidates for reality and to- lines now stand for relationships of deduction morrow’s phenomena. But behind this fail- and the continuous lines for relationships of ure also lay an inadequate intransitive di- analogy (to indicate their reversed relative mension, and in particular the absence of the importance). Ta has come to be established concept of objects apart from our changing as real, and in this case also is perceivable. In knowledge and possibilities of perception of the course of this process facts Ea have been 0 them. Campbell’s theory has been extended corrected and now become facts Ea. Ta now in two ways. Some have rectified his first provides one of the sources for a new model mistake but not his second, and viewed sci- designed to explain phenomena Ea. And the ence as a sequence of models, an unfolding process of checking its reality (which will al- process of shifts in intellectual fashion. Oth- most certainly modify our conception of it) ers have developed his theory in a realist way. has begun. Needless to say there will in gen- Harre, for example, has drawn attention to eral be more than one model for Ea. The the role of the existential questions prompted state of chemistry c.1930 provides an illus- by the creative use of analogies in the devel- tration of the model. Ta is Prout’s hypoth- 31 opment of science. By way of concluding esis and Tα the theory of sub-atomic struc- 0 my discussion of Campbell’s theory I want ture. Ea → Ea, consists in the elimination to sketch out such a dynamized realist ver- of the impurities that dogged the verifica- sion of it. tion of Prout’s hypothesis for over a cen- tury. And the new model might be the Bohr- 159:2/o In Diagram 3.3 below the dotted 218 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY

Rutherford model of atomic structure; which genuine and pseudofalsification, and hence conceived as a hypothesis about the internal for invoking the CP clause in defence of gen- structure of atoms is, we now know, false. eralizations in the former case. This calls The source of such models may lie either into question the necessity of deducibility, in some general conceptual scheme (such as not just its sufficiency in the explanation of atomism in chemistry) or some other science laws. or proto-science (such as the wave models of light and sound in particle physics). The 160:2/o Science needs a concept of pseudo- subject of such models is the unknown but falsification for three reasons, two of which knowable intransitive structure of the world. are epistemic and one of which is ontolog- It is by means of the experimental testing of ical. Firstly, because a theory may not be the hypotheses suggested by already existing at present sufficiently refined or developed to knowledge that new knowledge comes to be cope with anomalous counter-instances; that produced. is to say, every theory needs a ‘protective- belt’ for its development.32 Secondly, be- 159:3 Diagram 3.3 A Dynamic Realist De- cause the ‘facts’ may be wrong: either in velopment of Campbell’s Theory the simple sense that they are misrepresenta- 160:1 The problem of the surplus-element, tions of the phenomena or more profoundly and Hume’s challenge, has another aspect. because they depend upon false or inad- This turns on the question of what warrant equate observational theories.33 As is well we have for distinguishing between cases of known, every new theory is faced with innu- 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS219 merable anomalies and counter-instances of erative mechanism at work and this concept these kinds. They form in a sense the staple is irreducible to that of a sequence of events diet of normal science. A successful theory is then it is quite rational to uphold an ontolog- one which, like Newton’s, though it never re- ical distinction between cases of genuine and solves them all and generates new ones in the pseudo-falsification (in which, as exemplified process of their resolution, ‘turns each new by the case of Prout’s hypothesis referred to difficulty into a victory for its programme’.34 above, our epistemic distinctions too may be Thirdly, science needs a concept of pseudo- grounded). For we may readily allow that falsification because a countervailing cause the generative mechanism in virtue of which or interfering agent may be at work gener- natural necessity is ascribed is not under- ating the ‘counter-instance’. It is only un- mined by the instability of the conditions un- der closed conditions, as we have seen, that der which it operates. So that if a law has a theory can be given a fair test or that a been confirmed under closed conditions and crucial experiment—Bacon’s ‘instance of the there is no reason to suppose that the gener- fingerpost’35—becomes possible. ative mechanism at work in those instances has ceased to operate, the law that the con- 161:1 The problems of the necessity and cept of the mechanism grounds may be sup- universality of law are indeed inextricably posed to continue to apply outside the condi- linked, but not in the way Hume thought. tions under which it was confirmed, whether For if the surplus element in the analysis of or not the consequent of the statement hap- law-like statements is the concept of a gen- 220 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY pens to be realized. distinguishing necessary from accidental se- quences is solely pragmatic; that it is, as it 161:2/o By now it would, I think, be gen- were, a question of our greater attachment erally agreed that models play some cogni- to the former (Quine),36 or of the deeper en- tive role in science and that there is a fea- trenchment of their predicates in our con- ture about such models which renders them ceptual system (Goodman). Similarly it en- irreducible to the experiences that they are courages the idea of the CP clause as a de- in some way intended to embroider or ex- vice that can be relaxed or invoked, switched plain. (This feature is, I have argued, typi- off or on, according to whether or not we are cally an idea of a mechanism which would, if prepared to forego the falsified law-like state- it were real, generate the phenomena in ques- ment. This view carries the implication of tion.) But the representatives of the three course that the use of the CP clause is bound traditions in the philosophy of science differ to be more or less arbitrary or dogmatic.37 radically in their interpretations of the sta- And this in turn creates the Kuhn-Popper tus and role of such models, and of the irre- problem of the functions of dogma. ducible concept that constitutes its essential core. 162:2 The concept of the generative mech- 162:1 The classical positivist view is that anism may be given a firmer status, and the it is merely a heuristic device (Duhem, distinctions it grounds a better rationale, by Hempel and Brodbeck). This is liable to seeing its function as concerned essentially encourage the view that the rationale for with the development of science. Protec- 3.2. THE SURPLUS-ELEMENT IN THE ANALYSIS OF LAW-LIKE STATEMENTS: A CRITIQUE OF THE THEORY OF MODELS221 tion from pseudo-falsification then becomes is that of transcendental realism, science is protection from too easy or too early falsifi- concerned neither with the incessant accu- cation; that is, before the full potentialities mulation of confirming facts (or the inces- of the theory have been developed (Lakatos sant search for falsifying ones), nor even and Feyerabend). This view allows that our with its own growth and development, but knowledge is structured—that it contains, as rather with the understanding of the differ- it were, layers of different age. The condi- ent mechanisms of the production of phe- tions of knowing are here explicitly distin- nomena in nature. Thus it allows that un- guished from the conditions of being. But der certain conditions the concept of the gen- positivism still provides the underlying ac- erative mechanism at work may be given a count of the world. And because of this realist interpretation as a representation in the rationale of the concept of the genera- thought of the transfactually active causal tive mechanism, which forms the heart or structures of the world. The possibility of essential core of the theory, is still more or such an interpretation supplements internal less pragmatic, still science- or knowledge- or consistency and contextual plausibility as a man-dependent. constraint on the possible forms of theoreti- cal advance; and it constitutes the ultimate 162:32/o The third position consists in goal of all theory construction. coming to see not just our knowledge but the world itself as structured and differenti- 163:1 Now empirical realism generates the ated. According to this conception, which following dilemma: Either theoretical enti- 222 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY ties refer ultimately to experience, in which nature! case they can be eliminated. Or theoreti- cal entities constitute experience (in whole or in part), in which case they cannot be eliminated, but must, given the equation of 3.3 Natural Necessity empirical realism, constitute the world (in and Natural Kinds: whole or in part). Now as long as an ontol- ogy based on the category of experience is The Stratification retained there can be no grounds indepen- of Nature and the dent of man for ascribing necessity to some but not other statements. On the first horn Stratification of Sci- this generates the problem of what justifies our belief that the future will resemble the ence past, or the unobserved the observed, i.e. the problem of the induction. But on the second 163:2 In the process of the establishment of a horn it generates the problem of what jus- law of nature three questions may be asked: tifies the assumption of intellectual confor- 163:3 (i) is there an empirical regularity mity. And, on this horn, scientific change, or which constitutes a prima facie candidate for even dissent, actually constitutes (in whole a law? or in part) a breakdown in the uniformity of 163:4/o (ii) is there some reason, other than the regularity, why the predicates in- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE223 stantiated in the law-like statement should empiricist, transcendental idealist and tran- be conjoined? scendental realist philosophies of science. At 164:1 (iii) is this reason located in the en- the Humean level laws just are empirical reg- during powers of things and the transfactu- ularities. At the Kantian level both (i) and ally active mechanisms of nature? (ii) must be satisfied. Here we have what might be called the dual criterion theory of 164:2 If the answer to (i) is yes we have 39 38 law. I have already noted its vulnerabil- what might be called a ‘protolaw’. If the ity to Humean counterattack. At the level answer to (ii) is yes we have strong grounds of transcendental realism, a distinction is for a law. If the answer to (iii) is yes we drawn between the empirical identifiability have a law. Typically of course the reason in and the universal (transfactual) applicability question in (ii) will be provided by a model of laws; and the latter is seen to be a condi- of the connection between antecedent and tion of the possibility of the former. As the consequent, putative cause and putative ef- application of laws in open systems is jus- fect. The transition from (ii) to (iii) typi- tified, and presupposed by the intelligibility cally occurs when a realist interpretation of of experimental activity, the existence of an the mechanism posited in the model becomes empirical regularity or a constant conjunc- acceptable. tion of events is now not even necessary for 164:3/oo The answers to (i)–(iii) corre- the ascription of a law (see Table 3.1 below). spond of course to three levels of criteria I have argued that it is only at this level that for law, viz. those specified by the classical 224 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY a distinction between necessary and acciden- tal idealism Y N transcendental realism N tal sequences can be sustained. A sequence N Ea.Eb is necessary if there is a generative 165:1 Diagram 3.4 (not typed in) mechanism M such that whenever Ea, Eb 165:2 Diagram 3.5 (not typed in) tends to be produced; a sequence is acciden- 166:1 Now these three levels of criteria tal if this is not the case. Their difference generate and are generated by different views is represented in Diagram 3.4 below. Most of science. Thus whereas the classical em- events occur in open systems and must be piricist will ask merely: conceived, as argued in 2.6 above, as ‘con- 166:2 (i)* is there a regularity such that junctures’. This is illustrated in Diagram whenever C then E? 3.5 below. Necessity as such, like universal- ity, is thus ascribed essentially to the activ- 166:3 The transcendental idealist will ask ity of the mechanism; and only derivatively in addition: to some particular event sequence. For the 166:4 (ii)* given a regularity, is there an result of the activity of the mechanism will explanation such that we can render it in- in general be co-determined by the activity telligible to ourselves that whenever C then of other mechanisms too.40 E?. 166:5 The transcendental realist will how- 164:3 Table 3.1 Status of Constant Con- ever, after making an essential correction, go junction of Events Necessary Sufficient for one step further and ask: Law classical empiricism Y Y transcenden- 166:6 (iii) out of the plausible explana- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE225 tions for this regularity, is there one which sible hypothetical mechanisms by the cre- correctly describes the mechanism by means ative employment of his imagination (cf. Di- of which, upon the occurrence or obtaining agram 3.6) and subjects these ideas to rigor- of C, E tends to be produced? ous theoretical criticism and empirical test. These three phases of science are of course 166:7 That is to say, the transcendental those represented in Diagram 3.1 on page realist will demand that models be tested 114 above. not just for plausibility but for truth; i.e. for their adequacy in correctly describing the 166:8/oo For the transcendental realist real generative mechanism at work (if the then a model has a relationship with its sub- connection between C and E is necessary) ject as well as with its source. And it is such that when C occurs, E tends to be pro- within the nexus formed by this double ar- duced (is produced in the absence of inter- ticulation that new knowledge is produced. fering causes or the transformation of M). For new knowledge is doubly articulated, ar- That real things and generative mechanisms ticulated in two dimensions (transitive and must exist can be established by philosoph- intransitive): it is a socially produced knowl- ical argument. It is the job of the scientist edge of a natural (man-independent) thing. to discover which ones actually do. Given It is this bipolarity that a model expresses the identification of some prima facie non- in standing in two sorts of relationship: a re- random pattern in nature or protolaw the lationship of analogy with its source; and a scientist thus builds up ideas of various plau- relationship of adequacy (when it is) with its 226 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY subject matter. Many philosophical prob- be laid down for it. Science is work that re- lems arise from a misunderstanding of the quires creative intelligence, and there can be second relationship. It is not a relation- no mechanical surrogate for that. The idea ship of correspondence; the terms of the re- of an automatic science is a will-o’-the-wisp lationship are not necessarily like each other, that the philosophy of science has pursued, though pictures and iconic models may play with damaging consequences, since Bacon’s an important role in scientific thought.41 search for a ‘sure and certain method’ that Moreover there are no general philosophi- would eliminate the need for human thought, cal criteria for such judgements of adequacy; which of course inevitably entails the possi- they are necessarily intrinsic to the particu- bility of human error. lar science concerned. Analogy is one of the 167:1 Diagram 3.6 (not typed in) possible relationships that models may have with respect to their source. The existence of 168:1/o Most science proceeds by way of the first type of relationship (in the transitive a two-tiered method designed to identify in- dimension) is important in establishing both variances in nature, normally under condi- a constraint on the number of possible ex- tions which are experimentally produced and planations42 and an indispensable means of controlled, and to explain them by reference their production. It cannot be described at to enduring mechanisms.43 It is in the move- all adequately as one of coherence; and here ment from the identification of an invariance again no general philosophical criteria can to the mechanisms and structures that ac- count for it that the logic of scientific dis- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE227 covery must be found. Thus the observable sists in the discovery of the mechanisms re- reactions of chemistry, which are represented sponsible for chemical bonding and valency. in textbooks by formula such as 2Na + 2HCl This has been explained in terms of the elec- = 2NaCl + H2, are explained by reference tronic theory of atomic structure. Once the to the atomic hypothesis and the theory of reality of this explanation has been estab- valency and chemical bonding. The pat- lished, science moves on to the discovery of terns which constitute the explananda of the the mechanisms responsible for what hap- theory of valency are needless to say by no pens in the sub-atomic microcosm of elec- means superficially obvious or readily avail- trons, protons, and neutrons; and we now able. Both the concepts and the substances have various theories of sub-atomic struc- and conditions had and have to be worked ture. The historical development of chem- for, produced in the social activity of science. istry may thus be represented by the follow- The theory itself sets out to describe the ing schema: causal mechanisms responsible for the overt behaviour of the substances. Once its real- ity has been established (which justifies our 169:1 Stratum I 2Na + 2HCl = 2NaCl + assuming that chemical bonding occurs and H2 explained by Stratum II theory of atomic the laws of chemistry hold outside the labo- number Mechanism 1 and valency explained ratory) and the consequences of the theory by Stratum III theory of electrons and Mech- have been fully explored, the next task con- anism 2 atomic structure explained by Stra- tum IV [competing theories of sub- [Mecha- 228 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY nism 3] atomic structure] that level. The key move in this involves the postulation of hypothetical entities and 169:2 It should be noted that the historical mechanisms, whose reality can then be ascer- order of the development of our knowledge of tained. Such entities need not be smaller in strata is opposite to the causal order of their size,45 though in physics and chemistry this dependence in being. No end to this process has normally proved to be the case. The of the successive discovery and description species of explanation here identified itself of new and ever deeper, and explanatorily falls under a wider genus: in which the be- more basic, strata can be envisaged. Other haviour of individuals is explained by refer- sciences reveal a similar open-ended strati- ence to their natures and the conditions un- fication. Geometrical optics is explained in der which they act and are acted upon (see terms of Young and Fresnel’s wave optics; 2.3 above). which is explained in terms of the electro- magnetic theory of light; which can be ex- 170:1 Now for the transcendental realist plained in terms of the quantum theory of the stratification this form of explanation radiation.44 imposes upon our knowledge reflects a real 169:3/o A general pattern of scientific ac- stratification in the world. Without the con- tivity emerges from this. When a stra- cept of real strata apart from our knowledge tum of reality has been adequately described of strata we could not make sense of what the the next step consists in the discovery of scientist, striving to move from knowledge of the mechanisms responsible for behaviour at one stratum to knowledge of the next, is try- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE229 ing to do: viz. to discover the reasons why tive terms must have initially possessed a the individuals which he has identified (at meaning independent of them. This enables a particular level of reality) and whose be- us to see how meaning-change is possible, haviour he has described tend to behave the and indeed if the independence of predicates way they do. Without this concept the strat- is denied, inevitable in the transitive process ification of science must appear as a kind of of science. Similarly we can see how knowl- historical accident, lacking any internal ra- edge of newly discovered strata may correct tionale in the practice of science (if indeed knowledge of less fundamental strata, as con- it is not denied altogether in a reductionist cepts and measurement techniques are re- and ultimately phenomenalist account of sci- fined. Now if changing knowledge of strata ence). is to be possible the strata must not change with our knowledge of them. Thus the con- 170:2 As it is clear that the hypothetical cept of real strata apart from our knowl- entities and generative mechanisms imagined edge of them is necessary if both the ideas for the purposes of theory-construction must of scientific structure and scientific change, initially derive at least part of their meaning which are central to recent critical philos- from some other source (if they are to be ca- ophy of science, are to be intelligibly sus- pable of functioning as possible explanations tained. More generally, acknowledgement at all) theories must be already understood of the real stratification of the world allows before correspondence rules for them are laid us to reconcile scientific discovery (of new down.46 Equally this means that the descrip- 230 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY strata) with scientific change (of knowledge timate. For it will still remain possible for of strata). him that there are reasons, located at a still deeper level, for the causes of the phenomena 170:3/o Now the stratification of the world he has succeeded in identifying and describ- must be assumed by the scientists working ing. I will return to this point below. in any field, to be in principle unbounded. For it will always be possible for him that 171:1 Now the only kind of necessity that there are reasons, located at a deeper level, holds between events is connection by a gen- for the phenomena he has hitherto identified erative mechanism. But there are two other and described. But his knowledge may be in concepts of necessity applicable to the objec- practice bounded by semi-permanent tech- tive world order: there is the necessity im- nical or conceptual problems or by the do- plicit in the concept of a law, i.e. in the activ- main assumptions of his particular science; ity of a generative mechanism as such or the or by the fact that reality is itself bounded exercise of a thing’s tendencies irrespective at the level knowledge of which he has at- of their realization; and the necessity implicit tained. However, if the stratification of the in the concept of a thing’s real essence, i.e. world has an end, i.e. if there are ‘entities’ those properties or powers, which are most which are truly ultimate—and I can see no basic in an explanatory sense, without which reason for supposing this must be so—and it would not be the kind of thing it is, i.e the scientist has achieved knowledge at that which constitute its identity or fix it in its level, he can never know that the level is ul- kind. The first concept of ‘natural necessity’ 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE231 is clearly derivative from the second, depen- ean and Leibnizian levels respectively. The dent upon the contingent feature of the sys- transcendental idealist, as well as the classi- tem in which the thing’s behaviour occurs, cal empiricist, is, in virtue of his ontological viz. that it be closed (see 2.4 above). I am commitment, restricted to the first level of therefore going to refer to the second as the knowledge of the objective world order. I concept of natural necessity, and the third as shall argue that even at the Leibnizian level the concept of natural kinds. Knowledge of science remains empirical, so that the tran- natural necessity is expressed in statements scendental argument of Chapter 1 remains of causal laws; knowledge of natural kinds in valid; and that even at that level the deduc- real definitions. But natural kinds exist and tively justified prediction of events is impos- naturally necessary behaviour occurs inde- sible, so that the critique of philosophy of pendently of our definitions and statements science contained in Chapter 2 continues to of causal laws. apply with undiminished force. Moreover I shall argue that the concepts, such as that of 171:2/o Now in the transition from knowl- natural powers, that we need to render intel- edge of any one stratum to knowledge of the ligible the transition to the Leibnizian level next, knowledge of three levels of the objec- remain categorically valid even at that level. tive world order is progressively obtained: of relations between events, of causal laws 172:1 At the Humean level a pattern and of natural kinds. I am going to refer is identified or an invariance is produced. to these three levels as the Humean, Lock- (This, we know, empirical realism cannot 232 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY sustain.) We thus have a protolaw (at let sic structure, e.g. genetic constitution atomic us say Stratum I). This is to be explained structure or electric charge. Now it is contin- by reference to the circumstances and nature gent that x has the nature (e.g. constitution of the thing whose behaviour is described. or structure) that it has. But given that it The scientist never doubts for a moment that has, It is necessary that it behaves the way something is generating the effect in ques- it does. One criterion of this is our capacity tion. His problem is: what is? That is, why to deduce the tendency to B from N. This is does x behave the way it does, viz. B, in con- the Lockean level of knowledge. Note that at ditions C1 ...Cn? this level it is still contingent that the thing has the structure that it has. 172:2/o The first step in the scientific ex- planation of B is to ascribe a power (or li- 173:1 Now at the third Leibnizian level ability) of x to B, i.e. to do (or suffer) phi. possession of that structure or constitution This is to say, very roughly, that x does φ comes to be regarded as defining the kind in virtue of its nature N.47 The next step of thing that x is. Now it is necessary that is thus to investigate N (defining Stratum x has the structure it has if it is to be the II). This involves inter alia creative model- kind of thing it is. It is no longer contin- building and rigorous empirical-testing (cf. gent that hydrogen is a gas with a partic- Diagram 3.1). As a result of this investiga- ular atomic structure; rather anything pos- tion we may say x comes to do φ in virtue sessing that structure is hydrogen. That is, of its having a certain constitution or intrin- the criterion for the application of the con- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE233 cept ‘hydrogen’ ceases to be the lightest gas more the importance of viewing science as and become instead possession of that struc- a process in motion is clear. For if we stay ture. At this level the only contingent ques- at any one level, phase or moment of science tions are whether and where things of a given the idea that a definition may be arrived at kind exist.48 But note contingency still lies empirically will appear absurd. If it is ac- in the flux of the circumstances in which cepted, however, the reason why the laws of things act, so that events are still not de- nature cannot be deduced a priori from self- ductively predictable. That is, the ‘contin- evident axioms becomes clear. For the ax- gency’ of events deriving from open systems iomatic base of a science at any moment of applies even at the Leibnizian level, so that time, at any stratum of reality, is something laws must still be formulated as tendencies that has had to be worked for, produced, as (whatever their scope of application). More- part of the irreducibly empirical process of over it is important to see that knowledge at science. the Leibnizian level is, or may be, attained empirically. We may discover, quite empir- 173:2/o Scientists attempt to discover ically, that the most important explanatory what kinds of things there are, as well as property or real essence of hydrogen, identi- how the things there are behave; to capture fied as the lightest gas, is its atomic struc- the real essences of things in real definitions ture; and then attempt to express this dis- and to describe the ways they act in state- covery in a real definition of hydrogen. Once ments of causal laws. The real essences of things are their intrinsic structures, atomic 234 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY constitutions and so on which constitute the will return in paragraph 5 below real basis of their natural tendencies and causal powers. Thus there is no conflict 174:2/o The situation at the Humean level between explanatory and taxonomic knowl- is rather like that faced by the citizens of edge. Rather, at the limit, they meet in Koenigsberg who knew, from experience, the notion of the real essences of the nat- that there was no way of crossing each of the ural kinds, whose tendencies are described town’s seven bridges just once.49 See Figure in statements of causal laws. 3.1 below. At the Lockean level this fact is deducible from the topology of Koenigsberg, 174:1 At the Leibnizian level statements given Euler’s theorem. At the Leibnizian of law are substitution instances of necessary level, there is a necessary truth about a cer- truths about the individuals to which the re- tain physical set-up, whether or not there is a fer. For any individual which did not behave town called ‘Koenigsberg’ or any town at all in that way would not be an individual of to which it applies. At the Leibnizian level, that kind. They may thus be regarded as Mendeleyeev was able to deduce from his analytic truths. But they are arrived at in Periodic Table, interpreted as dealing with the transitive process of science a posteriori, atomic number and valency, the properties by empirical means. Thus a fully dynamic of several new elements. But it remained philosophy of science must take seriously the contingent whether, and if so where, there question ‘how is analytic knowledge arrived were elements in the world to which his pre- at a posteriori possible?’ To this question I dictions applied. Certain chess games have 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE235 only one possible solution. But it remains doubtless have been, for them, a rash conjec- contingent whether they are ever played. ture at the time. Moreover it is far preferable to the Humean alternative, viz. that when- 174:3 Figure 3.1. The Seven Bridges of ever men smoke opium they fall asleep. For Koenigsberg (not typed in) in the first place, the latter is untrue. Sec- 175:1 The concept of powers has played a ondly, it is less informative. It might be a key role in our analysis of science’s transition complete accident that everyone in the den from knowledge of one stratum to knowledge is asleep: the powers formula rules this out. of the next. To ascribe a power is to say It says that there is something about opium that a thing will do (or suffer) something, in virtue of which when men smoke it they under the appropriate conditions, in virtue tend to fall asleep. The connection is nec- of its nature. This is not, as is so often essary. But it is only a tendency. Thirdly, claimed, a pseudo-explanation50 or a purely the Humean formula is regulatively useless. verbal formula.51 Rather it is an indication The powers statement is by contrast quite of work to be done. Moliere’s doctors in Le suggestive. For it indicates the need for Malade Imaginaire have often been ridiculed an investigation into the chemical properties for speaking of opium as possessing a ‘dormi- of opium and the way they induce sleep in tive virtue’. But in doing so they left open men.52 In context, it constitutes an open ad- the possibility of an investigation, at some mission of ignorance. The Humean, on the future date, into the nature of opium with- other hand, must pretend that once he has out committing themselves to what would 236 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY his generalization there is nothing more to neous (or simultaneous) and unselfconscious be known. And if he should stumble upon response of the scientist to the identification a higher-order generalization this can be, for of protolaws; it represents, if you like, an at- him, only accidentally related to the origi- tempt to reconstruct the internal rational- nal one: there is no inner logic connecting ity of the inter-strata move. The concept of the two, or rationale by which science moves powers is not intended to figure in the dis- from the first to the second. course of science, but in the discourse of the philosophy of science (which is the former’s 175:2/o To this it might be objected that rational reconstruction). the concept of powers does not figure in the discourse of science. This is true. And the 176:1 It should perhaps be stressed here reason for it is of course that the scientist, that the stages of my rational reconstruction unlike Moliere’s doctors, is never just con- of the process of scientific discovery repre- tent to ascribe a power but moves immedi- sent phases of scientific activity; they cannot ately to the construction of possible expla- be identified with moments of chronological nations for it with the paradigms and other time. Thus most scientific work must oc- instruments of thought at his disposal. That cur, for reasons I will bring out in paragraph is his job. (Sometimes, however, when we 4, in the context of a research programme are completely at a loss we do just ascribe a designed to show that on the supposition of power.) The concept of powers is introduced the mechanism M the field of phenomena can precisely to describe this normally instanta- be rendered intelligible. Thus the identifica- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE237 tion of a protolaw normally depends upon ers, in pinpointing an essential moment in the prior existence of a conjecture or a hy- the transition from knowledge of one stra- pothesis of a mechanism intended to function tum to knowledge of the next, plays such as a possible explanation for the presumed a key developmental or strata-bridging role. protolaw.53 In this way, a powers statement is a promis- sory note cashed in the development of sci- 176:2/o To ascribe a power is to say that ence, a schematic explanation filled out in there is something about the thing, which the growth of our knowledge.54 may be unknown, in virtue of which it be- haves the way it does. The grounds for the 177:1 It is worth noting that the structure ascription of a power must thus be stronger of a powers ascription is well adjusted to ac- than the mere occurrence of a regularity. For commodate both falsification (obviously, as we must possess some reason to suppose the the hypothesized reasons may be subjected connection necessary (though in the limit- to independent tests) and meaning-change ing case this may just be the invariance of (less obviously). If meaning change is to an experimentally produced result). It is be- be possible, some elements of meaning must cause it indicates the power-ascriber’s belief remain constant through the change. Now in the existence of a reason, located at the if ‘x does B’ is analysed as ‘x is of such a next highest level of inquiry (in the nature nature N [defining Stratum II] that it will of the thing), whether or not the reason is do φ in conditions C1 ...Cn [defining Stra- currently known, that the concept of pow- tum I]’ we may allow that the meaning of 238 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY

‘phi’ remains constant while the meaning of we have climbed up to Stratum II, we can- the N component changes between theories not throw away the ladder, so to speak. To (and vice-versa). This applies even in the pursue the analogy, the ladder is a rope, case of simple descriptive observational pred- not a wooden one. For to make a powers icates such as ‘blue’. For ‘x is blue’ may be statement is to make a categorical statement analysed as ‘x looks blue [defining Stratum about the nature of the thing situated at I] in virtue of its reflecting light of a certain the level to which we have climbed. It is to wavelength [Stratum II]’. The simple theory make a statement about possibilities which that things look blue because they are blue are possessed by the thing quite categori- may then be replaced by the scientific the- cally, whether they are known (or actualized) ory that they tend to look blue in normal or not. Dogs do not lose their power to bark circumstances because they reflect light of when we understand how they do so, just as wavelength 4400A. Subsequently we may al- glass does not cease to be brittle when we low the latter to define the scientific use of know its molecular structure. ‘blue’; in which case of course it is no longer contingent that blue surfaces reflect light of 178:1 The ontological bases of powers just that wavelength. are the properties that account for them; i.e. the natures in virtue of which they are 177:2 Now although the concept of pow- ascribed. Now in the transitive process of ers serves this essential developmental func- science such natures may come to be qual- tion, it cannot be reduced to it. For when itatively described. When this happens it 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE239 will of course initiate a search for the higher- to render the thing or behaviour directly ac- order entities and mechanisms that account cessible to the scientist’s senses (so that it for them. But how does it happen? becomes the possible object of an act of im- mediate demonstrative reference). And it is 178:2 In general, at any one level, indi- the aim of the latter to produce the concept viduals must be identified and their normic of the object, so that the scientist is capable behaviour described. Now for a qualitative of an adequate description of it. Both are description of a thing or a dispositional ac- necessary for a qualitative description. count of its behaviour it must be present to the scientist’s senses and he must be able 178:3 It should be noted that the two to describe it correctly, i.e. as being of the kinds of criteria, viz. demonstrative and kind or type that it is. This will normally recognitive, are distinct. For my incapac- depend upon two kinds of work: practical ity to identify the chromosome structure by (experimental and technical) work, in which peering down an electron-microscope does the scientist’s causal and perceptual powers not mean that it is not a possible object are augmented (the latter with the aid of the of an act of demonstrative reference. It is construction of sense-extending equipment, present to my senses, whether I recognize such as microscopes); and theoretical work, it or not. Conversely to render it accessi- in which the scientist’s conceptual and de- ble to my senses is an independent labour scriptive powers are augmented. It is the (itself only possible if some concept of it is aim of the former to produce the object, i.e. possessed), requiring great ingenuity, just as 240 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY experimental production and control does, ground or auxiliary theories to specify the when we are concerned with the description conditions under which an object of the ap- of the law-like behaviour of some thing al- propriate type may be said to be present to ready identified. The production of the ob- the senses. In this sense they constitute, as ject and the production of its concept are it were, the criteriology of empirical science. thus independent tasks, each essential to a qualitative description of a thing or account 179:1 It is clearly essential to the theory of its behaviour. The thing must be there of scientific development proposed here that and I must know what kind of thing it is, i.e. imagined entities may come to be established how to describe it; in general this will involve as real. Now an entity may be ‘theoreti- a theoretical redescription of it. cal’ either in the sense that its existence is open to doubt (theoretical1) or in the sense 178:4/o Now it is important to realize that it cannot be directly perceived, either that though the production of the object unaided or with the help of sense-extending and the production of its concept are dis- equipment (theoretical2). The same distinc- tinct, the judgement that the object has been tion applies in the case of behaviour. Now produced itself depends upon a tacit theory an entity (or mode of behaviour) may be of vision and the instruments according to theoretical1 at t1 and perceived and ade- which its range is extended. The case of the quately described at t2, so that it then ipso electron microscope illustrates this very well. facto ceases to be theoretical1. The existence In general it is the function of such back- of bacteria, initially conceived as minute hos- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE241 tile micro-organisms, and molecules, initially netic field by a compass needle. That there modelled on material objects, came to be es- is a difference between the cases of detection tablished in this way. This is typical of sci- and perception is clear. In the case of detec- ence and shows once more the importance of tion the thing can be individuated only indi- viewing it as a process in motion. rectly, i.e. via the spatio-temporal framework or through its effects on particular things; it 179:2/o But if an entity cannot be per- cannot be the object of an act of immediate ceived, i.e. is theoretical2, does this mean demonstrative reference. Whatever the men- that it cannot be known to exist, so that tal imagery we use to think of a magnetic it must be theoretical1? If this were the field it can be present to us only through its case all theoretical2 entities would indeed be effects. On the other hand my incapacity hypothetical, and our knowledge would be to identify a bacterium under a microscope necessarily confied to the domain of observ- as being of a particular type, or even as be- able things, even if this were now regarded ing a bacterium at all, does not mean that as an expanding class. Fortunately this con- it is not present to my senses; and so ca- clusion does not follow. For theoretical2 en- pable of functioning as the object of a pos- tities may be known to exist indirectly, viz. sible act of immediate demonstrative refer- through the perception of their effects. The ence, although ex hypothesi I am incapable paradigm here is the case of the detection of intentionally performing it. of radio-active materials by a geiger counter, of electricity by an electroscope, of a mag- 180:1 It should be stressed that in the de- 242 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY tection case that something does exist pro- here. One is that there is a nature, suscepti- ducing the effect is not in question. Nor ble in principle to a qualitative description, is the fact that it exists and acts indepen- as yet unknown, which is the bearer of its dently of its detection. To say ‘electricity is causal powers. The other is that the na- what electricity does’55 is to collapse pow- ture of the thing just is its causal powers, ers to their exercise. Electricity is not what as in the case of physical field theories. At electricity does; but what it can do. The any moment of time a science may have to mode of reasoning employed in inferring the put down its ultimate entities just as pow- existence of causal agents through the osten- ers to produce effects, e.g. to affect observers sion of their effects is thus perfectly proper. and equipment, possible observers and pos- Hence though it is correct to say that when sible equipment, material things, in certain we cannot qualitatively describe the cause ways. About such entities all the scientist we know less about it than when we can knows is their powers. It always remains (given that in the latter case we know the possible that he will be able to achieve a thing’s causal powers as well) it is not true qualitative description of them, and he must to say that there is a cause is less certain. It strive to do so. On the other hand, it is also is just that in the detection case what we can possible that such entities are their powers. know about a thing is limited to its causal The scientist can never dogmatically elimi- powers. nate one of these alternatives in advance. If there is a frontier to possible knowledge of 180:2/o Now there are two possibilities 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE243 the world the scientist can never know when the epistemic one. he has reached it. But whatever is responsi- ble for the world as manifest must possess 181:1/o The general thrust of my argu- causal powers which are continually being ment in Chapter 2 was against reduction- exercised; it must be co-extensive with space ism. How does this square with my empha- and continuous with time. It must be struc- sis on strata of knowledge? It will be re- tured and complex; it cannot be atomistic or membered that I did not deny the possibil- event-like. The concept of a field of poten- ity of an explanatory reduction but stressed tial seems closest to meeting these require- (a) the need for a well-defined reductans (so ments.56 However it seems to me there is no that a reduction could not in general be a reason in principle why there should not be means of acquiring knowledge of a higher- strata of fields (of perhaps radically different order or less fundamental stratum); and (b) kinds), forever unknown to us. It should be that a reduction left the reality of the higher- noted that only the identification, not the ex- order entities intact, at least in as much as istence of fields depends upon the existence they were causal agents capable of acting of material things in general. Here again the back on the materials out of which they are order of dependence in being is opposite to formed (see 2.5 above). It is clear that I the order of dependence of our knowledge of was there taking possession of causal pow- being. The ontological order is distinct from ers, and hence existence in time, as the most general criterion of reality. There is an asym- metry between space and time here. For 244 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY powers must be possessed and exercised in is satisfied. On it, quantum mechanics and time, but they need not be localized at any chemistry would belong to the same branch. point in space. Relations, for example, such But electromagnetism and mechanics, neu- as that of spin (in physics) and marriage en- rophysiology and psychology and (it will be dure through time and have causal effects. argued) psychology and sociology would be- But they have no position in space. Now in long to different branches. general a reduction is possible because the entities in terms of which the behaviour of 182:1 Changes of things are explained in the thing is explained occupy a different vol- terms of unchanging things. If there are ume of space, either larger or (more usually) ultimate entities they must be unchanging. smaller. Thus the possibility of a reduc- Atoms have already been disqualified as pos- tion implies in general that the individuals sible ultimate entities (see 2.3 above). So of the different kinds cannot be said to oc- ultimate entities must be powers; that is, in- cupy the same place at the same time and dividuals characterized solely by what they one not be part of the other. This gives us a can do. For if one could describe the chang- general criterion which imposes limits on re- ing states or conditions in virtue of which gresses of strata, i.e. upon the possibility of their powers were exercised they could not a sequence of (explanatory) reductions. For be ultimate (unchanging). In the last in- one could define a branch of science as a se- stance to be is just to be able to do. But ries of theories within which this criterion this does not rule out the possibility of a science of cosmology (which would be con- 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE245 cerned with the distribution in space and as any system of relations in which objects redistribution in time of the ultimate enti- stand to one another. And we can conceive ties) or of irreducibly historical branches of the possession and exercise of causal pow- science in which the ultimate entities were ers in time in ways, and at levels, forever Aristotelian or even Strawsonian individuals. unknowable to men. We can never know The transformation of the principles govern- where we stand absolutely in the chain of ing such things would in general have to be being. Despite this cosmic incapacity sci- conceived as conjuncturally determined open ence has succeeded in identifying strata of systemic events (see 2.6 above). In this way reality. Now a scientist never doubts for a a complex thing such as a person (or a so- moment that there are reasons for the be- ciety) could come to be the cause of its own haviour he has identified and described. It transformation. is in the search for such reasons, at a deeper level of reality, at present known to him only 182:2 Now it is because we are ourselves through its effects, that the essence of scien- material things that our criteria for estab- tific discovery lies. This search necessitates lishing the reality of things turn on the ca- the construction of both new concepts and pacity of the thing whose existence is in new tools. But, as what is produced must doubt to bring about (or suffer) changes in possess a material cause, the scientist stands its material constitution or the constitution for his essential task, in two systems of so- of some material thing. Space, for exam- cial relationships, depending necessarily on ple, might be regarded quite abstractly just 246 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY the work of others. the traditional empirical realist ones?

183:2 I think it has at least four substan- 183:1 I have argued that the concept of tial advantages. First, it reveals the essen- natural necessity is the concept of a real tial movement of science. Second, it allows generative mechanism at work, a concept room for the location of a surplus-element, which is applicable to the world quite inde- reflecting a difference independent of men, in pendently of men. And it is in virtue of their the analysis of law-like statements at any one connection by such a mechanism, of which level. Third, it alone is capable of sustaining knowledge may be attained in the social ac- the ideas of the necessity and universality of tivity of science, that necessity is properly laws, which are necessary for the rationality ascribed to some but not other sequences. of theory-construction and the intelligibility In paragraph 5 I will analyse and criticize of experimental activity. Finally, it alone is some objections to this concept of natural capable of accommodating the possibility of necessity and the related concept of natural the existence of entities and the necessary kinds. But I want to deal here with the fol- phase of the knowledge of entities which can- lowing basic objection to the account I have not be analysed as substances with qualities, proposed: If, as I have contended, at each but must be conceived as powers to produce stratum or level of reality an entity is identi- effects, powers which are possessed and may fied and its behaviour is described what pos- be exercised quite independently of their de- itive advantages does this account have over tection. (Needless to say, these advantages 3.3. NATURAL NECESSITY AND NATURAL KINDS: THE STRATIFICATION OF NATURE AND THE STRATIFICATION OF SCIENCE247 are not independent of each other.) of accidents. 184:1 Now it is our knowledge of the rea- 183:3/o Science never stops still for a mo- sons at Stratum II for the behaviour at Stra- ment. At whatever level we look, it always tum I that warrants our designating the be- involves something more than the empirical haviour as necessary. But the reasons for the realist concedes. For example, if we consider behaviour at Stratum II cannot be collapsed the phase of the identification of a protolaw into the behaviour at Stratum I or an inter- (which seems prima facie most susceptible to pretation or model of that behaviour consis- empirical realist analysis), we find the cate- tently with the intelligibility and rationality gorical clause implicit in a powers ascription, (respectively) of theory-construction or the representing the scientist’s instantaneous re- possibility of empirical test. Nor can such sponse to this situation, indicating his belief reasons be glossed simply as more fundamen- in the existence of a reason, located at the tal regularities,57 if they are to be subject next highest level of inquiry, for the predi- to experimental confirmation (or corrobora- cates being conjoined. Only the powers con- tion) . ceptual system is capable of giving an ac- count of the internal rationality of science, 184:2 I have already shown in detail that by which it moves from knowledge of one the empirical realist account of laws, and stratum to knowledge of the next, so display- hence the ontology that underpins it, is de- ing the actual historical development of the fective. Laws, I have argued, cannot be in- sciences as something other than a sequence terpreted as conjunctions of events, but must 248 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY be analysed as tendencies of things. If sci- mechanisms, I have argued, must be anal- ence is to be rendered intelligible the world ysed as the ways of acting of things; and their must be seen as one of persisting things, of operations must be understood in terms of differing degrees of structure and complex- the exercise of tendencies and causal pow- ity, to which powers and tendencies are as- ers. Tendencies may be possessed unexer- cribed; it cannot be reconstructed as a world cised, exercised unrealized, and realized un- of atomistic events apprehended in sense- perceived (or undetected) by men. experience. Briefly, to summarize my ac- count of laws: To invoke a law I must have grounds for supposing a generative mecha- nism at work. These comprise: (a) indepen- dent grounds, preferably under experimen- 184:3/o Finally, the empirical realist can- tally closed conditions, for the mode of op- not deal with the case of entities which just eration of the mechanism; (b) grounds for are their powers or about which all we know the satisfaction of the antecedent (or stimu- are their powers. He thus rules out dog- lus) conditions for the operation of the mech- matically, tout court the possibility of a cer- anism on the particular occasion in ques- tain kind of entity and a necessary phase of tion; and (c) the absence of specific grounds knowledge. In virtue of this he is no more for supposing a breakdown or transformation able to make sense of the frontiers of knowl- of the mechanism in that case. Generative edge, than show the mechanism by which sci- ences if it can and when it does, will advance. 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE249

3.4 The Social Produc- Science operates on given materials, includ- ing pre-existing theory and antecedently es- tion of Knowledge tablished facts, with given materials, i.e. by by Means of Knowl- means of an ensemble of intellectual and technical tools (including among the former edge paradigms, models, metaphors and analo- gies), producing new theories and facts. This section is omitted in the reprint of this chapter in [ABC+98]. 185:2 Science is produced by the imagi- 185:1 The basic conception of scientific native and disciplined work of men on what activity that I have been concerned to ad- is given to them. But the instruments of vance here is that is (consists in or involves) the imagination are themselves provided by work. Science, I have argued, must be knowledge. Thus knowledge is produced conceived as an ongoing process of trans- by means of knowledge. The objects from, formation, continually or essentially in mo- and by, which knowledge is generated are tion, in an attempt to capture (i.e. pene- thus always themselves social products (as trate and describe) the stratification of the is the knowledge generated). Thus science world. The logical structure of work is Aris- as a process is always entirely intrinsic to totelian. It depends, in particular, upon the ‘thought’. However, by perception and ex- co-presence in any given productive episode periment access to objects, viz. things and of both a material and an efficient cause. causal structures, existing independently of 250 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY thought may be obtained.58 And of such ob- But knowledge of existence, I have argued, jects knowledge may be achieved. Science is cannot be identified with demonstration of not an epiphenomenon of nature, for knowl- it. Causal powers, for example, can only be edge possesses a material cause of its own known, not shown to exist. Hence if, as I kind. But neither is nature a product of have suggested there are grounds for suppos- man, for the intelligibility of the scientific ac- ing, the ultimate entities in any one branch tivities of perception and experiment presup- of science are bare powers, they must nec- poses the intransitive and structured charac- essarily be undemonstrable. However, under ter of the objects of knowledge, viz. that they certain conditions, some states of things may exist and act independently of the operations be perceived, unaided or with the help of of men and the patterns of events alike. sense-extending equipment; and some causes may be demonstrated indirectly, i.e. through 186:1 Thus science, I have argued, presup- the ostension of their effects. But for an ex- poses the ontological independence and the istential or a dispositional claim to be con- possible disjuncture of the domains of the firmed or corroborated the states (or effects) real, the actual and the empirical at every and behaviour must be recognised or iden- stratum or level of reality. At each stra- tified as being of the asserted type. Hence tum scientists attempt to identify the enti- in general two kinds of criteria, viz. demon- ties responsible for what happens at the less strative and recognitive, must be satisfied fundamental stratum (their point of depar- for such a claim to be granted. Because ture) and describe their normic behaviour. 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE251 the theoretical and technical conditions un- us to identify them. Thus there is a di- der which such claims are made (and criteria rect link between the dynamic realist thesis elaborated) are themselves developing, our that the things and causal structures of na- knowledge may be extended; and because ture not only exist but act independently of they may be falsified if the criteria are not men and the conception of science as a so- met (or revised) our knowledge may be cor- cial activity sui generis in which both the rected. facts and the conjunctions that, when at- tainable, provide the empirical grounds for 186:2 The paradoxical air of talking causal laws are seen as social products. In of the correction of knowledge vanishes classical empiricism, in a subtle interchange, once the demand for extra-theoretical truth these ideas are crossed: so that facts and and intertheoretical synonymity is rejected. their conjunctions appear as naturally given Progress, I shall contend, can be shown to and things and causal structures as experi- have occurred but only from some particular ences of men. (In transcendental idealism, position, some specific vantage point, as it the former is seen as in part imposed by, were, in theoretical time. and the latter as unknowable to, men.) Now 186:3/o Science is explanatory, not simply the identification of the conditions of (knowl- descriptive. Explanation is achieved by ref- edge of) being with the conditions of experi- erence to enduring mechanisms. Such mech- ence in empirical realism leaves ‘theory’ with anisms exist as the powers of things and act a very uncertain status. For it must be ei- independently of the conditions that enable 252 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY ther reduced to, or grounded a priori in some to be used (for scientific work). That sci- necessary condition of, experience; so that it ence is ongoing implies that some individu- is either reducible or immutable. For tran- als do so. Knowledge shares a feature com- scendental realism theory is both irreducible mon to many social products then: namely and mutable. It is always there and liable that though it exists only in virtue of hu- to change, as part of our socially innate in- man activity, it is irreducible to the acts of tellectual endowment. It is this endowment men. For any cognitive act to be possible that we must draw upon as we attempt to there must be a material cause; some knowl- deepen our knowledge of the way things are edge established, given to us, already pro- and act in the world; and in so doing we can duced. No sum of individual cognitive acts continually add to and modify it. The exis- can yield knowledge, for the first member of tence of this stock, as a layered structure, is the series would already presuppose it. Ex- a necessary feature of any human cognitive perience is, on the other hand, susceptible to situation; so knowledge can never be seen as a purely individualistic analysis; ‘mass expe- a function of individual sense-experience. rience’ is clearly derivative and analysable as ‘the experiences of masses of individuals’. 187:1/o The necessity for a scientific train- It can thus be seen that underpinning em- ing shows that knowledge is a social product pirical realism is an epistemological individ- and cannot be conceived as purely individual ualism. That knowledge is not analysable acquisition. For it always stands to the in- in terms of individual experiences does not dividual as something that must be acquired 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE253 imply that it is not analysable in terms of completely definable in terms of experience; experience. But that the latter is the case i.e. it must consist of a web of empirical and can be seen by reflecting upon the consid- theoretical ideas. eration that the antecedent cognitive situ- ation of the individual would have at the 188:1/o Recent work establishing that sci- very least to contain one theoretical conjec- ence has a transitive (or sociological) dimen- ture, viz. that there were experiences of oth- sion and some facts about its nature has ers. Assuming that the category of experi- been widely regarded as shocking. That sci- ence was allowed to apply intersubjectively, ence is a social activity which shares many if all terms were explicitly (ostensively or op- of the characteristics of, and does not ex- erationally) defined all truths would be an- ist in isolation from others; that it depends alytic and all falsehoods contradictory, as a upon a whole complex of institutions, some claim about the facts would be implicit in of which have little interest in knowledge the meaning of every theoretical conjecture, ‘for its own sake’; and that in particular so that the point of appealing to experience circumstances its fortunes can depend upon would be lost; and theory would be incapable the accidents of particular men raises seri- of growing and developing. In short, if an- ous moral and political questions. Moreover tecedently established knowledge is to be ca- there are some particularly disturbing fea- pable of functioning as a material cause, the tures about current science. One need only layered stock must contain some terms not invoke, from the recent literature, the epi- thets of entrepreneurial science and shoddy 254 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY science, reckless science and dirty science,59 longer be distinguished by their content. For government science60 and mob science,61 re- experiences and the facts they generate must pressive science, Stalinist science and their now be viewed as socially produced and anodyne anarchist science62 to appreciate what is socially produced is socially change- this. However these problems do not flow able. There are no absolutely privileged from the social character of science per se, statements. The application of the cate- i.e. the mere existence of a transitive dimen- gory ‘empirical’ becomes relative and theory- sion, but from the present character of its so- dependent. Hence it cannot be used, with- cial character. The realization that science out a degree of circularity, to establish the has social problems could only be shocking if scientificity of one class of statements with one had been tacitly viewing it, in the style respect to another. Knowledge, viewed as a of Hume, as a kind of behavioural response transitive process, has no foundation—only to the stimulus of given facts and their con- a structure in time. The sciences have his- junctions. This is the positivist concept of tories, which like all histories are character- behaviourist or automatic science. It is a ized by both continuity and change; and in concept which can itself be used to disguise which, as in all histories, certain events stand embarrassing facets and rationalize the prac- out retrospectively as especially significant, tice of a science. e.g. the discovery of oxygen, the publication of The Origin of the Species, the Michelson- 189:1 Recognition of the transitive dimen- Morley experiment. (Later we shall have sion implies that scientific beliefs can no 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE255 to inquire into the grounds for their signifi- nomena in nature. cance.) 189:3/o In paragraph 3 I described science as the systematic attempt to capture the 189:2 Now the fact that scientists do stratification of the world. Only the concept not possess a special attitude or a superior of a real stratification allows us to sustain morality does not mean that science does not the idea of scientific progress, in a way which have a rationality of its own. Nor does the is both non-inductivist and consistent with fact that scientific beliefs cannot be distin- the possibility of scientific change. Knowl- guished by their content imply that scientific edge of new strata does not dissolve, though activities cannot be distinguished by their it may occasion a correction of, knowledge structure or their aim. There are two er- of old strata. Nor does it render the less rors here: the first is to suppose that sci- fundamental strata illusory. We do not need ence is not a social activity in the fullest the metaphor to which so many writers have sense (exactly what this entails we have yet found it necessary to resort in order to rec- to see). The second is to suppose that it oncile progress with change: viz. that of an is not equally a social activity, quite unlike asymptotic approach to the truth. any other, sui generis: namely a social ac- tivity whose aim is the production of the 190:1 Two fallacies must be most assidu- knowledge, with the cognitive tools at its ously avoided. The first is to suppose that disposal, of the enduring and transfactually science grows but does not change. The active mechanisms of the production of phe- second is to suppose that science changes 256 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY but does not grow. It is the fact that sci- science, not the intransitive world of things. ence grows, i.e. that in the transitive pro- Does this mean that it is wrong to talk of the cess of science new strata and dimensions scientist explaining events describing mecha- (or branches) of reality are discovered, that nisms, etc.? No: provided we remember that means that scientific change can be accom- what is explained in any concrete scientific modated as a fact of history without sac- episode is always the event known under a rificing the idea of scientific progress. But particular description. This does not mean progress can only be shown to have occurred that the event is, or that we must think of it from some substantive theoretical stand- as if it were its description. On the contrary, point or position. There is no Archimedean the ontological independence of the event is point outside theoretical time. But knowl- a condition of the intelligibility of its descrip- edge changes as it grows. For knowledge at a tion. But here, as elsewhere, it is the task of new level may lead to a revision, correction philosophy to analyse concepts, such as that or modification of knowledge at the previ- of an event, which can only be used syncat- ous level. For what is explained is never the egorematically in science. ‘pure’ phenomena, but always the phenom- ena read in a certain way: i.e. facts. The 190:2/o Scientific activity is continuous. scientist seeks to describe the mechanisms This has meant that ‘refutations’ have nor- generating the phenomena; but the results mally taken the form of ‘replacements’. Now of his activity belong to the social world of it has been pointed out that cases of both ‘inconsistency’ and ‘meaning-change’ can be 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE257 drawn from the history of the sciences. For ing must be shared in common.65 But that example, Newtonian physics corrected Ke- the subsequent divergent development of the pler’s and Galileo’s laws;63 and the concepts theories may result in their eventually be- of ‘mass’ employed in classical dynamics and coming ‘incommensurable’. So that ‘incon- the theory of relativity are radically differ- sistency’ and ‘incommensurability’ refer to ent.64 Now given that these are the most ob- distinct moments of the scientific process.66 vious ways in which scientific changes occur, (Something similar must be true of the nor- great care must be taken about the way in mal process of education.) which they, and their relationship, are for- mulated. Kuhn, Feyerabend, and others, 191:1/o Theory without experiment is have claimed that theories may be so rad- empty. Experiment without theory is blind. ically different in meaning as to be literally But in the historical development of the sci- ‘incommensurable’. To this there is the ob- ences experiment and theory are often out vious objection that if they were literally in- of step. Michelson and Morley did not see commensurable, i.e. shared no elements of their experiment as a refutation of the other, meaning in common, it is difficult to see how and Michelson never in fact relinquished his scientists could have had grounds for prefer- belief in it.67 On the other hand Prout’s hy- ring one to another. It is clear that at the pothesis could not be vindicated until the moment of ‘falsification’, when one theory is invention of physical techniques of chemical replaced by another, some elements of mean- separation, a century after its formulation. It is not just the experimental results but 258 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY what is done with them that counts. On the ing kind. The discovery of X-rays illustrates other hand, the nonfulfilment of experimen- both these facets. Indeed Lord Kelvin ini- tal expectations can always be explained in tially thought that Roentgen had devised an terms of the deficiency of experimental tech- elaborate hoax.68 (b) consists in the appli- niques. cation of the discovery and/or change to ac- count for (and perhaps correct) currently es- 192:1 Scientific activity is itself differen- tablished facts and generate new ones. tiated into periods or better ‘phases’ (so as not to identify them with chronologi- 192:2/o Needless confusion has been en- cal time): viz. into (a) phases of discovery gendered by the failure to distinguish mod- and/or change and (b) phases of r applica- els, theories, paradigms, etc. Very roughly, tion. Both are necessary. I use these charac- a theory is a model with existential commit- terizations in preference to the emotive and ment; that is, a model conceived, and meant somewhat misleading terms ‘revolutionary’ to be taken, as true; i.e. a model in which the and ‘normal’. (a) consists in the produc- entities posited and mechanisms described tion of the knowledge of a new stratum or are conceived as real. It is relatively easy level and/or the radical revision of knowl- for the scientist to invent models, but much edge at the current one. This is often pre- more difficult for him to construct theories. ceded by the hint or glimpse of a new level There were several models of the aether, but or by a crisis induced by the proliferation never a satisfactory theory of it. Diagram 3.7 of anomalous facts of a particularly disturb- illustrates a schema for the development of 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE259 science. A general conceptual scheme (ab- larian metaphysical paradigm. breviated here to G.C.S.) or metaphysical framework, such as that provided by atom- 193:1/o Puzzles or problems are the con- ism, ‘begets’ (logically, not temporally) a re- crete working data of the scientist. His im- search programme, such as that associated mediate task is their clarification and res- with the attempt to explain phenomena by olution. For this he must use the tools reference to the primary qualities of matter. he has at his disposal: established results, The research programme in turn generates facts, and theories, promising hypotheses a theory and/or a sequence of theories ei- and half-tried (or forgotten) ideas, the avail- ther intended for different fields (or different able formal and technical equipment and strata) and/or in competition with and re- usually some methodological (or heuristic) placing each other. At the centre of theory- paradigm, concretely embodied perhaps in construction is the process of model-building some exemplary piece of work. It is be- and technical innovation required for the em- cause such paradigms are shared that there pirical testing of the various models. These are intersubjective criteria for assessing the levels are never in complete harmony. Some scientific adequacy of his work. The scien- hypotheses, seemingly necessitated by the tist’s work normally takes place within a def- facts, are always out of line with the theory inite institution, a disciplinary matrix and generating the facts. Gravity, for example, is governed by what some continental writ- could never be assimilated to the corpuscu- ers have called the problematic of his sci- ence (that is, roughly the structured field 260 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY within which alone meaningful questions can taken from an adjacent field, as in the case of be asked or problems posed, expressing the Pasteur who was convinced, in opposition to dominant theoretical concerns of the time). orthodox opinion at the time, that fermenta- The research programme may have its own tion could not be caused by chemical agents methodological paradigm, such as Newton’s alone but had to be explained by dissym- Principia or Durkheim’s Suicide. Now if metrical forces associated with the activity the problems generated by work at phase (a) of living organisms.69 cannot be resolved by the material available 193:2 Diagram 3.7. The Internal Struc- to the scientist in his own field he must nec- ture of Science essarily draw on another. This provides the rationale for paramorphic model-building, a 193:3 Nb. so called ‘non-normal’ science role for analogies and metaphors (the mod- corresponds to phase (a) viz. of discovery els of discursive thought), leading to existen- and change: in it models are invented and tial novelty (e.g. ‘what are the punctuation subjected to empirical tests with the aim of marks in a genetic code?’). Incidentally, this theory construction. so called normal’ sci- also helps to explain why theoretical inno- ence corresponds to phase (b) viz. of ap- vations are often made by individuals origi- plication: in it puzzles generated by phase nally working in fields adjacent to the field (a) are resolved. the structure of theory- they innovate, as in the case of Dalton (a construction: models + facts + theories + meteorologist); or with a strong ‘prejudice’ techniques [within G.C.S. + Res. Prog.] –¿ new theories. the structure of puzzle-solving: 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE261 facts + theories + techniques + method- anisms of nature. Now from the perspective ological (and heuristic) paradigms –¿ reso- advanced here an event, such as the discov- lution of puzzles. critical science has no for- ery of oxygen, is significant not just because mal structure but may call into question any it refuted the phlogiston theory of combus- level, including established theories, success- tion, but because it constituted a decisive ful research programmes and even the G.C.S. moment in the transformation from one way or metaphysical framework, leading to the of doing chemistry, viz. that associated with replacement or development of theory, re- the theory of elective affinity, to another, viz. search programme or G.C.S. (e.g. Mach, Ein- that represented by post-Daltonian atomic stein). chemistry; that is, because it constituted a transformation in the ongoing activity of 194:1/o Conceiving science as work read- chemistry. ily lends itself to Aristotelian schematiza- tion.70 The material cause is antecedently 195:1 Knowledge does not exist in a third established knowledge, facts and theories; world.71 Rather, it exists in our world, em- the efficient cause is the methodological bedded in the scientific community. Without paradigm or generative theory at work in the men there would be no knowledge, only its theoretical and experimental activity of men; traces. In this sense it depends upon men. the formal cause new knowledge, facts and But though it exists only in virtue of the theories; and the final cause knowledge of thoughts and actions and products of men, the enduring and transfactually active mech- it is irreducible to them. For though it would 262 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY not exist without the activity of some men, losophy of science would depend not only its pre-existence is a necessary condition of upon the development of an adequate philo- the activity of anyone. It is a public mix sophical ontology, but upon the development that always antedates the individual. Now of an adequate philosophical sociology too. it is not necessary that science should con- This must consist in an answer to the ques- tinue, i.e. be ongoing. It is contingent that tion: what must society be like if science (as it is. But given this men must reproduce (or a specific kind of social activity) is to be pos- more or less transform) the knowledge that sible? It must satisfy the desiderata of being is given to them. Men do not construct their a structure irreducible to but present only in knowledge: they reproduce or transform it. its effects. Society can only be known, not This is another way of saying that any knowl- shown, to exist. It exists only in virtue of edge that there is must possess a material the intentional activity of men but it is not cause. Now in general for scientific activity the result (or the cause) of their intentional to be continuous all levels depicted in Dia- activity. Sociology and psychology thus con- gram 3.7 must be represented. The trouble stitute distinct branches of science, in the with social science, for example, is not that it sense of the criterion developed in paragraph has no (or too many) paradigms or research 3 above. Sociology is not concerned with programmes; but rather that it lacks an ad- masses of individuals or mass behaviour; but equate general conceptual scheme. with the persistent relationships between in- dividuals. Such relationships would not exist 195:2/o I said in 1.6 that an adequate phi- 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE263 without their relata but they do not depend myth of creation (logical or historical), i.e. for their efficacy upon any particular relata, the possibility of a methodologically individ- any particular named individuals. ualist reduction. It is not necessary that that society should continue. But if it is to do so 196:1 Now the autonomy of sociology and then men must reproduce (or more or less psychology accords well with our intuitions. transform) the structures (languages, forms Thus we do not suppose that the reason of economic and political organization, sys- why the war is fought is the soldier’s rea- tems of belief, cultural and ethical norms, son for fighting it, just as the reason why etc.) that are given to them. The Newto- the bar of chocolate is wrapped need not be nian revolution in sociology consists in com- the chocolate wrapper’s reason for wrapping ing to see that it is not necessary to explain it (though it depends upon the latter). I do society as such; but only the various struc- not have to know the laws of supply and de- tures responsible for different societies and mand to buy a mackintosh or to know the their changes. The problem of how men re- deep structure of language in order to use it. produce any particular society belongs to a The deep structure of language may indeed linking science of social psychology. As so impose limits (like natural structures) upon conceived, society may be regarded as an en- the kinds of speech acts I can perform but it semble of powers which exist, unlike other does not determine what I say. This concep- powers, only as long as they are exercised; tion of social science thus preserves the sta- and are continually exercised via (i.e. in the tus of human agency, but does away with the 264 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY last instance through) the intentional action always there; which men must reproduce or of men.72 partially transform; but which would not ex- ist without its ‘functionaries’. It is method- 196:2/o Established facts are social prod- ologically incorrect to search for an efficient ucts. Understanding their logic may help cause of society, though society depends nec- us to clarify the relationships between men essarily upon the efficient activity of men. and society and men and the world. Here But a reading depends upon antecedent so- the metaphor of a reading may be used. cial activity; the acquisition of a language by Its adequacy depends upon the existence of the reader. It is in this sense that the facts both a given language and an independent always depend upon social activity. In ex- text. (But the metaphor is misleading in one perience the skilled scientist reads the world way: in that the text of nature exists inde- as if it were a text in an attempt to under- pendently of any language.) It is this that stand the mechanisms of the production of makes possible talk of a correct, rather than phenomena in nature. But his own reading just commutatively successful, reading. Men depends upon the mechanisms of the repro- never create this language. For it always duction and transformation of language, of preexists them. But it exists as an actual, knowledge and of society. i.e. ‘living’, language only in virtue of, and changes with, their uses of it. Thus if soci- 197:1/o To return now to science: it is ety is represented by the model of a language contingent that science is continuous but it may be regarded as a structure which is given that it is men must reproduce (or 3.4. THE SOCIAL PRODUCTION OF KNOWLEDGE BY MEANS OF KNOWLEDGE265 more or less transform) the knowledge that ditional nothing can be. If all knowledge is is given to them. The condition that sci- (equally) conjectural, no statement can be ence be continuous is equivalent to the con- refuted. Of course it is always possible that dition that all knowledge possesses a mate- the scientist is mistaken in any particular be- rial cause. Hence the criterion of adequacy in lief (and a good scientist is continually alive the transitive dimension of the philosophy of to this possibility). But in order to demon- science is that the account of science should strate a mistake some proposition must be be capable of sustaining the concept of it as asserted (some theory accepted and frame- an ongoing social activity. It is here that work worked within). In order to learn from dyadic theories of falsification, that is, the- our mistakes we must know that (and when) ories that conceive falsification as consisting we are mistaken. Lacking from fallibilism, as in a confrontation between a single theory from classical empiricism, is the key concept and a set of facts fail. If science is to be con- of knowledge necessarily possessing a mate- tinuous, refutations must be replacements; rial cause: antecedently established knowl- which means that always more than one the- edges; science’s means of production. It is ory must be involved. But, related to this, not necessary that a scientist works within is an even more basic objection to fallibilism an particular framework or accept any par- as such. For the refutation of any theory ticular theory; but it is necessary that he presupposes the acceptance of the refuting works within (accepts) one. observation statement. If everything is con- 198:1 Underpinning empirical realism is a 266 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY model of man in which men are seen as sen- define the world then the world must be sim- sors of given facts and recorders of their con- ilarly composed: of atomistic and discrete stant conjunctions: passive spectators of a events (or momentary states) independent of given world rather than active agents in a each other. If knowledge is to have its foun- complex one. This model plays a role at least dations in experience and experience is to de- as important as that played by the classical fine the world then both the ultimate items paradigm of action and the celestial closure and objects of knowledge must be atomistic discussed in Chapter 2. Together they form and independent of each other. This creates a complementary triangle (see Diagram 3.8) the problem of what grounds we can have for moving from the observed to the unob- 198:2 Diagram 3.8. Three Sources of Em- served, or from the actual to the possible pirical Realism (and thence to the counter-factual). Falli- 198:3/o Implicit in empirical realism is a bilism, which shares this model, can no less conflation between a ground of knowledge, escape this problem. For the refutation of viz. experience, and the world. If experience theory T at time t1 by an observation state- is to be capable of playing the role tradition- ment is consistent with its corroboration by ally assigned to it of grounding our knowl- that statement after t1 unless we are justified edge (in whole or in part) then the items of in moving from the observed to the unob- which it is composed must be perfectly sim- served, and from the actual to the possible. ple and atomistic; i.e. insusceptible to fur- (Unless induction is justified or nature is uni- ther analysis or justification. But if it is to 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED267 form we can never know that a ‘mistake’ is knowledge of, necessary connections between a mistake; so we can never put our mistakes matters of fact; behind us.) 199:4 (ii) if there were necessary correc- tions between matters of fact they would have to be known a priori; so science could 3.5 Objections to the not be empirical; Account of Natural 199:5 (iii) men are never directly aware of any causal power or agency or necessary Necessity Proposed connections between matters of fact, so these concepts cannot be justified by experience 199:1 Having outlined the principal advan- (though they may be explained by it; or are, tages of my account of natural necessity and for the neo-Kantian, imposed upon it). natural kinds (on pp. 183–5 above), I now 199:6 The argument for thesis (i) is typ- want to consider some objections to it. In ically constructed as follows: there is noth- Chapter 4 I will consider the conditions of ing inconsistent about the supposition that the plausibility of these objections. the cause of a phenomenon, say putting a 199:2 The chief Humean counter- kettle of water on the stove and heating it, arguments may be put in the form of three should not be accompanied by the effect in theses: – question. It is conceivable that water might 199:3 (i) there can be no, or at least no freeze instead of boil when it is heated. Now 268 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY thesis (i) is, as stated, highly ambiguous. It terms of our knowledge of the necessity of the is not clear whether it is an ontological or an connection would be to commit the epistemic epistemological thesis (this ambiguity is of fallacy (see 1.4 above). There is a real differ- course explicit in the way I have formulated ence, quite independent of men, between the it); whether the ‘necessity’ is logical or non- fact that when I heat the kettle of water it logical; and whether the ‘matters of fact’ are boils and the fact that it boils when the time events and states of affairs or the statements is half-past two or the colour of my socks is describing them. Before returning to the ar- blue. The necessary connections that bind gument, we must see exactly what is at stake some but not other events together (which in it. are the enduring mechanisms of nature) are quite independent of our knowledge of them. 200:1 Now, it will be remembered, that for the transcendental realist to say that a 200:2 Statements clearly belong to the sequence Ea.Eb is necessary is to say that epistemic not the ontological order; and log- there is a generative mechanism at work such ical connections hold only between state- that when Ea occurs Eb tends to be pro- ments, not between events and states of af- duced (is produced in the absence of inter- fairs. Hence the prima facie absurdity of fering causes). If there is such a mechanism those who, in attempting to refute Hume, the sequence is necessary; and its necessity try to establish that nomic necessity is, or is quite independent of any knowledge of it. may be, a species of logical necessity.73 Nat- To analyse the necessity of the connection in ural necessity is not logical necessity. Natu- 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED269 ral connections hold between things, events, ence; so that statements can be classified as states of affairs and the like; logical connec- definitions, deductive consequences of true tions between propositions. Moreover there theories and simple protolaws according to could be a world without propositions, in the position they occupy (at any moment of which the concept ‘logical connection’ had time) in the development of our knowledge. no application. The laws of logic are not Hence the deducibility of a tendency from a features of the world, nor are they imposed nature may serve as a criterion at the Lock- upon it. Rather, we must say: the world is ean level for our knowledge of natural neces- such that changes in it can be consistently sity, just as a correct definition may serve described. as a criterion at the Leibnizian level for our knowledge of natural kinds. But whether 200:3/o Neither natural necessity nor or not a sequence of events is necessary is knowledge of natural necessity can be iden- quite independent of the logical status of the tified with logical necessity. But our capac- proposition used to express it; which is a ity to deduce the Wiedmann-Franz law from function of the way it is described in the con- Drude’s theory of electrical conductivity may text of our knowledge; which in turn may be serve as a criterion of our knowledge of the shown to have a certain rationale in the de- necessity the theory describes. I suggested velopment of science. in paragraph 3 above that three levels of knowledge of the objective world order can 201:1 Some causal statements expressing be distinguished in the development of sci- necessary connections are logically necessary 270 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY and some are logically contingent.74 For the essary connections between some (but not Humean, however, logical and natural ne- other) events, science would be impossible; cessity are easily confused. For given the and that in science the most stringent cri- isomorphic relationship between knowledge teria for knowledge of natural necessity may and the world assumed in empirical realism be satisfied. and restricting our knowledge of nature to the protolegal phase of science (see page 172 201:3 Unless there were necessary connec- above) he naturally comes to regard relation- tions between matters of fact neither con- ships between events as characterizable in firmation nor falsification would be possi- the same kind of way as the statements ex- ble. For without them no confirmation in- pressing their relationships are at that phase stance adds any probability whatever to any typically, though not invariably, character- inductive instance.75 On the other hand for ized; namely as contingent. But it is into it to be rational to reject what is falsified this very same trap that defenders of the en- it must be assumed that a hypothesis which tailment view of natural necessity fall. has been false in the past will not suddenly become true in the future.76 Whether the 201:2 I shall construe thesis (i) as an epis- conclusions of inductive arguments are weak- temological claim to the effect that knowl- ened to probability judgements or it is denied edge of necessary connections between events that science is inductive in nature there must is impossible. And I will attempt to refute be necessary connections between matters of it by arguing that unless there were nec- fact. Such necessary connections are pro- 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED271 vided by enduring mechanisms. Moreover, if nothing other than the powers or ways of experimental science is to be possible, there acting of things. Thus, if science is to be pos- must be necessary connections between some sible, there must be a relationship of natural but not other events. This implies a dynamic necessity between what a thing is and what principle of indifference: to the effect that a thing can do; and hence between what a mechanisms not only endure but are trans- thing is and what it tends to do, in appro- factually active. Neither their enduring nor priate conditions. The deducibility of a ten- their transfactual activity is in need of ex- dency from a nature thus constitutes a crite- planation. rion for our knowledge of natural necessity. Events are necessarily connected when nat- 202:1 Unless there were necessary connec- ural tendencies are realized. tions between matters of fact we could have no knowledge, even particular knowledge (in 202:3/o With this in mind, let us return as much as this depends upon inferences be- to a detailed examination of the argument yond what is immediately observed), of the for thesis (i). Is it conceivable that water world. For science to be possible then the should not boil when it is heated? Now it world must consist of enduring and transfac- might be said straightaway that it is incon- tually active mechanisms; and there must be ceivable to suppose that water might not boil necessary connections between some but not when it is heated. Since anything that did other matters of fact. not boil when it was heated could not prop- 202:2 Natural mechanisms are of course erly be said to be ‘water’ at all. That is 272 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY that, in Lockean terminology, ‘boiling when per’, ‘a metal’ or ‘petrol’ respectively; since heated’ specifies part of the nominal essence the point of referring to the particulars con- of water; or we could say with Putnam that cerned in those ways would be gone.78 A ‘water’ functions as a ‘law-cluster concept’.77 magnet that could not magnetize, a fire that Now the strength of this reply should not be cannot burn or a pen that can never write under-rated. I have no doubt that we should would not be ‘magnets’, ‘fires’ and ‘pens’ at ordinarily say something on these lines. In- all. Things must satisfy certain criteria for deed, unless we have some criteria for the them to be (correctly identified as) the kinds correct application of the term ‘water’ there of things they are. By far the most impor- is no reason why we should use it to refer tant of such criteria are those that depend to substances which as a matter of fact boil upon their powers to affect other bodies (a when heated rather than to say desk lamps class which may be extended, analytically, to or Saturday afternoons (which do not boil include their powers to affect observers un- when heated). And such criteria would be der specified conditions in certain standard at least in part dispositional; appearances, ways). notoriously, can be misleading. Litmus pa- per that does not turn red when dipped into 203:1 Such a reply will not however satisfy acid, a metal that does not conduct electric- the Humean (particularly if he believes that ity, or petrol that does not explode when definitions are merely matters of convention ignited could not be said to be ‘litmus pa- and cannot express empirically ascertained truths about kinds of things). More to the 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED273 point it will not satisfy the scientist: for, ac- ditions and a closed system (so as to elimi- cepting that ‘boiling when heated’ specifies nate the possibility of intervening causes) it part of the nominal essence of water, i.e. the would seem that we have the following alter- criteria for the identification of a substance natives: as ‘water’, he will want to know what it is 204:1 (a) our explanation was false; about water in virtue of which it boils when 204:2 (b) the fact that it was intended to it is heated. That is, he will set out to con- explain, viz. that water boils when heated, struct an explanation, in terms of the molec- was false ular and atomic structure of water, from which he can deduce its tendency to boil 204:3 (c) the particular concerned had when it is heated. Now it is clearly inconsis- been wrongly identified: it was not a sam- tent with this explanation to suppose that ple of water after all; water might freeze, blush shyly or do any- 204:4 (d) the particular concerned had thing else rather than boil when it is heated. changed; so that it had ceased to be water That is, if the explanation is correct water by the time it was heated. must boil when it is heated. 204:5 Now the Humean asks us to imag- ine, and inductive scepticism requires that 203:2/o Suppose however we came across it be possible, that the cause event occurs a stuff which in all other respects looked and and the effect event fails to materialize. Let behaved like water but which did not boil me call this the critical situation. Now I when it was heated. Assuming standard con- want to argue that, given only that possi- 274 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY

bility (a) is ruled out, so that we have a cor- the time it froze at t2. Here again it is not rect explanation, the critical situation is im- true to say that water did not boil when it possible; that is, it is not possible that the was heated. For by the time it froze it had cause event occurs and the effect event fails become something else, say ‘retaw’. Hence to materialize—in our example, that water given only the possibility of a realist interpre- is heated and does not boil. tation of the entities postulated in the expla- nation, the conditions for inductive scepti- 204:6 Let me show this. If the explana- cism cannot be satisfied. If there is a real rea- tion is correct water must boil rather than son, located in the nature of the stuff, inde- freeze when it is heated (though of course pendent of the disposition concerned, water the converse is not the case); so possibility must tend to boil when it is heated (though (b) is ruled out. Consider (c), the misiden- in an open world any particular prediction tification of the particular concerned. Now may be defeated). The stratification of na- in this case it is not true to say that wa- ture thus provides each science with its own ter did not boil when it was heated. For internal inductive warrant what did not boil was not water but only something which looked, and perhaps other- 204:7/o Now it might be objected that I wise behaved, like it, say ‘nwater’. Finally have omitted from my list of alternatives the consider (d), a change in the particular con- possibility of the explanation, though correct cerned: what was water when it was put into up to time t1, subsequently breaking down. the kettle at time t1, ceased to be water by But this possibility equally does not satisfy 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED275 the requirements of the critical case. For, ergy in microphysics) and for the agent or now at Stratum II (defining the Leibnizian mechanism which brought about the change. level of the particular movement of science The principles of substance and causality are with which we are here concerned), noth- interdependent and complementary. Things ing which did not possess the molecular and persist (and continue to act in their normal atomic structure that water has been discov- way) unless acted upon; and their changes ered to possess could be said to be ‘water’. are explained in terms of the action of per- So, here again, it would not be water that sisting (and transfactually active) things. was freezing. A stuff remains water only so If science is to be possible changes must long as its nature (or real essence) remains be transformations, not replacements; and unchanged. (Of course scientists could make transformations must be effected by the ac- a taxonomic change, but this does bear upon tions of causes (causal agents). Things can- the argument. not pass clean out of existence or events hap- pen for no reason at all. These are ideals of 205:1 It is of course possible that the na- reason. But if science is to be possible our ture of some particular will be transformed: world must be such that they hold. This en- in which event, scientists will search both for tails that it must be a world of enduring and an underlying substance or quasi-substance continually acting things. It is o course true which preserves material continuity through that it is impossible to prove that cases of ex change (e.g. a gene pool through species nihilo production and miracles cannot ever change, an atom in chemical reactions, en- 276 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY happen. All we can say is that they cannot its nature (in this case, molecular structure) be known to happen. For it always remains will be transformed, so that it ceases to be possible for the scientist that what appears an individual of that kind, and wondering to be a case of an ex nihilo production or whether, while remaining an individual of a miracle at time t1 can come eventually at that kind, it will cease to behave in the way t2 to be explained in terms of the transfor- that it has tended to behave in tr the past. mation of real things and the action of real The point is even clearer if we generalise it, causes upon them. so raising the questions of the boundaries of kinds and of the scope of application of 205:2/o I have argued that provided we laws. The difference is between wondering have a correct explanation the critical sit- whether water will cease to exist; and won- uation cannot occur; that, for example, as dering whether, while continuing to exist, it long as the particular stuff remains water it will stop boiling (in exactly the same circum- must tend to boil when heated. But it might stances) when it is heated. be urged if, as I have acknowledged, the nature of some particular may be changed 206:1 It might be objected that while what does this not open the floodgates of induc- I have said clearly covers case (d), viz. that tive scepticism once more? The answer is no: of a particular changing, I have not taken for there is a big difference between wonder- the possibility of case (c), viz. that of a par- ing whether some particular will be so acted ticular being misidentified, of nwater being upon by real causes in its environment that mistaken for water, seriously enough. What 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED277 is to prevent us continuously misidentifying ter boiling and its being heated. particulars in just this way? Now just as par- ticulars may be transformed, so they may 206:2/o Nevertheless despite this irrel- be misidentified. But the situation the in- evance to our present concern scepticism ductive sceptic asks us to imagine only gets about particular knowledge can and should of the ground if we assume that the rele- be met. It might be met in the following vant particulars have been correctly identi- way: Any argument in which the case for fied. The problem of induction is the prob- the general misidentification of particulars lem of what guarantee we have that the un- is stated itself presupposes the capacity to observed will resemble the observed, or the identify certain particulars, namely words as future the past; it is not the problem of what tokens of a type and hence possessing a cer- guarantee we have that we have correctly ob- tain standard meaning in a given context. served the observed or correctly described Hence no argument for the general misiden- the past. The suggestion that what I have tification of particulars can be consistently here may in fact be a piece of lead piping is stated. If this argument does not carry con- irrelevant to the question of what warrant I viction try to imagine a world in which we have for assuming that water will continue (a) systematically (b) at random misidenti- to boil when heated or for supposing that fied (α) some particulars (β) all particulars there is a necessary connection between wa- (a’) all the time (i) some of the time. A world in which we systematically misiden- tified some given class of particulars (such 278 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY as books as saucers and vice versa) would however, that if science it to be an ongoing just be a world in which objects had different concern it cannot persistently demand and names. But a world in which our misidenti- persistently return negative verdicts. fications were haphazard or universal is not coherently conceivable. It makes no sense to 207:1 It might be objected to my refuta- say that a particular has been misidentified tion of thesis (i) that I have not considered unless one is prepared to say in what respect the possibility that the explanation, which it has been misidentified. This itself presup- gives each science at any moment of time its poses the capacity to identify the particular own inductive warrant, is incorrect. Now it as of a certain type. Of course our capac- is of course always possible that we are mis- ity to identify particulars presupposes the taken in our explanation of why water must extended or dynamic principle of substance boil when heated; that our description of the enunciated above, namely that things per- mechanism in virtue of which it does so is sist and continue to act unless acted upon, wrong. But this is a general condition of and hence in this way it presupposes the all knowledge; it does not bear on the argu- existence of necessary connections between ment of thesis (i), which concerns the special matters of fact. It is up to the criteriology difficulty of knowledge of necessary connec- of empirical science to determine whether a tions between matters of fact. I have already particular has been misidentified or a per- argued against the idea that all knowledge ceptual report is nonveridical. The point is, is conjectural on the grounds that refuta- tions presuppose acceptances (progress re- 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED279 quires a material cause). But whether or ist when tables and chairs do; we need only not my account of the transitive dimension of worry about whether electrons will cease to the philosophy of science is accepted, refuta- exist when atoms do. It is contingent that tions presuppose necessary connections be- the world is such that science is possible. But tween matters of fact. given that it is the dynamic principles of sub- stance and causality that I have formulated 207:2/o I have argued that scepticism must be true of it. about change, about our capacity to identify 208:1 Three further forms of Heraclitean particulars and about the possibility of non- scepticism are possible in which we could be conjectural knowledge as such are all distinct invited to imagine that our world is replaced from the special kind of scepticism involved (a) by a totally different one; (b) by one in in thesis (i), which is scepticism about the which the principles of substance and causal- possibility of knowledge of necessary connec- ity no longer held; and (c) by one in which tions between matters of fact. I have shown science ceased to be possible. I shall argue how the second and third forms of scepti- that the replacements envisaged in (a) and cism, though irrelevant to thesis (i), may be (b) are impossible, but that I am precluded averted. But how can Heraclitean scepticism by my own premises from saying anything be countered? Changes in things, I have ar- about (c). gued, are explained in terms of unchanging things. The world is stratified. We need only 208:2/o In (a) it is supposed that our worry about whether atoms will cease to ex- world could be replaced by a totally differ- 280 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY ent one; but to which, once it had come into long after mankind has perished things will being, inductive techniques could be reap- persist and continue to interact in the world plied. Now this is not an intelligible suppo- that we once lived in. This leaves us with sition, not only because scientific continuity (c), about which I have said my premises would be lost during the replacement (so it preclude me from speaking. But a moment’s would make little sense to talk of reapplying reflection shows that (c) is devoid of interest inductive techniques), but because there is for us. It is an empty counterfactual. For no possible way in which such a replacement we know as a matter of fact that our world could be affected save by the action of real is one in which science is possible. Hence to causes.79 In (b) it is supposed that our world assert the possibility of a world without sci- might be replaced by one to which the princi- ence is merely to reassert the contingency of ples of substance and causality do not apply. the circumstance that makes a study of the Now although the existence of our world is conditions of the possibility of science possi- contingent, given that it exists the supposi- ble. tion that it might be replaced in this way is not an intelligible one. Transcendental real- 209:1 I have established that we can have ism demands that we reason from the effect, (and that science actually possesses) knowl- science, to the condition of its possibility, viz. edge of necessary connections between mat- a world of enduring and transfactually active ters of fact. And I have shown how inductive mechanisms. So we can rest assured that scepticism proper, namely that arising from the assumption of the possibility of what I 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED281 have called the critical situation, viz. the ral necessity, can come to be attained empir- occurrence of the cause event and the non- ically; that is, how a posteriori knowledge of occurrence of the effect, can be allayed, viz. natural necessity is possible. by the provison of an adequate explanation; and how the other forms of scepticism of- 209:3/o As there is some misunderstand- ten confused with inductive scepticism can ing about the role of the concept of essence be countered. I now turn to theses (ii) and (and, as we shall see, the nature of defini- (iii) which the Humean uses to bolster his tion) in science, some preliminary termino- central contention. logical clarification is necessary. The nomi- nal essence of a thing or substance consists of 209:2 Thesis (ii) alleges that if there were those properties the manifestation of which necessary connections between matters of are necessary for the thing to be correctly fact they would have to be known a pri- identified as one of a certain type. The ori, so that science could not be empirical. real essences of things and substances are It is clear that this argument trades on a those structures or constitutions in virtue tacit conflation of logical and natural neces- of which the thing or substance tends to sity and the identification of the resultant behave the way it does, including manifest concept with that of the a priori. To refute the properties that constitute its nominal it, I will have to show how knowledge of the essence. Science, I have argued, seeks to natures or real essences of things, which I explain the properties of things identified have argued ground our ascriptions of natu- at any one level of reality by reference to 282 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY their intrinsic structures, or the structures though it does not presuppose that the real of which they are an intrinsic part (defin- essence or nature is known. Thus we are ing the next level of inquiry). Thus the justified in classifying alsations, terriers and dispositional properties of say nickel, e.g. spaniels together as different varieties of the that it is magnetic, malleable, resistant to same species dog because we believe that rust, melts at 1445 degrees C and boils at they possess a common genetic constitution 2900 degrees C are explained, in the con- which, despite their manifest sensible differ- text of post-Daltonian atomic theory, by ref- ences, serves to differentiate them from the erence to such facts about its intrinsic struc- members of the species cat. A chemist will ture as that its atomic number is 28, its classify diamonds, graphite and black carbon atomic weight is 58.71 and its density is 8.90. together because he believes that they pos- The atomic constitution of nickel is its real sess a real essence in common, which may essence. But it was discovered a posteriori, be identified as the atomic (or electronic) in the transitive process of science. And it it- structure of carbon, of which these are al- self constituted an explanandum of the next lotropic forms. To classify a thing in a par- phase of scientific inquiry. ticular way in science is to commit oneself to a certain line of inquiry. Ex ante there 210:1 In general to classify a group of will be as many possible lines of inquiry as things together in science, to call them by manifest properties of a thing, but not all the same name, presupposes that they pos- will be equally promising. Thus if one’s con- sess a real essence or nature in common, 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED283 cern is to account for the manifest proper- thing. ties of cucumbers it is clearly preferable to classify a 12 in. long green cucumber un- 211:1 The distinction between real and der the sortal universal ‘cucumber’ rather nominal essences should not be confused than under the universals ‘green’ or ‘12 in. with that between real and nominal defi- long’. Not all general terms stand for nat- nitions. Real definitions are definitions of ural kinds or taxa; because not all general things, substances and concepts; nominal features of the world have a common expla- definitions are definitions of words. (Nom- nation. Carbon and dogs constitute natural inal essences are the properties that serve to kinds; but tables and chairs, red things and identify things). Real definitions, in science, blue, chunks of graphite and fuzzy dogs do are fallible attempts to capture in words the not. The justification of our systems of tax- real essences of things which have already onomy, of the ways we classify things, of the been identified (and are known under their nominal essences of things in science thus lies nominal essence) at any one stratum of re- in our belief in their fruitfulness in leading ality. As so conceived, they may be true or us to explanations in terms of the generative false (not just or even—more or less useful). mechanisms contained in their real essences. The atomic weight of copper is 63.5. It would Not all ways of classifying things are equally be wrong to claim that it was 53.4 or alter- promising because not all sets of properties natively that 63.5 was the atomic weight of individuate just one and only one kind of tin. Of course this fact was discovered a pos- teriori; but it may now be said to constitute 284 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY part of the real definition of copper. If the ing social activities require justification. real essence of copper consists in its atomic (or electronic) structure, its nominal essence might consist in its being a red sonorous 211:2/o On the view advanced here sci- metal, malleable and a good conductor of ence consists in a continuing dialectic be- electricity etc. Something that did not sat- tween taxonomic and explanatory knowl- isfy these properties could not properly be edge; knowledge of how the things there are said to be ‘copper’. But conversely just be- behave. It aims at real definitions of the cause the word ‘copper’ in science has a his- things and structures of the world as well as tory, and at any moment of time a use, the statements of their normic behaviour. The nominal essence of copper cannot suddenly source of the failure to see this is the ontol- be designated by the use of ‘reppoc’ or ‘tin’. ogy of empirical realism which reduces things Nominal definitions in science cannot there- to qualities, taxa to classes, enduring and ac- fore be conceived as stipulative, arbitrary or tive mechanisms to constant conjunctions of matters of convention. Although there is a independent and atomistic qualia.81 Now if sense in which any other symbol could have the world consists only of qualia and qualia been used to refer to copper; given this usage are independent of one another then the par- and that history ‘copper’ cannot be replaced ticular names that we give to qualia cannot by ‘bronze’ or ‘female’ for no reason at all.80 matter and all qualia will appear on a par. Changes in the definitions of words in ongo- On this conception, predicates must be inde- pendent of one another and classification is 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED285 ultimately arbitrary. nomena of the world, including (upon anal- ysis) the genesis and behaviour of ordinary 212:1 Now just as it is a mistake to as- things. The dialectic of explanatory and tax- sume that science is concerned with any and onomic knowledge must thus be formulated all behaviour it is a mistake to assume that as follows: science is concerned with the be- it is concerned with any and all things. Sci- haviour of things only in as much as it casts entists do not seek to describe the behaviour light upon their reasons for acting and hence of or to classify common objects like tables upon what kinds of things there are; and sci- and chairs, though the laws of physics and ence is only concerned with things of a par- the principles of scientific taxonomy (e.g. the ticular kind, in as much as they constitute identification of a table as an oak one) may the reason for some pattern of normic be- be brought to bear on them. Now from the haviour and thus themselves become an ap- fact that tables have no real essence it does propriate object of inquiry. not follow that carbon has none. Electrons are not related in the same way as games. A 212:2/o The importance of taxa in science resemblance theory of universals works best may be expressed by saying that what is non- for the complex Strawsonian individuals of accidentally true of a thing is true of a thing ordinary life. But the universals of inter- in virtue of its essential nature. A thing est to science are real: they are the gener- acts, or at least tends to act the way it is. ative mechanisms of nature which account, It should be stressed that the difference be- in their complex determination, for the phe- tween a thing which has the power or tends 286 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY to behave in a certain way and one which ‘what is an atom?’ but only with ques- does not is not a difference between what tions of the kind ‘how can the energy of the they will do, since it is contingent upon the sun be made useful?’ or ‘under what con- flux of conditions whether the power is ever ditions does an atom radiate light?’82 Pop- manifested or tendency exercised. Rather, it per’s ‘methodological nominalism’ seems to is a difference in what they themselves are; be based on the idea that to suppose that i.e. in their intrinsic natures A copper vase things have essences is to suppose that it remains malleable even if it is never pressed is possible to give explanations which are out of shape. It is contingent whether an ‘ultimate’ in the sense that they are insus- electric current is ever passed through a cop- ceptible in principle of further explanation per wire. But it is necessary, given its elec- (which is what he calls ‘essentialism’).83 Al- tronic structure, that it be a good conductor though Locke may have held this view, is it of electricity. We know how things will be- certainly no more a necessary feature of the have, if certain conditions materialize, if we concept of real essence than it is a necessary know what the things are. But we can only feature of the concept of behaviour to sup- know what things are a posteriori, via the pose that because a thing can be described as empirical process of science. behaving in a certain way the behaviour it- self cannot be subject to further explanation. 213:1 This view may be contrasted with It is clear that to suppose that things have the idea that scientists are not concerned real essences is not to suppose that the real with questions such as ‘what is energy?’ or 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED287 essences of those things cannot be explained solid objects fade gaseously away into empty in terms of more fundamental structures and space. The second involves the belief that things. to suppose that there are natural kinds is to suppose that these kinds are fixed, and 213:2/o Two other arguments sometimes is in particular to rule out the possibility of invoked against the concept of real essences a mechanism of evolution.86 Again, this is should be mentioned. The first depends completely unwarranted. For natures may upon the assumption that differences in na- change; and whether, and if so the ways in ture are continuous, not discrete; that ‘God which they do, are matters for substantive makes the spectrum, man makes the pigeon- scientific investigation. No spectrum exists holes’;84 so that ‘genera, species, essences, between men and apes but that does not pre- classes and so on are human creations’.85 I clude the possibility of a mechanism of evo- can find no possible warrant for such an as- lution (involving a whole sequence of ‘miss- sumption. Taken literally, it would imply ing links’). What happens in such cases is that a chromosome count is irrelevant in de- that biologists posit a novel entity, a gene termining the biological sex of an individual, pool, as the underlying continuant through that the class of the living is only conven- the species’ change. The objection is only tionally divided from the class of the dead, valid at the level of ultimate physical en- that the chemical elements reveal a con- tities since necessarily if such entities exist tinuous gradation in their properties, that tulips merge into rhododendron bushes and 288 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY they must be enduring. That is, a thing must tend to act the way it does if it is to be the kind of thing it is. If 214:1/o Scientists attempt to discover the a thing is a stick of gelignite it must explode real essences of things a posteriori, and to if certain conditions materialize. Since any- express their discoveries in real definitions thing that did not explode in those circum- of the natural kinds. From a description of stances would not be a stick of gelignite but the nature of a thing its behavioural tenden- some other substance. Now given the satis- cies can be deduced. When such tendencies faction of the criteria for the identification are realized the events describing the stim- of a substance, say water, and the recording, ulus or releasing conditions for the exercise preferably under experimentally closed con- of the tendency and its realization may be ditions, of its most significant and suggestive said to be necessarily connected. Thus sci- behavioural properties, scientists move im- entists can come to possess knowledge of nec- mediately to the construction and testing of essary connections between events as a result possible explanations for the protolaws iden- of an a posteriori process of discovery. Scien- tified. But if there is an explanation, located tists are not content to collect conjunctions m the nature of the stuff or the system of of events. Rather they try to discover the which the stuff is a part, whether or not it natures of things. Given this, no problem of is known by men, water must tend to boil induction can arise. Since it is not possible when it is heated. It is the real stratification for a thing to act inconsistently with its own of nature that justifies induction in science. nature and remain the kind of thing it is. 3.5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY PROPOSED289

It is not we that impose uniformities upon causal agents in a world of other causal the world, but nature that makes induction agents; and that unless we were so aware we (properly circumscribed) a rational activity could not act intentionally, or come to know for men. ourselves as causal agents at all. (Projec- tive explanations of our idea of necessary 215:1 The third Humean counter- connection are clearly anthropocentric.) argument is that we are never directly However for the transcendental realist this aware of any necessary connection between is incidental. For, for him, the status of matters of fact or causal power or agency the concept of necessary connection is clear: so that these concepts cannot be justified it has been established, by philosophical by experience. Thesis (iii) thus completes argument, as applicable to some but not a triangle, whose other sides are theses other sequences of events as a necessary (ii) and (i). It could be argued that we condition of the social activity of science. (It are sometimes directly aware of necessarily should be stressed that this does not mean connected sequences (see 2.3 above), and that any particular science has correctly that we are sometimes directly aware of identified, let alone adequately described, the exercise of causal powers (though the the necessary sequences: it is a condition of powers themselves can only be known, not the possibility of science.) Thus the concept shown, to exist; i.e. we are never directly of natural necessity does not have to be aware of causal powers as such).87 It seems justified in terms of or traced back to its clear that we are aware of ourselves as 290 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY source in sense-experience; though there and I will relate my resolution of the problem must be a scientific explanation of how we of induction to the problem of universals. I come to possess the concept. shall argue that the condition of the intelli- 215:2 That science has a posteriori knowl- gibility of the traditional problem of induc- edge of necessary connections between mat- tion is an ontology of atomistic events and ters of fact is a proposition that can be given closed systems; but that in our world induc- no further justification. tive reasoning may be shown to have a ratio- nal place. 3.6 The Problem of In- 216:1 The traditional problem of induc- tion is the problem of what warrant we duction have for reasoning from particular instances to general statements (induction proper) or 215:3/o In the concluding section of this from observed to unobserved or past to fu- chapter I intend to argue that traditional ap- ture instances (eduction). Now it is clear proaches to the problem of induction fail; to that unless we (sometimes) have some such reveal a crucial ambiguity in the formulation warrant nothing can be justified, shown to of the problem; and to show how transcen- be mistaken or called into doubt: memory dental realism can resolve it. In doing so I cannot be relied upon, a mistake demon- will be bringing together my critiques of the strated or grounds for a sceptical conclusion ideas of the actuality and contingency of law; given. (Why, for instance should the fact 3.6. THE PROBLEM OF INDUCTION 291 that my senses have deceived me in the past ditional problem of induction are of course: be a ground for believing that they will do (i) to deny that science is inductive in na- so in the future?) Indeed complete scep- ture (e.g Popper); (ii) to justify induction ticism about induction seems literally un- inductively (e.g. Black, Braithwaite); (iii) to thinkable.88 So pervasive a feature of our so- strengthen the premises of inductive argu- cial life is inductive-type reasoning that it ments, so that they became in effect en- seems patently unsatisfactory to be told that thymematic deductive arguments (e.g. Mill); it is just a contingent fact about the world (iv) to weaken the conclusions of induc- that induction is successful.89 If inductive- tive arguments to probability judgements type reasoning is necessary, then it seems (e.g. Carnap); (v) to justify induction prag- incumbent upon us to ask what the world matically or vindicate it (e.g. Reichenbach, must be like for it to be possible; and what Salmon); (vi) to dissolve the problem, i.e. to must have been assumed (inter alia about claim that it is a pseudo-one (e.g. Strawson, the world) for the problem to have remained Edwards). The objections to (ii)-(v) are well intractable. The answer to the first ques- known; they all in one way or other beg the tion will constitute a set of synthetic a pri- point at issue which is: ori truths about the world; the answer to 217:1 (A) the problem of what warrant we the second a set of synthetic a priori truths have for supposing that the course of nature about received philosophy of science will not change. 216:2/o The standard responses to the tra- 217:2 (A) must of course be distinguished 292 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY from: statements). However, for a relatively obser- 217:3 (B) the problem of what warrant we vational statement to refute a general law- have for believing some proposition, state- like statement (or theory) it must be pre- ment or theory true. supposed that the course of nature will not change so that the experimental and obser- 217:4 For to say that the special theory vational context in which the refuting obser- of relativity refuted Newtonian mechanics is vation statement is true ceases to be true. not to say that the course of nature changed: Popper would reply to this objection as fol- it is to say that Newtonian mechanics was (at lows: ‘. . . there is a logical asymmetry: least in one respect) wrong all along. I will one singular statement—say about the peri- leave aside considerations pertaining to ques- helion of Mercury—can formally falsify Ke- tion (B) and the rationale for distinguishing pler’s laws; but these cannot be formally it from question (A) until Chapter 4. verified by any number of singular state- 217:5/o Popper claims to have solved the ments. The attempt to minimize this asym- problem of induction, by accepting Hume’s metry can only lead to confusion’.90 But conclusion that induction cannot be justi- the decision to accept the singular statement fied but denying that science is inductive in about the perihelion of Mercury as falsify- nature. According to Popper science pro- ing Kepler’s laws presupposes that in ex- ceeds by the refutation of bold conjectures actly the same circumstances Mercury’s per- (general statements) by their deductive con- ihelion would behave in exactly the same sequences (singular relatively observational 3.6. THE PROBLEM OF INDUCTION 293 way. The asymmetry is there alright. But would be incapable of sustaining scientific or what warrant is there in Popper’s system social life. For the unexpected is a poten- for supposing that nature is uniform so that tially infinite class. No inductive rule could its course will not change, in the way Hume be operationalized in a counter-inductive and Goodman invite us to imagine, so that world. ‘Expect the unexpected’ can only be our best-falsified theories (astrology, Marx- applied ex post, after the unexpected has ac- ism, psycho-analysis, Newtonian mechanics) tually happened. The fact that induction is become true? Whatever the merits of Pop- (sometimes) successful places a constraint on per’s philosophy of science, his claim to have the world in which we live. For in some con- solved the logical problem of induction is ceivable worlds induction would be unsuc- manifestly untenable and based on a confu- cessful or unoperationisable. Hence it would sion of problems (A) and (B) seem reasonable, and indeed necessary, to isolate the conditions that must obtain for 218:1 (vi) also fails, for a number of rea- induction to be successful. If there are neces- sons. First, it seems possible to imagine sary conditions for the success of induction, worlds in which induction would be unsuc- they must constitute the missing ‘justifica- cessful not just ‘counter-inductive’ worlds, tion’. Thirdly, it is clear that not all induc- in which the unexpected always happens, tive arguments are equally good. To appeal but capricious worlds, for which no kind of to induction as an institution does not help rule could be formulated.91 Secondly, even us to decide between good and bad induc- a straightforwardly counter-inductive world 294 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY tive arguments (any more than appeal to the eralization concerned is a law of nature.93 Of law helps us to decide between good and bad course ex ante when inductive reasoning oc- laws). Finally, induction is not always justi- curs we do not know whether the sequence is fied. If it is not always justified there must be necessary. But it is justified if it is necessary. conditions under which it is justified, about And to justify it we need only have grounds which the approach represented by (vi) has for supposing that it is necessary. Now I have nothing to say. already argued that (α), though represent- 218:2 Induction, I have said, is not al- ing an important moment in the process of ways justified. In general induction is only scientific discovery, carries too little ontolog- justified if we have some reason other than ical bite to justify the assumption of a law positive instances for the generalization con- of nature (see paragraph 2 above). The gen- cerned. Two kinds of reasons are distinguish- erative mechanisms of nature are of course able: – nothing other than the powers or ways of acting of things. 218:3/o (Alpha) a plausible model or hy- pothesis of a mechanism by means of which 219:3 Now eduction, or inference from we can render it intelligible to ourselves that particular instances to other particular in- when E then E .92 a b stances, is only justified if in addition the 219:1 (β) knowledge of the mechanism system in which the consequent event oc- which given Ea generates Eb. curs is closed. The crucial ambiguity in the 219:2 Induction is only justified if the gen- formulation of the problem of induction to 3.6. THE PROBLEM OF INDUCTION 295 which I referred earlier now becomes clear. this does not legitimate predictions). It turns on the question of whether the gen- 219:4/o Induction is only justified if the eralization referred to (or in eduction as- generalization is a law of nature and educ- sumed) is an empirical or a normic state- tion is only justified if the system is closed ment, a statement about the conjunctions or (so that the tendency designated in the law events or the tendencies of things, a state- statement must, given the occurrence of the ment about actualities or possibilities. Now antecedent, be realized). Induction is jus- a belief in the uniformity of nature is quite tified because nature is stratified. Now we misplaced if it is a belief in the invariance do not need to know what the structures of patterns of events (or experiences). For are to know that nature is stratified (We do the non-invariance of their patterns is, I have not need to know what the explanation is to shown in Chapter 2, a condition of the possi- know that there is an explanation.) We know bility of science. A belief in the uniformity of that nature is stratified because its strati- nature is only rational if it is a belief in the fication is a condition of the possibility of invariance of structures. The eventsequential science-in-general. And we know that sci- past is an unreliable guide to the future. In- ence is possible in general because it in fact stead what we require, and in small measure occurs. To know that induction is justified actually possess, is a knowledge of the invari- we do not need to know what any particular ant tendencies and natures of things (though explanation is. 220:1 Now if we know what the correct 296 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY explanation is we do not need to reason in- tion, when it is, a rational activity for men. ductively. And when we need to reason in- ductively we do not know what the correct 220:2 I have of course already argued that explanation is. Given then that our knowl- the conditions for inductive scepticism only edge that nature is stratified is secured a pri- obtain if we deny the possibility of (β) above. ori how can we justify any particular piece of If we allow (Beta), i.e. a realist interpretation inductive reasoning? By giving grounds for of the entities postulated in scientific the- supposing that there is an explanation, lo- ory, then we do have a reason independent cated in the nature of things (at the next of the facts identified at any one level of re- highest rung of reality), for the generaliza- ality as to why one but not another sequence tion claimed. It is the possibility of an ad- of events must be forthcoming (if the system equate explanation, in terms of real invari- is closed). Now as the argument for induc- ances from which the behaviour concerned tive scepticism turns on the alleged impossi- can, normically understood, be deduced that bility of knowledge of necessary connections must justify any piece of inductive reasoning between matters of fact and I have demon- in science. Thus it is the possibility of the strated how we can (and do) come to have satisfaction of a deductive criterion that jus- such knowledge a posteriori in science, it re- tifies induction in science. But we know that mains only for me to examine the conditions the world must be such that this is so It is of the plausibility of the traditional problem the structure of the world that makes induc- of induction and to show how the galaxy of problems in its wake can be rationally re- 3.6. THE PROBLEM OF INDUCTION 297 solved. are causally inert: they possess neither lia- bilities nor powers. Now step (β)* involves 220:3/o The problem of induction arises if experimentation. It is condition of the intel- we restrict the grounds for a generalization ligibility of experimental activity that causal to its instances, i.e. if we accept Nicod’s cri- structures not only persist but act indepen- terion of the evidence for a law. It is resolved dently of the patterns of events. Thus the in two steps: (α)* by allowing a model of a world is open, and the laws of nature must generative mechanism or structure to sup- be analysed as the tendencies of things. The ply the missing reason that the coherence of dynamic realist principles of substance and scientific practice demands, and in particu- causality to which I have been working may lar to provide a crucial part of the grounds thus be stated as follows: the world con- for a law; and (β)* by allowing that under sists of enduring and transfactually active certain conditions, i.e. if certain criteria are things (substance) which endure and act in satisfied, such models held out in the scien- their normal way unless acted upon (causal- tific imagination as plausible representations ity). Effects presuppose both continuants of the real mechanisms of nature may come and causes. They must occur in things and to be established as real. Mechanisms are be brought about by things (other than po- enduring; they are nothing but the powers sition in space or moment in time). On the of things. Things, unlike events (which are other hand only effects need explaining. changes in them), persist. Their persistence does not need explaining. Space and time 221:1/o The condition of the intelligibil- 298 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY ity of the problem of induction is an ontol- trast (and unbridgeable gulf) between our ogy of atomistic events and closed systems. particular and our general knowledge of the For without closed systems there is no reason world, transcendental realism allows knowl- for the past to resemble the future and with- edge both of things and of their powers or out atomistic events there is reason why it ways of acting. In place of the analysis of should. The grounds for the atomistic ontol- laws as constant conjunctions of events, tran- ogy of empirical realism disappear when we scendental realism analyses laws in terms of realise that sense-experience is neither the the tendencies of things which may be exer- only ground nor the only source of knowl- cised unrealized and realized unperceived by edge; and that it is analysable neither in men. Science becomes a social activity, dif- purely atomistic terms94 nor as a happening ficult and discriminating; not an automatic, to passive men.95 The grounds for the ac- individualistic affair. Science is explanatory tualist ontology of closed systems disappear non-predictive. when we realize that in general, outside as- tronomical contexts, they need to be exper- 222:1/o I now want to show how the re- imentally established. In place of the ontol- placement of the empirical realist ontology of ogy of experience and atomistic events con- atomistic events and closed systems by the stantly conjoined, transcendental realism es- transcendental realist ontology of persisting tablishes an ontology of complex and active and transfactually active things allows us to structures and things. In place of the con- resolve the problems and paradoxes associ- ated with the problem of induction. This 3.6. THE PROBLEM OF INDUCTION 299 problem, I have argued, only arises if we become green next Christmas (Goodman’s deny the possibility of a reason, located in paradox);96 no reason why the sighting of the enduring nature of some thing, for the a black raven should confirm the proposi- behaviour concerned. In its sharpest form tion that all ravens are black better than it may be expressed as follows: if all predi- the sighting of a red herring or a white shoe cates refer ultimately to experience and ex- (Hempel’s paradox);97 no reason to suppose periences are independent of each other, as that if I had gone for a walk in the rain five they must be if they are to ground (in part or minutes ago I would in fact have got wet (the in whole) our knowledge of the world, then problem of subjunctive conditionals). Now predicates must be independent of one an- there are two ways of meeting these absurdi- other. There can then be no reason for ex- ties. The first is to hold that the paradoxes pecting one rather than another set of ex- and problems stem from the insertion of arti- periences so that for all we know predicates ficial predicates and fanciful conjectures into may become associated in entirely new ways. already functioning and well-connected sci- Thus there is no reason why cabinet minis- entific contexts, for which no positive reason ters should not suddenly start bearing figs can be given. The trouble with this line of re- or Mancunians disintegrate when exposed to sponse is that it is still vulnerable to the ob- the sun (the problem of induction); no rea- jection that there is no ground, independent son why emeralds examined after A.D. 2000 of custom or convention or past practice or should not turn out to be blue or blue things mob psychology, for expecting one sequence 300 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY of events rather than another. And besides though there may be specific problems of the nature of the ‘connection’ predicates are identification and special reasons for expect- supposed to enjoy is unclear. The second ing change. Things persist. They are natu- is the transcendental realist line. This line ral endurers and their changes are explained holds that there are objective connections in in terms of unchanging things. What is the the nature of things, which may be identi- rationale for this resolution? The scientific fied as enduring mechanisms, which bind or explanation of scientifically significant be- link some but not other events and states of haviour is in terms of invariant principles of affairs. I will now sketch the transcendental structure. Thus the scientist assumes that realist resolution of these problems. there is something about metals (their pos- session of free electrons, perhaps) in virtue 223:1 It is physically impossible for a cab- of which it is not possible for them not to inet minister to bear figs; that is, nothing conduct electricity. Their possession of free which bore figs could properly be said to be electrons is the invariant principle of struc- a cabinet minister at all. Desk lamps cannot ture. There is something about cabinet min- fly or walk about the room, just as Man- isters (their genetic constitution, perhaps) in cunians do not disintegrate when exposed virtue of which it is not possible for them to to the sun. A particular must tend to be- bear figs; just as, if Socrates is a man he have in certain ways if it is to be of the must die. kind that it is. On an ontology of things the general problem of induction cannot arise, 223:2/o On an ontology of things Good- 3.6. THE PROBLEM OF INDUCTION 301 man’s paradox cannot arise. Now either ‘all tirely different kind of thing or kind). Thus a emeralds are green’ is law-like or it is not. genuine Goodman-type problem could arise. If it is not the Goodmanesque alternative However it would be a specific problem, it- ‘all emeralds are grue’ is equally admissible. self presupposing the existence of both a con- For it is then ex hypothesi purely accidental tinuant and a cause. Moreover no predicate that all emeralds happen to be green. On such as ‘grue’ could ever be admissible to sci- the other hand, to suppose that ‘all emer- ence. Since the mere passage of time cannot alds are green’ is law-like is to suppose that constitute a cause. It would have to be a co- there is a reason, located in its crystalline incidence that emeralds examined after A.D. structure of chemical composition, why it 2000 looked blue Dates can be at best only differentially reflects light the way it does. proxy causes. Now given that structure, emeralds must, to normal observers under standard conditions, 224:1 Hempel’s paradox may be resolved look green. So anything which looked blue quite simply once the significance of his intu- could not possess that structure, and hence ition, viz. that propositions about shoes and would not be an emerald at all. Now of herrings are irrelevant to the truth of propo- course occasionally we may have grounds for sitions about ravens, is grasped. If laws are supposing that a particular and even a kind statements about things and there must be will cease to exist, i.e. be transformed into a some reason other than instances for accept- different thing or kind (or even into an en- ing them, then Hempel’s paradox may be re- solved as follows: If ‘all ravens are black’ 302 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY is law-like there must be a reason, located all. in the nature of ravens (not in the nature 224:2/o The problem of subjunctive con- of black), why ravens are black. ‘All ravens ditionals is easily and rationally resolved on are black’ is a truth about ravens, not about an ontology of things. To assert a law of colour. Hence the contrapositive ‘all non- nature is to ascribe a possibility to a thing— black-things are non-ravens’ has no bearing a possibility which is possessed by the thing, on it. The logical subject of a law of nature and has a real basis in the enduring nature of is a (natural kind of) thing. Hence there is the thing, whether it is exercised or not. To a logical asymmetry built into its structure, assert a subjunctive conditional is just to say reflecting the site of the mechanism desig- that the possibility possessed by the thing nated, in virtue of which its terms are not would have been exercised, had the condi- equivalent and contraposition is prohibited. tions in fact been different. I would have got To put this another way: the mechanism wet alright, rain being what it is. that, to use Strawson’s term, ‘collects’98 red under herring or white under shoe is either 225:1 The source of these problems lies in entirely different from the mechanism that the reduction of things to qualities and laws, collects black under ravens or else where as which are statements about things, to con- in the shoe case the ‘connection’ is entirely junctions of events. This is reflected most accidental there is no mechanism involved at sharply in the failure to sustain the idea of the necessity of law. But side by side with these well-known problems is a less well- 3.6. THE PROBLEM OF INDUCTION 303 known set (due to the tacit assumption, by of what justifies the assumption that laws almost all philosophers of science, of closed will continue to hold outside the laboratory. systems), which turn on the failure of the ac- This is resolved by allowing (or rather see- tualist ontology of empirical realism to sus- ing that it is a condition of the intelligibil- tain the idea of the universality of law. Lack- ity of experimental science) that things en- ing from the former set is a criterion for dure and continue to act in their normal way distinguishing necessary from accidental se- outside as well as inside the laboratory (as quences (depending upon a concept of the they will do in the future as in the past) stratification of the world); lacking from the unless, as may sometimes happen, they are latter set is a criterion for distinguishing themselves transformed. Corresponding to open from closed systems (depending upon a the problem of subjunctive (and counterfac- concept of the differentiation of the world). tual) conditionals we have the problem of For empirical realism all sequences are acci- normic (and transfactual) ones. Correspond- dental and the world is closed; for transcen- ing to the paradoxes of confirmation, para- dental realism some sequences are necessary doxes of falsification. (Laws and theories are and the world is open. straightaway falsified by any open-systemic instance, just as they are straightaway con- 225:2/o Let us briefly note these homo- firmed by any contrapositive instance, if we logues of the well-known problems on the regard laws as empirical statements.) Corre- universality axis. Corresponding to the sponding to the problem of justifying the use problem of induction we have the problem 304 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY of hypothetical entities in theory construc- tivity presupposes that they act, in space tion (which Hempel has called ‘the theoreti- and time, independently of the patterns of cian’s dilemma’)99 we have the problem of events they generate. Now the use of general justifying the use of the CP clause in theory terms in identifying these objects presup- application (which we could call ‘the engi- poses that they fall into natural kinds. But neer’s dilemma’, or the problem of the ap- it is not possible to say anything in general plied scientist’s excuse). All these problems about the number of kinds there are or about can be rationally resolved by an account of the numbers in any particular kind. Now the science which sees it as an attempt to pen- things posited by science in its investigations etrate ever deeper into the nature of things may be quite recondite and abstract with re- and to describe more adequately the things spect to our ordinary experience. It is wrong of nature. to think of them as necessarily like material objects—they may be powers, forces, fields 226:1 The Humean analysis of laws is a or just complex structures or sets of relation- failure: it does too little and too much. The ships. Their metaphysical character, which causal contingent is neither contingent nor justifies us labelling them as ‘things’ to mark actual, but necessary and real. their insusceptibility to analysis as ‘events’ 226:2 The intelligibility of perception pre- or ‘experiences’, lies in their persistence and supposes that objects persist, in space and transfactual activity. This entails that they time, independently of our perception of persist even when they do not act, and act them. The intelligibility of experimental ac- 3.6. THE PROBLEM OF INDUCTION 305 in their normal way in the flux of condi- kinds of things. This is consistent with the tions that co-determine the actual outcome view of ordinary things as subject to dual of their activity. Things, as so conceived, (and multiple) control, perhaps by princi- must be complex and structured; in virtue of ples of relatively different kinds. Laws do which possibilities may be ascribed to them not describe the patterns of events. Rather, which may be unexercised or exercised unac- we could say, they describe the normic be- tualized or actualized unperceived by men. haviour of novel kinds and impose con- On this account of science the actual is seen straints on familiar things. Ordinary things as an instance of the possible; and a normic may be conceived, metaphysically, as com- mood is added to the hierarchy of condition- pounds. This allows us to make sense of the als marking the space of possibilities exer- individuality of historical particulars; just as cised but unactualized. the conception of ordinary events as ‘con- junctures’ (see 2.6 above) allows us to make 226:3/o On the account of laws advanced sense of the uniqueness of historical events. here they cannot be identified with constant conjunctions of atomistic events or regarded 227:1 If the ordinary things of the world as reporting correlations between either in- are compounds then it is natural that they dependent or equivalent variables. On the should share nothing in common except re- contrary, they must always be grounded in semblances. But just as only some events some conception of an explanatory mecha- are significant in science, although all in nism and ascribed, as tendencies, to specific principle may be explained by it, so with 306 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY things. Ordinary things have a genesis and rational theory-construction in science. To their changes may be rationally explained stress, nothing can be said about the number (in terms of continuants and causes) by ref- or variety of real universals there are. But erence to the exercise of the tendencies of it is clear that ‘table’ and ‘red’ are not real things which share a common identity, i.e. universals; and ‘gene’ and ‘molecule’ are. which fall into a natural kind. Scientifically significant generality does not lie on the face 227:3/o A similar trilemma may be ap- of the world, but in the hidden essences of plied to our explanatory knowledge of the things. world. Either explanation is arbitrary or it is not (arbitrariness is suggested by the prob- 227:2 How can this be shown? Either clas- lem of induction or any of the paradoxes sification is arbitrary or it is not. If it is non- discussed above). If it is non-arbitrary the arbitrary it must be based on a relationship ground for the explanation is either imposed of resemblance (similarity) or identity. If it by men or it exists in the world. If it is im- is only based on the assumption of a rela- posed by men we are left without any ra- tionship of resemblance there is no rationale tionale for experimental activity, the process for the stratification of science. On the other of testing human constructions against the hand if it is based on an assumed theoretical world. Predicates are not independent of identity then we do have a rationale for the each other and classifications are not arbi- move from manifest behaviour to essential trary in science because there are necessary nature that we have seen lies at the heart of connections in the world and things fall into 3.6. THE PROBLEM OF INDUCTION 307 natural kinds. ‘that a flash or a bang occurs does not en- tail that anything flashes or bangs. “Let 228:1 I can now return to the question there be light” does not mean “let some- I asked at the beginning of this chapter. thing shine” ‘.100 But a scientist can never To what, in our ascription of laws, is ne- rest content with effects: he must search cessity and universality properly ascribed? for causes; and causes reside in or consti- The answer is to the transfactual activity tute things. Charged clouds, magnetic fields of things, i.e. to enduring mechanisms at and radio stars can only be detected through work. For these ascriptions to be possible the their effects. But this does not lead us to world must be composed of enduring mech- deny their existence, any more than we can anisms which act independently of men; sci- rationally doubt the existence of society or of ence must be an ongoing social activity; and language as a structure irreducible to its ef- men must be (in the sense indicated in 2.5 fects. There could be a world of electrons above) free. without material objects; and there could 228:2 Now it is because we are mate- be a world of material objects without men. rial things, possessed of the senses of sight It is contingent that we exist (and so know and touch, that we accord priority in verify- this). But given that we do, no other posi- ing existential claims to changes in material tion is rationally defensible. It is the nature things. But scientists posit for these changes of the world that determines which aspects both continuants and causes, some of which of reality can be possible objects of knowl- are necessarily unperceivable. It is true that 308 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY edge for us.101 But it is the historical devel- Maxwell, pp. 482–97. opment of the various sciences that deter- 6 (a) and (b) correspond of course to Hume’s mines in what manner and to what degree two definitions of ‘cause’. See D. Hume, Treatise, these possibilities are taken up by men. p. 172 and Inquiry, pp. 76–7. 7 I owe this term of G. Buchdahl, op. cit., p. 27 and passim. 8 See e.g. R. B. Braithwaite, op. cit., Chap. 8; 3.7 Footnotes to Chap- C. G. Hempel, op. cit., Chap. 12; and E. Nagel, op. cit., Chap. 4. ter 3 9 The Jesuit mathematicism Clavius demon- strated this fallacy in Osiander’s apologetic pref- ace to Copernicus’ De Revolutionibus. Osiander 1 This is the ontological form of Hume’s doc- had argued, as Galileo was later invited to be- trine that events ‘seem conjoined, but never con- fore the Inquisition, that the helio-centric theory nected’. See D. Hume, An Enquiry Concerning was merely a mathematically adequate represen- Human Understanding, p. 74. tation of the facts of planetary motion that made 2 F. Bacon, Novum Organum. no claim to be true. Clavius pointed out that it 3 To borrow Toulmin’s useful concept. See S. was never a good argument in favour of a theory Toulmin, Human Understanding, Vol. I, p. 158 that it ‘saved the appearances’, as a true result and passim. could be derived from any number of absurd or 4 W. E. Johnson, Logic, Vol. III, Chap. 1. false premises. (Cf. J. Losee, An Historical In- 5 R. Chisholm, ‘The Contrary to Fact Con- troduction to the Philosophy of Science, pp. 44– ditional’, Mind 55 (1946), reprinted in Readings 5.) Indeed even if we exclude all premises which in Philosophical Analysis, eds. H. Feigl and G. we know to be false or which are not explicitly 3.7. FOOTNOTES TO CHAPTER 3 309 defined there will still be an infinite number of Solution, Chaps. 4 and 5). sets of premises from which the facts can be de- 11 J S. Mill, A System of Logic, Vol. I, p. 378. duced, provided we allow for the introduction 12 E. Nagel, op. cit., pp. 64–5. of artificial predicates such as Hesse’s ‘tove’ (M. 13 I. Lakatos, op. cit., p. 116. B. Hesse, Models and Analogies in Science, p. 14 N. Goodman, op. cit., pp. 92–122. 30), of which place- and time-dependent predi- 15 For, as Craig’s theorem shows, if it does cates such as Goodman’s ‘grue’ (N. Goodman, not the theoretical component is then completely Fact, Fiction and Forecast, p. 74) merely form eliminable. See W. Craig, ‘The Replacement of a special class. Hence deducibility cannot pro- Auxiliary Expressions’, Philosophical Review 65 vide a sufficient criterion for choosing one set (1956), pp. 35–55. of premises rather than another (the source of 16 N. R. Campbell, The Foundations of Sci- Goodman’s paradox) or for justifying one state- ence, esp. Chap. 6. ment rather than another as law-like. 17 Ibid. pp. 126–40. 18 N. R. Campbell, What is Science?, p. 99. 19 N. R. Campbell, Foundations, pp. 132–7; and M. B. Hesse, op. cit., pp. 35–43. 10 This is of course a very poor best. For 20 N. R. Campbell, op. cit., pp. 243–56. (i) the simplest of any small number of explana- 21 J. Tyndall, ‘Scientific Uses of Imagination’, tions is not necessarily the best (cf. M. Bunge, Fragments of Science for Unscientific People, p. The Myth of Simplicity, pp. 51–134); (ii) there 131. will still be an in principle infinite number of 22 P. K. Feyerabend, ‘Problems of Empiri- equally simple explanations, if we restrict our- cism, Part II’, The Nature and Function of Sci- selves to formal or syntactical criteria alone (cf. entific Theory, ed. R. G. Colodny, p. 317. J. J. Katz, The Problem of Induction and its 23 See e.g. M. B. Hesse, op. cit., pp. 35ff. Cf. 310 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY also P. K. Feyerabend, ‘Problems of Empiricism’, Empiricism, Part II’, pp. 292ff. op. cit., pp. 173ff. 34 P. S. de Laplace, The System of the World, 24 P. Duhem, op. cit., p. 7. Bk. III, Chap. II. 25 M. Scriven, ‘Truisms’, p. 450. 35 F. Bacon, op. cit., Bk. II, Aphorism 26 ‘This’, says Mach, ‘is all that natural laws XXXVI. are’, op. cit., p. 192. 36 ‘Any statement may be held true come 27 Indeed one might be tempted to see the dif- what may, if we make drastic enough changes ference as merely one of taste or temperament elsewhere in the system’, W. V. O. Quine, From as when Duhem compared the ‘rolling drums’, a Logical Point of View, p. 43. ‘pearl beads’ and ‘toothed wheels’ of the me- 37 See e.g. K. R. Popper, Logic of Scientific chanical models of English physicists such as Discovery, p. 42 and pp. 80–2; and T. W. Hutchi- Maxwell, Kelvin and Lodge with his own Carte- son, The Significance and Basic Postulates of sian conception of an axiomatic electricity. See Economic Theory, pp. 40–6. I have of course op. cit., pp. 70–1. argued (in 2.4 above) on quite distinct realist grounds that once the irrationality of pseudo- 28 R Chisholm, op. cit., p. 496. falsification is granted the CP clause becomes 29 See e.g. E. Nagel, op. cit., p. 75; or S. Toul- superfluous. min, op. cit., p. 185. 38 R Harre, op. cit., p. 132. 30 Ibid., p. 185. 39 This theory is most clearly stated in R. 31 R. Harre, op. Cit., esp. Chaps. 2–3. Harre, op. Cit., Chap. 4. Although Harre is I 32 I. Lakatos, op. cit., pp. 134–8. think logically committed to, and may he pre- 33 As Feyerabend has put it: a theory may pared to accept, transcendental realism in the be in trouble only because of ‘the backwardness form in which it is developed here, he does not of the observational ideology’. See ‘Problems of say how laws ‘explain away counter-instances’ 3.7. FOOTNOTES TO CHAPTER 3 311 and so achieve universality, ibid. p. 92. The plausibility of a possible explanation cannot 40 It should be remembered that transcenden- be identified by purely syntactical or formal cri- tal realism not only warrants subjunctive and teria alone but depends upon a complex relation- counterfactual statements (where antecedents ship between what is so far known about the pro- are uninstantiated) but normic and transfactual cess generating the behaviour in question and es- ones (where consequents may be unrealised). tablished explanation patterns drawn from anal- This is another nail in the coffin of deductivism. ogous fields. It is thus in part a function of the For at level (iii) a law may be upheld even when existing knowledge in which the predicates oc- P is true and Q is false; which is of course the curring in the possible explanations are already only case, according to the principle of mate- embedded, so that the paradoxes of confirma- rial implication, when a conditional is false. The tion, etc. that flow from the insertion of artifi- moral is that falsification always depends upon cial predicates into already-functioning and well- the non-formal requirement that the system in connected scientific contexts cannot (at least at which the putative counter-instance occurs be that point of application) arise. Cf. R. Harre, closed. ‘Surrogates for Necessity’, Mind 1973, pp. 355– 41 See e.g. N. R Hanson, ‘A Picture Theory of 80. Theory Meaning’, The Nature and Function of 43 Ibid. p. 366. Scientific Theories, ed. R. G. Colodny, pp. 233– 44 Cf. M. Bunge, op. cit., p. 38. 73. 45 Cf. G. Schlesinger, ‘The Prejudice of 42 I have argued in paragraph 2 above that Micro-Reduction’, B.J.P.S., Vol. 12, pp. 215–24. without such a constraint on the content of pos- 46 Cf. K. Schaffner, ‘Correspondence Rules’, sible explanations, sorting them with respect to Philosophy of Science, Vol. 36 (1969), pp. 280– their plausibility, there will be an infinite number 90. of possible explanations, even of equal simplicity. 47 Given the B is law-like and allowing for 312 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY open systems we must say: x tends to do φ nomic Man, Chap. 7. in virtue of its nature N. A discussion of the 50 See e.g. A. Flew, An Introduction to West- rather complex relationship between tendencies ern Philosophy, p. 49; or E. Nagel, op. cit., p. and powers must be postponed to the appendix 37. to this chapter. For the moment they may be 51 See e.g. L. Kolakowski, Positivist Philoso- regarded as a class of powers whose exercise is phy, p. 34. normically qualified. But this is not a complete 52 Cf. R. Harre, Principles of Scientific Think- analysis. For a power may be exercised when ing, pp. 274–5. the behaviour is not law-like, so that it would 53 Cf. K. R. Popper, Conjectures and Refu- be wrong to attribute a tendency. The logic of tations, Chap. 5. However the protolaw itself power ascriptions, their role in science and the when it finally emerges, pari passu with its ex- ontological status of powers will be discussed be- planation after the limitations and modifications low. necessitated by the experimental process, may 48 The second question is both distinct from be in a form far more complex and refined than the first and important. Because it raises the that in which it was originally conceived (cf. S. question of the range or scope of application of K¨orner, Experience and Theory, passim). The the statements expressing the tendencies of the normal response to a (genuine) counter-instance individuals concerned. It cannot be assumed is modification within a continuous research pro- that all tendencies will be spatio-temporally uni- gramme, rather than (as is implied by naive fal- versal; for individuals and kinds may be trans- sificationism) the complete abandonment of the formed in time and bounded in space. A law may original conjecture and its replacement by a to- of course be universal (transfactually applicable) tally different one (cf. I. Lakatos, op. cit.). within its range and restricted in this way. 54 Cf. R Harre, op. cit., p. 275. 49 Cf. M. Hollis and E. J. Nell, Rational Eco- 55 See A. J. Ayer, The Fundamental Ques- 3.7. FOOTNOTES TO CHAPTER 3 313 tions of Philosophy. jective Knowledge, Chap. 5. 56 See R. Harre and E. H. Madden, ‘Natural 64 See P. K. Feyerabend, ‘Problems of Em- Powers and Powerful Natures’, Philosophy, Vol. piricism’, op. cit., pp. 168–71, and T. S. Kuhn, 45 (1973), esp. pp. 223–30. op. cit., pp. 100–1. 57 See e.g. P. Achinstein, Law and Explana- 65 I have already suggested in paragraph 3 tion, pp. 13ff. above the role that the concept of powers might 58 As the immediate objects of perception are play here. normally assumed to be short events or momen- 66 It is his failure to see this that I think tary states, perceptual access to things presup- leads Feyerabend into error. For he wants to say poses a resolution of the problem of induction (a) that there is bound to come a time when (to be discussed in paragraph 6 below). the ‘alternatives’ do not share a single state- ment (including observation statement) in com- 59 J. R. Ravetz, op. cit., esp. pp. 47–59. mon, yet (b) we could still ‘choose’ between the 60 N. Chomsky, ‘Objectivity and Liberal theories, viz. in terms of the uninterpreted sen- Scholarship’, American Power and the New tences that the scientists testing them would be Mandarins, pp. 23–129. motivated to produce in observational contexts, 61 K. R. Popper, ‘Normal Science and Its op. cit., pp. 214–5. Now what is objectionable Dangers’, Criticism and the Growth of Knowl- about this suggestion is not only that such unin- edge, eds. I. Lakatos and A. Musgrave, pp. 51–8. terpreted sentences could never provide grounds 62 P. K. Feyerabend, ‘Against Method’, Min- for a choice (cf. e.g. D. Shapere. ‘Meaning and nesota Studies in the Philosophy of Science, Vol. Scientific Change’, Mind and Cosmos, ed. R. G. IV, eds. M. Radner and S. Winokur, pp. 17–130. Colodny, p. 61) but also the idea that we could 63 See e.g. P. Duhem, op. cit., Chaps. 9 and ever be in a position to make such a choice. For 10, or K. R. Popper, ‘The Aim of Science’, Ob- this involves the hypostatisation of a whole his- 314 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY torical process of meaning-change and its encap- ‘Are there Necessary Connections in Nature?’ sulation in a single notional moment of judge- Philosophy of Science, Vol. 37 (1969), pp. 385– ment. In this way it involves a relapse back 404. to the pre-relativistic notion that we can make 74 Cf. ‘Tania pushed the door open’ logically judgements outside some particular theory and implies ‘the door opened’. As Davidson has put some particular position in theoretical time. it: ‘the truth of a causal statement depends upon 67 Cf. I. Lakatos, op. cit., pp. 159–65. what events are described; its status as ana- 68 Cf. T. S. Kuhn, op. cit., p. 59. lytic or synthetic depends upon how they are 69 See e.g. M. J. Mulkay, The Social Process described’, op. Cit., p. 90. of Innovation, p. 12. 75 Cf. M. Fisk, op. cit., p. 390. 70 This has been noted by J. R. Ravetz, 76 Cf. R. Harre, Surrogates for Necessity’, p. op. cit., pp. 116–18. But his use of Aristotle’s 380. schema differs substantially from mine. 77 H. Putnarn, ‘The Analytic and the Syn- 71 See K. R. Popper, Objective Knowledge, thetic’, Minnesota Studies in the Philosophy of esp. Chaps. 3 and 4. Science, Vol. III, eds. H. Feigl and G. Maxwell, 72 For the concept of powers continually ex- p. 376. ercised we have of course groomed the concept 78 Cf. E. H. Madden, ‘Hume and the Fiery of tendency. Furnace’, Philosophy of Science 1971, p. 66. 73 See e.g. A. C. Ewing, The Fundamen- 79 It is of course inconceivable that a fun- tal Questions of Philosophy, pp. 159–81 and B. damental entity or entities should act inconsis- Blanshard, Reason and Analysis, Chaps. 11– tently with its (their) nature. Hence in the 12. Cf. also N. Maxwell, ‘Can there be nec- last (non-Laplacean) instance everything is as it essary connections between successive events?’, must be. B.J.P.S. Vol. 19 (1967), pp. 1–25; and M. Fisk, 80 For a discussion of the history of ‘copper’ 3.7. FOOTNOTES TO CHAPTER 3 315 see M. Crosland, Historical Studies in the Lan- ory, Chap. 9, Pt. II, esp. guage of Chemistry and R. Harre and E. H. Mad- 90 K. R. Popper, op. cit., p. 41, n. 8. den, Causal Powers, Chap. 1. 91 M. Hollis, ‘Reason and Reality’, P.A.S. 81 To use Goodman’s very useful term. See 1967–8, esp. pp. 282ff. N. Goodman, The Structure of Appearance, p. 92 See e.g. G Harman, ‘Enumerative Induc- 130 and passim. Goodman himself attributes the tion as inference to the Best Explanations, Jour- term to C. I. Lewis, Mind and the World Order. nal of Philosophy 65 (1968), pp. 529–33. 82 K. R. Popper, The Open Society and Its 93 The stringency of this requirement may Enemies, Vol. I, p. 32. be relaxed in the kind of way indicated in 2.6 83 K. R. Popper, Conjectures and Refuta- above when we move from scientific contexts to tions, Chap. 3, esp. p. 102. the rough-hand-ready generalizations of every- 84 A. Flew, op. cit., p. 450. day life. 85 Ibid., p. 449. 94 See e.g. M. Vernon, The Psychology of Per- 86 S. Toulmin, op. cit., pp. 135–6. ception. 87 E. H. Madden and p. Hare do not clearly 95 See esp. J. J. Gibson, The Senses Consid- distinguish powers from their exercise in their ered as Perceptual Systems. criticism of this Humean argument in ‘The Pow- 96 See N. Goodman, Fast, Fiction and Fore- ers That Be’, Dialogue 1971, pp. 12–31. cast, pp. 73–80. 88 Cf. K. Campbell, ‘One Form of Scepticism 97 See e.g. C. G. Hcmpel, op. cit., Chap. 1. about Induction’, Analysis, Vol. 23, pp. 80–3, 98 P. Strawson, Individuals, pp. 167ff and pas- reprinted in ed. R. Swinbourne, The Justification sim. of Induction, pp. 144–8. 99 C. G. Hempel, op. cit., Chap. 8. 89 Strawson, An Introduction to Logical The- 100 Cf. P. Strawson, op. cit., p. 46 (my em- 316 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY phasis). 3.8 Appendix: Natu- ral Tendencies and Causal Powers

229:1 I have argued that there is a pair of ontological distinctions, presupposed by the intelligibility of experimental activity and scientific training, at any one stratum or 101 Cf: ‘In the Newtonian world and in New- level of reality, between structures and the tonian science . . . the conditions of knowledge do not determine the conditions of being; quite events they generate (and things and the the contrary, it is the structure of reality that changes that occur in them) and the experi- determines which of our facilities of knowledge ences men have of them. These distinctions can possibly (or cannot) make it assessible to may be conveniently expressed by the for- 1 us. Or, to use an old Platonic formula: in the mula Dr ≥ Da ≥ De, where the special Newtonian world and in Newtonian science, it is case Dr = Da = De, assumed to be sponta- not man but God who is the measure of things’, neously satisfied by empirical realism, has in A. Koyre, ‘The Influence of Philosophical Trends on the Formulation of Scientific Theories’, The fact to be worked for in the social activity of Validation of Scientific Theories, ed. P. G. Frank, science. The possibility of Dr 6= Da implies p. 199. that not all events, and that of Da 6= De 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 317 that not all experiences, are equally signifi- tated of course by the fact of open systems. cant epistemically (see 1–6 above). Now the I will call this concept of tendency the pri- postulation of this novel non-Humean on- mary concept. But there is another concept, tology of structures and generative mecha- in principle distinct from this, in which it nisms raises the question of how they ought functions so as to pinpoint the enduring ori- to be analysed. Possibility Da 6= De im- entations, rather than the possibility of the plies that they cannot be reduced to qual- transfactual activity, of things. On this con- ities; Dr 6= Da that they cannot be anal- cept a tendency is something more than a ysed as dispositions. I have suggested that power. It depends upon distinguishing from structures and generative mechanisms must within the class of actions naturally possi- be analysed as the tendencies and powers ble for a thing (which constitute the total- of enduring and transfactually acting things. ity of its powers), in virtue of its being the In this appendix I want to clarify these con- kind of thing that it is, those which are typ- cepts, and to defend the transcendental real- ical, usual or characteristic of that thing as ist ontology from some possible misconcep- distinct from others of its kind. It is the tions. function of this second concept to individ- uate natures within kinds, species r within 229:2/o A ‘tendency’ as I have been using genera, individuals within classes, etc. It is the term up till now is just a power which made possible by the fact that some com- may be exercised unrealized, a power normi- plex structured objects reveal, in virtue of cally qualified. This qualification is necessi- 318 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY their pre-formed structure, what I am going her power). There is a real difference in the to metaphorically characterize as an ‘onto- kind of behaviour, and the state of the thing logical preference’ for some but not other of whose nature is referred to, in the two cases. the natural possibilities open to them. A Men but not dolphins, can (i.e. possess the thing possesses powers in virtue of its falling power to) smoke; but some men are non- into a natural kind tendencies in virtue of its smokers. To attribute a tendency (in the sec- being one of a type within that kind. All men ond sense) is not just to normically qualify (living in certain kinds of societies) possess the exercise of a power; but to say that some the power to steal; kleptomaniacs possess the of the intrinsic enabling conditions of a rela- tendency to do so. tively enduring kind for the power’s exercise are satisfied; that the thing is predisposed or 230:1 The distinctiveness of the two con- oriented towards doing it, that it is in some- cepts of tendency is clear. To say Tania thing of a state or condition to do it. pushed the door open completely explains why the door is open and implies that she 230:2/o In distinguishing in this way be- can do it, i.e. has the power to do it. But tween tendencies and powers we are able to to say that she tends to push the door open avoid the dilemma of supposing either that is to say something more; which cannot be all behaviour is law-like or that some events analysed as when she exercises her power to are uncaused. For a power may be exer- push the door open, it tends to open (which cised and completely explain an event when is just to normically qualify the exercise of the behaviour is not law-like (typical) and so 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 319 cannot be seen as the exercise of a tendency which are possessed by the thing whether or in the second sense (which I will henceforth not they are known by men; so powers can- denote as tendency2). Particular circum- not be reduced to their exercise or our igno- stances may account for the exercise of a rance. power; whereas they must be invoked to ac- 231:2 Now if powers are possessed by count for the non-realization of a tendency , 2 things which act in open systems their exer- if the conditions for its realization are satis- cise must be normically qualified; and they fied. must be seen as tendencies1. This is the concept of tendency I have been using so 231:1 To say that a thing has the power far. In order to apply any tendency or to do 0 is to say that it will do phi, in the 1 normic statement we must know when the appropriate circumstances, in virtue of its antecedent or stimulus conditions for the nature (e.g. its intrinsic structure or genetic mechanism it designates are satisfied. But constitution). I have suggested in paragraph this does not warrant the prediction of the 3 above that in indicating the existence of a tendency’s , fulfilment, i.e. the consequent’s reason, at the next highest level of reality, for 1 realization; which depends upon the system the manifest behaviour a power statement being closed, and in particular upon the non- constitutes a schematic explanation which is intervention of countervailing causes. filled out in the growth of science. To ascribe a power is to make a statement about pos- 231:3 If things are complex and pre- sibilities which may not be actualized and formed they may be in a relatively endur- 320 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY ing state or condition to exercise some but fied and that when the other conditions are not other of their powers. This is the con- satisfied it will do it unless it is prevented. cept of tendency2—the concept of tendency normally employed in ordinary life. When 232:1 Notice that a tendency2 statement such things act in open systems the exercise presupposes a power statement but the con- of tendencies2, will of course also have to be verse is not the case. For a thing must not normically qualified. To say that a thing, X, only possess the power but be in something has a tendency2, to do φ is thus to say : of a condition to exercise it. Now lt might 231:4 (i) X has the power (or liability) to be thought, in view of this, that though do (or suffer) phi; both tendencies, and powers designated pos- sibilities, tendencies designated less remote 231:5 (ii) X is in an enduring condition 2 ones, possibilities closer to or in the process to do phi, i.e. it is predisposed or oriented of realization. But this metaphor must be towards doing phi; interpreted very carefully. For the greater 231:6 (iii) X will do phi, given an appro- ‘proximity’ of the tendency’s realization is priate set of circumstances, in virtue of its still a purely non-epistemic one. For I may predisposition, in the absence of intervening be absolutely certain that the stimulus or (or countervailing) causes. other conditions will be unsatisfied or that 231:7/o That is, it is to say that X can do countervailing causes will intervene so that it, that most (or the most important) of the the tendency2 will be unexercised or if exer- intrinsic enabling conditions for it are satis- cised unrealized respectively. Knowledge of 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 321 a tendency, is quite distinct from judgement tributing it must be based entirely on our about the tendency’s2, realization. theoretical knowledge of the thing. In gen- eral, the attribution of tendencies , requires 232:2 It is a mistake to think of offset- 2 more about things to be known than is the ting causes as relatively short-run retarding case in the attribution of powers. For we barriers which are sooner or later overcome must know that the enabling conditions for by the superior staying power of the ten- the power’s exercise are satisfied and hence dency. This is to tacitly conflate epistemic what they are. and natural possibilities. Equilibria may last for ever. A related mistake is to iden- 232:4/o Tendency ascriptions may be tify the ‘lawfulness’ implicit in a tendency , 2 2 merely generic. There may be no unique statement with the regularity or repeatabil- set of conditions under which a tendency ity of a syndrome of behaviour. For some 2 is exercised; and the way in which it is re- tendencies , (to explode to commit suicide) 2 alised, if it is, may depend upon the partic- have of necessity only a single manifestation ular circumstances in which its exercise oc- and cannot be fulfilled twice. curs. A kleptomaniac’s tendency is not nor- 232:3 A tendency2 may be revealed in ac- mally fixed on a particular shop and it is tion, even when it is unexercised; in which not normally knowingly exercised under the case our grounds for attributing it must be eye of the law. (If it is, he ceases to be a indirect. If it is not shown at all, i.e. has kleptomaniac and becomes something else.) no overt expression then our grounds for at- Similarly the way in which Fido expresses his 322 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY

generic tendency2 to bark may depend upon tendency2 were unrealized, set us off in the the particular factor that excites it. He may wrong direction, on a search for some ex- bark viciously at an intruder but convention- ceptionless generalization which covered that ally at the postman, fearfully at the moon case too; so providing an empty and ever- but affectionately at another dog, arrogantly expanding redescription rather than an ex- at a cat and playfully at an old shoe. planation. Generality in nature lies in things not conditions. Explanation in terms of 233:1 Both (ii) and (iii) play an impor- the powers/tendencies conceptual network tant role in indicating a characteristic ex- avoids the circularity and triviality of expla- planatory problem and the appropriate di- nations conforming to the Popper-Hempel rection of research. Thus the problem in- model, where an event cannot be explained dicated in (iii) is not: what accounts for other than by or without subsumption under the fulfilment of tendencies, given that the generalizations. The former is always insuf- conditions for their fulfilment are realized? ficient and the latter mostly impossible. So But rather, what accounts for their nonful- room must be found for the concept of a gen- filment in these circumstances? (The for- erative mechanism, activated under closed mer needs no special explanation.) And it conditions, explaining regularities when they tells us what to look for: namely, the pres- occur; and the concept of an agent bringing ence of interfering causes. It should be noted about and so explaining an event in a given that a formulation which only included ref- situation without this being interpreted as erence to the circumstances would, if the 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 323 involving a claim to a regularity. appropriate circumstances if, at the time, he possesses in addition a set of overriding or 233:2/o Offsetting causes are often as- more compelling reasons for not acting in sumed to be always extrinsic. But the cause that way. The idea of a perfect balance of of a failure of a car to move when the gear opposed causes is expressed in the theoreti- is m neutral is not something distinct from cal concept of an equilibrium. . and extraneous to the mechanism responsi- ble for its normal motion. Science is never 234:1 It is the specific role of (ii), I sug- concerned with just listing causes but seeks gest, to indicate the existence of a level of always to relate them, even when seemingly activity within the thing which is ensuring opposed or contradictory, to common struc- or has ensured the satisfaction of the intrin- tures. Now intrinsic offsetting causes may sic enabling conditions for phi. That is, it or may not directly interfere with the oper- directs our attention to the mechanisms re- ation of the mechanism responsible for the sponsible for those relatively enduring states satisfaction of the intrinsic enabling condi- of a thing that distinguish the possession tions. If they do then we must say that the of a tendency2 from that of a mere power. tendency2 is no longer possessed, as condi- The tendency2 statement is action-oriented tion (ii) ceases to be satisfied. But not all by contrast with the simple power statement intrinsic offsetting causes are like that. Thus in two respects then: firstly, in indicating the a person may possess a reason for acting in past or present activity of generative mech- a certain way and not act in that way under anisms responsible for the persisting orien- 324 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY tations of things- and secondly, in indicating of social policy. But this merely underscores the possibility of their transfactual activity, the latter’s distinctiveness from science. For i.e. the exercise of these tendencies irrespec- science is concerned with the behaviour of tive of their fulfilment. A simple power state- things only in as much as it casts light on ment is, on the other hand consistent with their reasons for acting and hence upon what completely quiescent or dormant things or they are. things which have a level of activity suffi- cient only for the . retention of that power. 234:3/o The underlying action indicated in the tendency2 statement may be continu- 234:2 Now in an important sense it is not ous or sporadic; and the onset of a tendency2 whether the tendency2 is realized or not that may be gradual or sudden. Moreover some constitutes the scientifically interesting thing of the mechanisms responsible for the pos- about it (for this will always be to a greater session of tendencies2, e.g. those posited in or lesser extent circumstantially determined) Freudian theory, may have ceased to oper- but the reason for the relatively enduring ori- ate long before the time of their manifes- entation that differentiates the possession of tation. The grounds for our attribution of a tendency, from that of a power. Thus in the tendency must then of course lie en- the case of the kleptomaniacs, the scientist tirely in our evidence for the pre-formed in- will be concerned with the reasons for his trinsic structure rather than in our evidence tendency2, whether he suppresses it or not. for the present higher-order activity of the Of course this need not be the chief concern thing. Such tendencies2 are latent It should 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 325 be stressed that we cannot assume that the things. But within the category of extrinsic underlying action occurs in systems which causes we must remember to include pulls are themselves anything other than open; as well as pushes, structural relationships as just as we cannot assume that underlying well as the momenta of other things. the action of those mechanisms are anything other than more basic ones. 235:2 Both components (ii) and (iii) of the analysis of tendencies2 encourage a further 235:1 Some tendencies2 are powers which investigation of powers: viz. into those re- are held in abeyance; and are straightaway sponsible for the persisting orientations of exercised when the impediments to their ex- things and into those whose exercise is re- ercise are removed. Other tendencies2 are sponsible for the non-realization of known powers which require for their exercise the tendencies2, when the conditions for their active stimulus of other things. One might in realization are satisfied. To sum up then a the former case talk of ‘releasing rather than tendency2 statement says there is a level of ‘stimulus’ conditions. Releasing or stimulus activity, perhaps unknown, intrinsic to the conditions may be intrinsic or extrinsic to thing, such that it is predisposed to perform the thing possessing the tendency2. Spon- an action of a certain type. Its chief func- taneity, self-determination and acting for a tion is to indicate the existence of a level of reason all fall under the former paradigm. It activity within the thing such that it is ori- is better to distinguish between powers and ented towards some rather than other of the liabilities than between active and passive natural possibilities open to it. In this way 326 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY it leads us to a more precise specification of of scientific practice requires that we view the natures of particular things (or groups) them theoretically as at any moment of time within kinds. unaffected by the closure or otherwise of the systems in which they occur. Clearly this 235:3/o I now want to turn to some pos- does not mean that natures and generative sible misconceptions arising out of my use mechanisms must be viewed as eternal or of the primary concept of tendency (which as completely unaffected by the actions of I shall not denote by a subscript) and the things or their operations. Powers may wane concept of power. The intelligibility of ex- and disappear. Mechanisms may be trans perimental activity and the possibility of the formed. Such transformations must them- practical application of our knowledge de- selves be analysed as events in open systems; pends, I have argued in Chapter 2, upon the and are in principle explicable in terms of a assumption of the ontological independence network of normic statements. The chang- of causal laws (which I have analysed as the ing is explained in terms of the unchanging, tendencies of things) from the patterns of but the unchanging need only be relatively events. That science can job back and forth so. Some fields may be irreducibly histori- from open to closed systems depends entirely cal. But the very move, which I have argued upon .the consideration that the natures of is essential to our understanding of science, things and the generative mechanisms of na- from talking of the occurrence of events to ture persist and act in open and closed sys- the things and structures that account for tems alike. That is to say, the intelligibility 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 327 them presupposes that they are relatively dental realist are conditions for the experi- persistent. A non-enduring thing would be mental testing, not the applicability of a law. a mere event, a totally affected mechanism To say that laws are applicable if and only if incapable of production. their ‘boundary’ as well as their initial condi- tions are satisfied is useless because it is men 236:1/o Now it is vital to distinguish the that ensure the satisfaction of the boundary initial conditions whose instantiation war- conditions. And this activity is only intelli- rants the application of a normic or tendency gible if laws continue to apply in open sys- statement, i.e. of a causal law, from the tems, where the so-called ‘boundary condi- complete atomistic state-description that the tions’ are not satisfied. We isolate systems empirical realist needs to sustain his analy- to understand mechanisms; but the mecha- sis of laws. For on the transcendental realist nisms endure and act quite independently of view a law is applicable if the mechanism it our activity. designates is set in motion, i e. if the condi- tions explicitly mentioned in the protatis of 237:1 The ordinary things or standard the law-like statement are satisfied, whether particulars of the world are metaphysically or not the system is closed i.e. whether or compounds, just as ordinary events are con- not what are sometimes referred to as the junctures. The philosophy of science has boundary conditions’ are satisfied, and hence noted, quite correctly, that the objects of sci- whether or not the consequent is realized. entific thought are ‘ideal’ or abstract with The ‘boundary conditions’ for the transcen- respect to such things and events. But the 328 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY transcendental realist sees such objects as bridging role. For in hinting at the perhaps real. For him the world is composed of real to be discovered reason [at Stratum II] for things and generated by real mechanisms. It behaviour [at Stratum I] it represents an im- is the world itself, not our thought of it, that portant moment in the process of scientific is abstract and ideal. The world consists of discovery. Its structure reflects this, in in- chemical elements, electrons, quarks; its phe- corporating both a characterization at the nomena are generated by actions, reactions, science s point of departure and a-reference forces, fields. The received tradition in the to its intended destination. To say that X philosophy of science has seen such things has the power to do φ is to say that it will do as inventions of men designed to gloss, and φ in the appropriate circumstances in virtue account for deviations in, our familiar expe- of its nature (e.g. structure or constitution);2 riences. But transcendental realism, starting that is to say it will do it in virtue of its being from the premise of the contingency of our the kind of thing that it is. own experience, sees nature as real; and sci- ence as our persistent effort to understand 237:3/o Now to ascribe a power is to sup- it. pose that there is a real basis for the posses- sion of that power independent of whether 237:2 In paragraph 3 I argued that the the power is exercised or not.3 But it is not concept of powers, in combining a specific possible to say whether that real basis can attribution of behaviour with an unspecific be qualitatively described or not. Some real attribution of structure, plays a key strata- bases, such as the molecular structure ac- 3.8. APPENDIX: NATURAL TENDENCIES AND CAUSAL POWERS 329 counting for the fragility of glass, can be already argued that if science is to be possi- qualitatively described; others, such as the ble there must be necessary connections be- structure of a gravitational or magnetic field, tween some but not other events; and that can not. Hence it is contingent whether such necessary connections are provided by we can know about a thin anything other the enduring mechanisms that bind some but than its causal powers; and it is contingent, not other events together and that exist as when we can not, whether a thing is any- the powers of things. Now the knowledge of thing other than its causal powers. Now the chemical nature of opium which allows us Mackie has objected to the concept of pow- to deduce its tendency to produce soporific ers on the ground that if it is to form the ba- effects in men is not a priori; but discovered sis for a new metaphysics of science it must a posteriori, in the ongoing process of sci- violate ‘Hume’s principle that there can be ence. Once we know a thing’s nature then we no logical connections between distinct ex- can deduce its tendencies, how it will behave istences’.4 In support of this he asks ‘what CP. But typically in science when we ascribe would be the point of showing that opium powers we do not know a thing s nature but contains morphine which (despite its name) only believe that it is in virtue of the thing’s is only contingently related to sleep, if we nature, whatever it is, that it does what it knew already that opium contained an in- does (i e that its behaviour is necessary, not trinsic power whose presence entailed the accidental). A power-statement then acts as production of sleep?’5 Now I have of course an imperative for scientists to find, and as a 330 CHAPTER 3. THE LOGIC OF SCIENTIFIC DISCOVERY temporary place-holder for, that explanation Belief, Truth and Knowledge, esp. pp. 11–16. which, by capturing the essence of the thing, 3 Cf. D. Armstrong, A Materialist Theory of will allow the most stringent possible crite- Mind, p. 86. rion for our knowledge of natural necessity 4 J. L. Mackie, Truth, Probability and Para- dox, p. 137. to be satisfied. 5 Ibid., pp. 143–4. 238:1 When we know what a thing is we know what it will tend to do, if appropriate circumstances materialize.

3.9 Footnotes to Ap- pendix of Chapter 3

1 Dr = domain of the real; Da = domain of the actual; and De = domain of the empirical in the sense specified in 1.6 above. 2 Cf. R. Harre, Powers’, B.J.P.S. Vol. 21 (1970), p. 85 and Principles of Scientific Think- ing, p. 270. See also D. Armstrong, Dispositional Properties’, Analysis 22 (1961–2), pp. 44–46 and Chapter 4

Metaphysics and the Philosophy of Science

239:1 If science is to be possible men must of states and events in the world in the sense possess certain essential powers. Among of bringing about effects which but for their these is the power of affecting the sequences action would not have been realized. In this

331 332 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE way men contribute to the universal mael- quire and use language. The latter implies strom of existence. More specific to men is at the very least that we are material things their power to initiate and prevent change with a point of view in space and an exis- in a purposeful way. The possession of this tence in time, so that we must be able to power seems to stem from the fact that men communicate with each other from differ- are material things with a particular degree ent spatio-temporal locations. And the for- of neuro-physiological complexity which en- mer implies at the very least that we must ables them to monitor and control their own be able to communicate with each other on actions. Foremost among the powers nec- the basis of the possession of differing cogni- essary for science and, as far as we know, tive equipment, so that reindividuation and distinctive of men is their power of inten- recharacterization must be possible and the tional action, which enables them to act self- definitions of terms must be at least partially consciously on the world: that is not just to open. However language is by no means monitor and control their performance, but the only vehicle of thought; nor is language- to monitor the monitoring of their perfor- using our only intellectual skill. Pictures, di- mance: to plan, to act and so to make an agrams and iconic models play, it has been anticipatory commentary come true. argued, an indispensable role in scientific thought. 239:2 Among the powers falling within this genus perhaps the most basic and cer- 239:3/o Now endowed with our ensemble tainly the most studied is the power to ac- of intellectual powers we are able not just to 333 describe the patterns of events but by imag- ble objects of self- criticism; and inter alia ining structures (which may come to be es- through our capacity to make public our tablished as real) to grasp the mechanisms monitored commentaries upon these moni- of their production. tored performances open to the criticism of others. Simple rudimentary perception, such 240:1 Closely related to these powers is as sniffing, and basic actions, such as rais- our capacity to design, manufacture and use ing one’s arm, which depend in general only tools. This inter alia enables us to act at a upon a first-order monitoring, are in them- distance; and in the course of science it has selves relatively insignificant in science. The enormously extended our powers of percep- chemist who sneezes while conducting an ex- tion and detection. More generally, it has periment is aware of himself in a certain way, enormously increased our powers of inter- but not in the way that the physiologist un- vention in and control over, as we say, ‘the derstands all sneezes, including his own. For course of nature’. the latter knows the mechanism of sneezing, 240:2 In all these activities—involving while the former merely experiences its ef- perceiving, manipulating, discoursive and fects. (The chemist, unlike the layman, may pictorial thought—the role of the nature of however be aware of the cause, in the sense men as mechanisms capable of second-order of antecedent or stimulus conditions for the monitoring and feed- back is vital. In virtue mechanism, in this case). of it our diverse perceptual, manipulative and cognitive performances become possi- 240:3/o Scientific perception is non-simple 334 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE then in the sense that it must normally be causal laws. Its significance - the significance theoretically-informed. And most scientific of experimentation—lies in the consideration actions are non-basic (in Danto’s sense)1 in that it gives us access to causal structures that they consist in doing things to bring that exist and act independently of the ex- other things about (i.e. phiing to ψ),2 that is periment. In the same way the significance initiating sequences of events whose outcome of perception lies in the consideration that it is both distant and planned. It is through gives us access to things that exist indepen- our acquired skills of perception that we dently of it. And in a sense scientifically- come to be in a position to formulate propo- relevant perception depends upon an an- sitions concerning the behaviour of things, to terior disruption too—of the sequence of identify and describe the flux of events. But common-sense experiences in a scientific ed- it is through our manipulative powers that, ucation or training. by interfering with the course of nature (this 241:2 Now the crux of my objection to the flow of events), we are able to check the re- doctrine of empirical realism should be clear. ality and study the operation of the hypo- It is that it cannot sustain the intelligibility thetical generative mechanisms that in the of perception and experimental activity; and scientific imagination we picture as respon- that in positing a correspondence (or even in sible for their behaviour. its positivist form an identity) between: 241:1 In such a move we disrupt the se- 241:3 (a) sense-experiences, which are in quences of events to identify the underlying general only made significant by the trans- 335 formation of antecedent common sense, and thing or phenomenon (or explaining the their objects, viz. events and the states of same event) in a different way, as well as from things, as expressed in the concept of the different spatio- temporal locations. (b) de- empirical world; and pends upon the intelligibility of the idea of 241:4 (b) constant conjunctions, which are causal structures existing and acting where in general only made possible by human ac- no empirical regularities prevail. Both de- tivity, and causal laws, as expressed in the pend upon the non-identity and possible dis- idea of the actuality of causal laws juncture of the terms of the pairs; that is, upon the possibility of one varying with- 241:5 it makes impossible both scientific out the other. Thus we must have a con- change (at least in our descriptions of possi- cept of facts as social products and what is ble objects of experience) and the scientific described and/or explained as independent explanation of things existing and acting in of men. And we must have a concept of open systems (that is, in systems where in- the mechanisms that generate phenomena ir- variant conjunctions of events have not been reducible to the phenomena they generate. found or made to prevail). In this way em- And so two dimensions must be established pirical realism comes to seriously understate in the philosophy of science: a transitive di- the critical significance and scope of applica- mension, in which experiences and conjunc- tion of science. tions of events are seen as socially produced; 241:6/o (a) depends upon the possibil- and an intransitive dimension, in which the ity of identifying or characterizing the same 336 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE objects of scientific thought are seen as gen- together with the celestial closure and the erative mechanisms and structures which ex- classical paradigm of action, the analogical ist and act independently of men. grammar’3 of empirical realism, the scien- tific substance that lends plausibility to its 242:1 Now empirical realism is character- metaphysics, that gives credence to its philo- ized by the absence of an intransitive dimen- sophical form. sion and of a transitive dimension with re- spect to experience. But I have already ar- 242:2 If there is a basic or fundamental gued that this merely results in the genera- level of knowledge, at which what is ex- tion of an implicit ontology, based on the cat- pressed is certain and given independently egory of experience and an implicit sociology, of any human activity and reflects (and ulti- based on the category of men: that is, in the mately constitutes) the world, then its con- generation of an ontological atomism and an stituents must be atomistic. For if they epistemological individualism (at least with were not atomistic they would be them- respect to experience). I have further sug- selves susceptible of analysis, and so require gested in 3.4 above that underpinning em- justification. Hence it is the characteris- pirical realism is a particular conception of tics of the concept of knowledge to which men (at least with respect to experience): empirical realists have subscribed, that ac- in which men are seen as passive sensors of counts for the disastrous ontology that we given facts and recorders of their given con- have examined. It is the desire for incorrigi- junctions. This model of man constitutes, ble foundations in a level which constitutes 337 the world that generates the ontology of is determined (in part or in whole) by expe- atomistic events and closed systems respon- rience or (ii) it is fixed a priori as a necessary sible for the problems and errors analysed in condition of experience or (iii) it is free (in Chapters 2 and 3. Against this I have argued the sense of unconstrained by experience).4 that knowledge depends upon knowledge- I said at the outset that part of my inten- like antecedents and that the world is inde- tion was a philosophy for science. What, pendent of men. But if it is the requirements then, are the implications of these three po- of an incorrigible ground for knowledge of sitions for scientific practice? It is clear that the world in empirical realism that generates if knowledge is regarded as justified in terms the implicit ontology of empirical realism, it of given experience we have the makings of is the model of man necessary to sustain the what is in effect a conservative ideology, in incorrigibility of this ground that forms the which the current experiences of a science are lynchpin of the tradition; and hence explains rationalized in being thought of as natural, in the last instance the doctrines and prob- given or implied by the very nature of things lems we have been criticizing and examin- themselves. If (ii) is seriously considered it ing This is of course as one would expect would tend to have the same conservative ef- given the epistemological bias and anthro- fect. It is less obvious that (iii) would too— pocentricity of our philosophical thought. until we remember that science is an insti- tutionalized activity in which if to be free 243:1 Empirical realism is faced with the is to be unconstrained by the possibility of following trilemma: either (i) our knowledge 338 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE critical experience a self-perpetuating dogma of their own underdevelopment and partly may ensue. Examples could be drawn from by the consideration that science in general the history of physics and chemistry and eco- and social science in particular is also part nomics to show how empirical realism has of their own field of inquiry. The resonance functioned in this way as a conservative ide- between Skinner and Hume or Marshall and ology for science. Mill is too clear to need further remarking. 244:1 In addition to the effect of its con- 243:2/o A philosophical system may serve cept of knowledge and its own substantive to rationalize the practice of a science in an- scientific analogies a philosophy of science other way, viz. through its own substantive may of course also, and most obviously, ra- scientific analogies and the correspondence tionalize the practice of a science through or resonance they find in the science. Thus its own conception of the methods of science empirical realism readily finds an echo in and the appropriate objects of scientific in- mechanistic explanations in physics and psy- quiry. If I am right in my argument the chology and reductionist programmes in so- methodology of empirical realism is not that ciology and biology. One would expect such of science but if it were to be acted upon it a link to be particularly strong in the case would have the most deleterious effects on of the social sciences, if only because of the its practice. greater sensitivity on the part of social sci- entists to the philosophy of science—a sensi- 244:2 Now if empirical realism depends tivity partly explained by the apprehension heavily for its plausibility upon its analogical 339 grammar, it could be asked upon what sub- the social sciences and at the very least de- stantive scientific analogy does transcenden- void of the same significance in psychology tal realism depend? The answer is, I think, as it possesses in physics and chemistry, I none. All philosophical argument that is not want to round off my argument by consider- explicitly transcendental depends, I think, ing whether something analogous to the con- upon more or less strained scientific analo- trolled investigation of nature, making pos- gies or tacitly presupposed substantive the- sible the experimental confirmation and fal- ories. But a transcendental argument whose sification of theories, might be possible in premise is explicitly stated need not depend the social sciences and psychology and other upon any particular theory, other than those fields where experimental activity is impos- bound up in the activity that is its object sible or more or less seriously circumscribed. and which is its task to explicate. 245:1 Three points are clear. First, that 244:3 The central argument of this study, there is a general problem of confirmation (or establishing an ontological distinction be- corroboration) and falsification in the non- tween causal laws and patterns of events (the experimental sciences. Second, that though independence of the domains of the real and we can assume that there are explanations the actual, the irreducibility of structures for social and psychological phenomena (and to events), has turned on the possibility of under social and psychological descriptions), experimental activity. Now as it is clear we cannot assume that the social or psy- that experimental activity is impossible in chological sciences have got anywhere near 340 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE them. Thirdly, that any adequate solution to tual performance, but such a study is not the methodological problems (and in partic- likely to proceed vary far unless the sepa- ular problems of confirmation and falsifica- rate factors are themselves reasonably well tion) of the non-experimental sciences must understood.5 depend upon a more adequate conception of natural science than that which has so far informed discussion of them. 245:4 Now if I am right in arguing that the significance of experimental activity in nat- 245:2 An awareness of the general prob- ural science is that it gives us access to en- lem of confirmation and falsification in the during and transfactually active structures non-experimental sciences is shown by the and that it is only under closed conditions following quotation: that confirmation and falsification of theory 245:3 The linguist . . . is studying one fun- is possible, then we are in a better position damental factor that is involved in perfor- to see that the central problem of the psy- mance, but not the only one. This idealiza- chological and social (and other non- experi- tion must be kept in mind when considering mental) sciences is that of devising (or recon- the problem of confirmation of grammars on structing) an analogous procedure of inquiry the basis of empirical evidence. There is no and selectively empirical confirmation (and reason why one should not also study the in- falsification) and to appreciate the great gulf teraction of several factors that are involved that must separate them, in the absence of in complex mental acts and that underlie ac- such a procedure, from the sciences of na- 341 ture. a certain domain of phenomena. I suggest that they are falsified by their incapacity 245:5/o In the case of psychology the to explain in a non-ad-hoc way a range of agent’s capacity to give a commentary phenomena that takes on a special signifi- on intentional action might provide an cance for the agents that participate in the experiment- analogue, though one cannot forms of social life they define. Thus it was rule out the possibility that new concepts the mass unemployment of the 1930’s that will have to be given by the investigator to demonstrated the inadequacy of the neo- the agent (even e.g. in the case of the identifi- classical system and provided the motor for cation of emotions). Much social science can the Keynesian innovation which showed how be seen to depend upon attempted real def- an unemployment equilibrium was possible. inition of forms of social life which have al- Clearly I have no space to defend or elabo- ready been identified under certain descrip- rate these suggestions here. It is sufficient tions and are known by the agents who par- for our purposes merely to note the prob- ticipate in the social activities concerned un- lem: what are the enduring and transfactu- der those or other descriptions. Social struc- ally active ‘mechanisms’ of the sciences of so- tures, unlike natural structures, cannot ex- ciety and man? Transcendental realism con- ist independently of their effects. Thus real ceives the various sciences as unified in their definitions of concepts such as capitalism, method but specific to (or differentiated with democracy, power, love can only be Justi- fied by their capacity to render intelligible 342 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE respect to) their particular objects. that is necessary for the generation of event- 246:1 Now once generative mechanisms invariances under significant descriptions, are seen to be the objects of scientific that is of the activity involved in (ii)-(iv). thought, it can be seen that four questions Thus the stimulus and enabling conditions can be asked of any generative mechanism for the operation of the mechanism must be G: – satisfied; the mechanism must be isolated 246:2 (i) is the mechanism enduring? and the flux of conditions held constant or otherwise controlled; and skilled observers 246:3 (ii) is the mechanism operating? must be present to perceive or detect it. And 246:4 (iii) are the results of the activity of secondly, neglect of the possibility of non- the mechanism unaffected by the operations enduring mechanisms, of laws which though of others (of either the same or different universal and normic in form (i.e. transfac- 246:5 (iv) are the results of the activity tually applicable) are themselves bounded in of the mechanism perceived or otherwise de- space and restricted in time. On the tran- tected by men? scendental realist ontology, the description 246:6/o Now if one assumes, as the ac- of what the world must be like if science is tualist does, that laws are empirical invari- to be possible, the classical principle of indif- ances (universal empirical generalizations) ference (or invariance) applies only to struc- one cannot sensibly ask these questions. tures, not to events. But structures may Two consequences should be noted. First, themselves be transformed; and so concepts a neglect of the conscious human activity 343 of diversity and change, like that of struc- sions apodeictic. I have shown the struc- ture, squeezed out by the implicit undiffer- tured and intransitive character of the ob- entiated ontology of empirical realism, may jects of scientific enquiry to be a condition come to occupy as significant a place in on- of the intelligibility of experimental activ- tology as the concept of indifference. ity and the social nature of knowledge to be a condition of the intelligibility of scientific 247:1/o I have stressed throughout this training. (The laboratory and the classroom study the diversity of phenomena and the are the two most under analysed, and yet autonomy of the various sciences (both from the two most obvious, sites of science). I one another and with respect to common have isolated the conditions of the plausibil- sense), the opacity of their concepts and the ity of the doctrine of empirical realism and strangeness of the objects with which they shown then to be very special. A conception have to deal. Natural processes are indepen- of both nature and our knowledge of nature dent of man but man is himself in, and in as differentiated and stratified has been ad- continual interaction with, nature. Nature vanced. And I have developed two criteria itself is diversified and complex. But I doubt for the adequacy of any account of science: whether our concept of nature can be under- viz. its capacity to sustain the possibility of a stood in isolation from our concept of man; world without men and the impossibility of and I doubt whether we understand the lat- knowledge without antecedents. These es- ter. The scope of this enquiry has how ever tablish the necessity for both of what I have been strictly philosophical; and its conclu- 344 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE called the intransitive and transitive dimen- any one level of reality to the construction sions of the philosophy of science: any ac- and testing of possible explanations and thus count of science that does not view knowl- the discovery of the mechanisms responsi- edge as socially produced and the objects of ble for them. This process necessitates the knowledge as independent of men must be construction of both new concepts and new ruled out as a possible account of science. tools (or the resurrection or refinement of old ones). The aim of science is the discovery 248:1 If science is to be possible the world of the mechanisms of the production of phe- must be one of enduring and transfactually nomena in nature; and it proceeds by way of active mechanisms; and society must be a a dialectic of taxonomic (or descriptive) and structure (or ensemble of powers) irreducible explanatory knowledge, in which the con- to but present only in the intentional ac- flicting principles of empiricism and rational- tion of men. Science must be conceived as ism can be reconciled, a dialectic which has an ongoing social activity; and knowledge no foreseeable end. as a social product which individuals must reproduce or transform, and which individ- 248:2/o In order to render intelligible sci- uals must draw upon to use in their own entific change and to reconcile it with the critical explorations of nature. Science is a idea of scientific progress we must have the process in motion, continually on the move concept of an ontological realm, of objects from manifest behaviour to essential nature, apart from our descriptions of them. We can from the description of things identified at then allow, for example, that theory Ta is 345

preferable to theory Tb, even if in the termi- ence we must always speak of them and know nology of Kuhn and Feyerabend it is ‘incom- them under particular descriptions, descrip- mensurable’ with it if theory Ta can explain tions which will always be to a greater or under its descriptions almost all the phe- lesser extent theoretically determined, which nomena p1 . . . pn that Tb can explain under are not neutral reflections of a given world. its descriptions Bp1 ...Bpn plus some signifi- Epistemological relativism, in this sense, is cant phenomena that Tb cannot explain. We the handmaiden of ontological realism and can speak in this way in the meta-language must be accepted. Now this does not mean of philosophy, and we must speak so if we that it is impossible to communicate between are to retain the idea of scientific progress different theoretical or conceptual schemes without falling back on the idea of certain or that a scientist cannot know the same ob- foundations of knowledge or theory-free ex- ject under two or more different descriptions. perience. It is the intuition of this neces- To show the difference between say Newto- sity that accounts, I think, for the readiness nian and Einsteinian dynamics and that the with which some philosophers of science have latter is an advance on the former a scien- embraced Tarski’s theory of truth.6 But this tist must be capable of doing so. Similarly theory cannot help us to resolve the problem though there is no guarantee of successful posed by the apprehension of the general rel- communication between the adherents of two ativity of our knowledge: viz. that whenever different conceptual schemas, there is no in- we speak of things or of events etc. in sci- evitability about failure (It is difficult to un- 346 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE derstand the concept of total failure.) Epis- thought). temological relativism insists only upon the impossibility of knowing objects except un- 249:1/o Transcendental idealists are fond der particular descriptions. And it entails of saying that either knowledge must con- the rejection of any correspondence theory form to objects or objects conform to knowl- of truth. A proposition is true if and only edge: that either how we speak must be a if the state of affairs that it expresses (de- function of things or things must be a func- scribes) is real. But propositions cannot be tion of how we speak.7 But this dichotomy com pared with states of affairs; their rela- is bogus. Science is an activity, a process in tionship cannot be described as one of corre- thought and nature which attempts to ex- spondence. Philosophers have wanted a the- press in thought the natures and constitu- ory of truth to provide a criterion or stamp tions and ways of acting of things that exist of knowledge. But no such stamp is pos- independently of thought. Thought has a re- sible. For the judgement of the truth of a ality not to be confused or identified with the proposition is necessarily intrinsic to the sci- reality of its objects: knowledge may change ence concerned. There is no way in which without objects and objects change with- we can look at the world and then at a sen- out knowledge. There is no correspondence, tence and ask whether they fit. There is just no conformity, no similarity between objects the expression (of the world) in speech (or and thought. Thoughts are only like other thoughts objects (including thoughts) simi- lar to or identical with other things. Things 347 exist and act independently of our descrip- gard some kinds of statements as grounds tions but we can only know them under par- for other kinds of statements, etc. In par- ticular descriptions Descriptions belong to ticular the question of what is capable of in- the world of society and of men; objects be- dependent existence must be distinguished long to the world of nature. We express [our from the question of what must be the case understanding of] nature in thought. for us to know that something is capable of independent existence Thus electrons could 250:1 Science, then, is the systematic at- exist without material things; but we could tempt to express in thought the structures not know this proposition, let us say P, un- and ways of acting of things that exist and less there were material things. The truth- act independently of thought. The world is conditions for our knowledge of P are not the structured and complex and not made for same as the truth-conditions for P. There men. It is entirely accidental that we exist, could be a world without men; but there and understand something about our bit of could not be knowledge without antecedents. it. It is important to avoid the epistemic fallacy here. This consists in confusing the ontological order with the epistemic order, priority in being with priority in deciding claims to being, the question of what has rel- atively underived (or independent) existence with the question of what entitles us to re- 348 CHAPTER 4. METAPHYSICS AND THE PHILOSOPHY OF SCIENCE

4.1 Footnotes to Chap- ter 4

1 See A. C. Danto, ‘Basic Actions’, A.P.Q. Vol. 2 (1965), reprinted in The Philosophy of Action, ed. A. White, pp. 43–58. 2 Cf. G. H. von Wright, op. cit., pp. 66ff. 3 To borrow Buchdahl’s useful concept. See G. Buchdahl, op. cit., p. 3 and passim. 4 Radical conventionalists adopt a mixture of the first and third horns in the sense of allow- ing that we are free to decide what constitutes experience. 5 N. Chomsky, ‘Problems of Explanation in Linguistics’, Explanation in the Behavioural Sci- ences, eds. R. Borger and F. Cioffi, pp. 427–8. I have of course already argued against the use of the concept of ‘idealisation’ to refer to generative structures in 2.4 above. 6 See, e.g. K. R. Popper, op. cit., p. 224. 7 See e.g. H. Schwyzer, Thought and Real- ity: The Metaphysics of Kant and Wittgenstein’, Philosophical Quarterly, 1973, p. 205. Chapter 5

Postscript to the Second Edition

251:1 In this postscript I wish to clarify which I think further work is necessitated. ambiguities in two key terms in the text; 251:2 Two terms, viz ‘law’ and ‘cause’, comment on what I now see as its principal are used in a systematically ambiguous way weaknesses; and indicate some of the ways in throughout the book. In general I think that

349 350 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION the context always determines which usage is the transitive dimension by some term such intended. But my failure to distinguish them as ‘statement’. explicitly may have been a source of gratu- itous misunderstanding. 252:1/oo Secondly, I use the term ‘cause’ to refer both to the antecedent event, con- 251:3/o The term ‘law’ is customarily used dition or agent which triggers a mechanism to refer both to statements of law and to and to the mechanism (and a fortiori the law what such statements designate. A distinc- it grounds) itself. Thus I talk both of men as tion between the two is, of course, implicit in causal agents of sequences of events and of the basic distinction of the work, that is be- the causal laws to which their activity may tween the transitive and intransitive dimen- sometimes give them access, but of which sions in the philosophy of science. Yet the they are not of course (in general) the agents. term ‘law’, together with the terms describ- This duplex use of the concept of causality ing their characteristics (‘normic’, ‘transfac- at a crucial state in the argument may have tual’, etc.), is used indifferently in the text confused some readers. Thus I do not ar- to refer to both. A critic determined to be gue (on pp. 33ff), as e.g. Sharpe supposes, captious may find confusion here where none ‘that the causal agency involved in the sci- in fact exists.1 Were I to rewrite the book I entist’s intervention cannot itself be simply would restrict my use of the term ‘law’ to the reduced to sequence’.2 Rather, I argue that concept in the intransitive dimension, always causal laws cannot, if we are to render in- qualifying its use to denote the concept in telligible the significance of and necessity for 351 the scientist’s experimental activity, be re- ism.4 duced to sequence. My main criticism is of the Humean account of laws. Only in a sub- 255:1 In general the work has been sub- sidiary argument (see e.g. pp. 66, 117) do I ject to three kinds, or levels, of criticism: (I) object to the Humean analysis of the agency of the transcendental method employed; (II) involved. The former implies the latter, but of the particular arguments used; and (III) the converse is not the case (as is shown of the results obtained. I shall restrict my by the possibility of Davidsonian anomalous remarks to (I) and (II), since if the argu- monism).3 The importance of this is that ments are valid and the method sound, the many philosophers, particularly those influ- results must follow. It is, however, worth enced by hermeneutical thought, such as von mentioning that many critics of, or sceptics Wright and Apel, have recently drawn atten- about, (I) and/or (Il) have been willing to tion to our causal activity (in experimenta- concede that transcendental realism of the tion) in science, taking this as a criticism of sort elaborated here, or something very like the empiricist account of causality, without it, ‘has been implicit in the thinking [activ- recognising its significance (cf. p. 54 above): ity] of most creative scientists from Galileo to viz in yielding a definitive criticism of the our own time’.5 Now, if this is the case, and empiricist account of laws, and hence of the transcendental realism does successfully re- whole wretched ontology of empirical real- capture what Bachelard has called the ‘diur- nal’ philosophy of the scientists, i.e. the phi- losophy implicit in their spontaneous prac- 352 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION tice. And if it is also the case that the ity is invalid, and that my argument for its theses of transcendental realism depart rad- intransitive nature itself presupposes what I ically from the ‘nocturnal’ philosophy of the am trying to prove. philosophers, forged in the schools of empir- 256:1 Ruben claims that my contention ical realism, to which scientists too tend to that the experimental scientist produces a return when they reflect upon their practice, sequence of events is false. For, though he then the problem arises of deciphering the produces the antecedent a, and thus indi- meaning, and explaining the mechanism, of rectly produces the consequent b, it does not this discrepancy.6 To this problem I will re- follow, given that a occurs, that he produces turn. the sequence a, b.8 For it might well be the case, according to Ruben, that whenever a 255:2/o A good example of the kind of occurs (however produced) so does b. That criticism directed against transcendental re- is, I have not eliminated the possibility of alism at levels (I) and (II) is provided in empiricist reconstructions of causal laws as a recent book by Ruben.7 Ruben argues sequences. that though reality is indeed, as I claim, structured and intransitive, there cannot be 256:2/o But of course I have. For our any interesting non-question-begging valid causal activity is as much a necessary con- deductive arguments for this conclusion. To dition for the realisation of the consequent this end he seeks to demonstrate that my b, given that a occurs, as it is (in the con- argument for the structured nature of real- text of an experiment) for the occurrence of 353 the antecedent. I explicitly designate (on strued as making a claim about something p. 53 above) the former activity as ‘exper- that bears only a contingent relationship to imental control’ and the latter as ‘experi- the actual world (including that significant mental production’. Experimental produc- subset of it produced by human work). This tion is necessary to ensure the operation of claim, I have argued, is about the operation the mechanism; experimental control for the of a tendency or the working of a mechanism occurrence of a closure. My argument is that irrespective of its actualisation in any partic- without our causal activity, given a, b may ular outcome. And if that is the case, then not, and in general will not, occur. Patently, all the theories based on the flawed princi- if it is the case that our causal activity is ple of empirical-invariance—from the consis- necessary for the realisation of the conse- tency condition of monistic historiography quents of laws, they just cannot be glossed, of science to Feyerabend’s ‘dadaism’,9 from without absurdity, as empirical regularities. the Popper-Hempel theory of explanation to That is, if it is not the case that whenever standard (empirical realist) analyses of coun- a then b then in making a claim about a terfactuals (see p. 158 above), from the 2nd causal law, we cannot, if we are to sustain Analogy to Bachelard’s cogitamus10—must the intelligibility of the experimental estab- all be radically wrong. lishment and transfactual application of our knowledge, be making a claim about a se- 257:1 It is not clear to me why Ruben quence of events. Instead we must be con- should take it as an objection to transcen- dental arguments, and not say mathemati- 354 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION cal proofs, that ‘in a deductively-valid ar- way between different philosophical theories gument nothing can appear in a conclusion from the premiss, established as a result of which was not already, at least covertly, in the analysis into the conditions of the pos- the premisses’11. Still, if it is the case that sibility of change (and criticism) in science, philosophy is, as I have claimed it can be, ‘scientific change (and criticism) is only pos- a conceptual science, then, like any science, sible on the condition that there are (rela- it ought to be able to tell us something we tively) unchanging objects, existing and act- did not already know: it ought to be able ing (relatively) independently of the chang- to surprise us. Philosophy does so when it ing theories of which they are the objects’. (for the first time) makes explicit what is al- From this premiss (only if Q then P) to- ready presupposed by the activities in which gether with the minor premiss, viz that sci- we engage; or when, to put it another way, entific change (and criticism) occurs (P), my it shows the conditions of their possibility. conclusion (Q) does indeed follow. In this sense it is a touchstone of the validity of any 257:2/o Now Ruben contends that I as- argument, transcendental or otherwise, that sume in my argument (on p. 31) that sci- the conclusion be ‘implicit’ in the premiss. entific change consists in changing theories But the interest of a transcendental argu- about the unchanging, and do not (and can- ment clearly does not lie in the formal deriva- not, without circularity) prove it.12 Here it is tion of the conclusion, which is trivial; but in important to distinguish the concept ‘scien- the production of the knowledge of the ma- tific change’ which can function in a neutral 355 jor premiss(i.e. in the analysis). It is in this idealist interpretations of scientific change that the essence of conceptual discovery in . . . according to which there is no neu- philosophy lies. Now such knowledge cannot tral world “shared” by different theories or be held to be already contained in the an- paradigms’.13 So, to ‘complete’ (on which tecedently existing concept which figures in more anon) the argument, let us consider philosophical discourse as ‘scientific change’, them. Such theories posit either ‘incommen- as is shown precisely by the existence of non- surability’ or ‘Kuhn-loss’. Now it is pre- realist explications of it. (It would require cisely a condition of the intelligibility of in- the drastic expedient of some absolute ide- commensurability (non-inter-translatability alist thesis of the identity of opposites, or of the terms of the rival theories) that there perhaps a Platonic theory of anamnesis to exists a field of real objects with respect avoid this conclusion.) to which the rival theories are incommensu- rable. (As I have remarked elsewhere no-one 258:1 But what are my grounds for the bothers to say that the rules of cricket and major premiss in this case? It is clear how football are incommensurable.) In ‘Kuhn- on the assumption of the independent ex- loss’, on the other hand, if it is total, there istence of nature the possibility of conflict- are no objects in common, so that Ruben’s ing, differing, changing, clashing and incon- objection is sustained. But now no sense sistent descriptions all become readily intel- can be given to the concept ‘scientific change ligible. But to leave the matter there would (and criticism)’. For total Kuhn-loss in- be to overlook ‘just those ‘alternative neo- 356 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION volves neither transformation nor discursive order. intelligence, but an archetypal, intuitive un- 259:1 On the conception of philosophy at derstanding constructing its world in a sin- work in this book both the ultimate pre- gle synthetic act14-a possibility from which misses and the immediate conclusions of it is significant that even Feyerabend has philosophical considerations are contingent baulked.15 facts, the former (but not the latter) be- ing necessarily social and so historical. It is only in this relative or conditional sense that 258:2/o I do not think that any of the ob- philosophy can establish synthetic a priori jections so far directed at the main theses of truths (truths about the world investigated the book hold. Moreover I think the tran- by science). Philosophy operates by the use scendental method employed is, in general, of pure reason. But it does not operate by sound. It is, however, certainly the case that the use of pure reason alone. For it exercises the book contains no adequate defence, or that reason always on the basis of prior con- meta-philosophical justification, of the lat- ceptualisations of historical practice, of some ter.16 This is not the place to make good this more or less determinate social form. omission, which I now regard as the most serious in the work. A sequel, to be pub- 259:2/o Now philosophy as so conceived, lished shortly, will treat the matter fully. In can tell us that it is a condition of the possi- the meantime some brief comments on the bility of scientific activities phi, ψ, etc. that question ‘how is philosophy possible?’ are in the world is structured X and differentiated 357

Y. But it cannot tell us what structures it plicitly or explicitly deny it.17 (Thus if laws contains or the ways in which it is differenti- are identified with empirical regularities we ated. These are entirely matters for substan- have the absurdities that men, in their ex- tive scientific investigation. Scientific activ- perimental activity, create or even change ity is a contingent, hisorically transient af- the laws of nature; and that either outside fair. And it is contingent that the world is closed systems nothing governs phenomena, as described by X, Y, Z. But given φ and ψ, so that nature becomes radically indetermin- X and Y must be the case. A ‘deduction’, istic, or that as yet science has discovered no or demonstration, of this necessity, (which laws!) Misunderstandings about the inten- may be termed ‘transcendental’) and which tions of transcendental arguments typically is represented by the major premiss of a tran- stem from the failure to appreciate the crit- scendental argument when set out in its triv- ical contexts in which they are developed- ial ‘deductive’ form, will normally consist of against already existing philosophical theo- two parts, a straightforward ‘positive’ part ries. Thus it is certainly the case that there in which it is shown how e.g. X makes φ in- is no way of demonstrating the uniqueness telligible; and a ‘negative’ part in which it of the conclusion of such an argument in is shown how absurd, incoherent, counter- advance of every possible philosophical the- intuitive or counterfactual results flow from ory.18 But the transcendental consideration the failure to sustain X, typically expressed is not deployed in a philosophical vacuum: it in the form of one or more theories that im- is designed to replace or situate, an existing 358 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION theory; and may come, in time, to suffer a ticular social practices it provides, at least similar fate. Moreover both the acceptability schematically, a way of reconciling transcen- and acceptance of some piece of philosoph- dental and historical analyses of human ac- ical reasoning will depend upon the accept- tivities such as science.19 Secondly, philos- ability and acceptance respectively of the mi- ophy cannot anticipate the form of a suc- nor premisses concerned. Someone who de- cessful scientific practice; that is to say the nies that our knowledge is experimentally es- minor premisses of the arguments it uses tablished and practically applied and that may have to be developed afresh in the case science develops in time need be bound by of each specific science. Does this mean, none of the results of this book. Further it then, that transcendental realism can do no should perhaps be stressed that I have not more than paint its grey on grey? No. For demonstrated that transcendental realism is though I have developed my arguments in the only possible theory of science consistent this book mainly from a consideration of the with these activities; only that it is the only experimental sciences of physics and chem- theory at present known to us that is consis- istry, ideologies, derived from defective con- tent with them. ceptions of those sciences, weigh, like a dead hand, heavily on the shoulders of many of 260:1/o Some implications of this concep- the other sciences, and particularly of course tion of philosophy should be noted. First, the proto-sciences of society and man. Phi- by making the possiblity of philosophical dis- losophy can perform at least two tasks here. course contingent upon the actuality of par- 359

First, freeing these sciences from the intel- dental questions of the form ‘what must be lectual grip of theories secreted by the flat the case for φ to be possible?’ where ‘φ’ de- undifferentiated ontology of empirical real- notes some characteristic activity as concep- ism, it can set the terms for a more ratio- tualised in experience. The conclusion of the nal appraisal of the real problems they face. argument would be a statement of the con- For instance, in the absence of spontaneously ditions of possibility of the particular social occurring, and given the impossibility of ar- activity concerned. However, in opposition tificially creating, closed systems in the hu- to a neo-Kantian stream of thought, tran- man sciences, their criteria for the rational scendental realism allows that such condi- confirmation and rejection of theories cannot tions are real and subject to historical trans- be predictive and so must be exclusively ex- formation, so that the resultant hermeneu- planatory.20 But besides this negative func- tics is contingently critical.21 If this formal tion, another more positive possibility opens analogy between philosophical and social sci- up for philosophy too, For, it shares an affin- entific discourse should prove fruitful then ity with social science in that they both seek, philosophy could live down Hegel’s jibe (and as at least part of their project, to identify Bradley’s that it merely produces bad rea- and describe the conceptions of agents en- sons for what we believe on instinct) and gaged in social practices. Given this, the help to act not just as the underlabourer, possibility is bound to arise of posing for but as the midwife of a science, or group of social practices other than science transcen- 360 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION sciences. structures and a conception of scientific ac- tivity as work; and between the denegation 261:1/o The book aimed to present a sys- of ontology in empirical realism and the view tematic account of science. I believe the ac- of facts, and constant conjunctions, existing count to be substantially correct. But it was quite independently of men. But the per- an error to imply that the account could also ceptive reader will have noted an asymetry be complete. For if philosophy is indeed, as in the development of the argument. For I have been arguing, a conceptual science while the ontology of empirical realism has which takes as its premisses human activi- been explained in terms of a certain concep- ties, situated in time and subject to trans- tion of man, no explanation of the latter has formation, then there is a sense in which been given. It is in this area that I think the philosophy’s work can never be completed. most pressing problems for the further de- Philosophy is in history too. velopment of the Copernican Revolution in 262:1 Philosophy is not independent of the the philosophy of science lie. various sciences, but neither is it reducible to them. Now this book aimed to be a philos- ophy for the sciences (and against the ide- ologies that threaten them). Throughout it I have stressed the close connection that ex- ists between the transcendental realist on- tology of enduring and transfactually active 5.1. FOOTNOTES FOR THE POSTSCRIPT 361

5.1 Footnotes for the and valency. Krige then goes on polemically to ask how I would fit alchemy and phlogiston into Postscript my schema. But my schema is designed to il- lustrate the concept of the real stratification of 1 See e.g. C. Whitbeck, ‘Review of A Realist the world. And it is dear that I can only do Theory of Science’ (hence-force R.T.S.), Philo- this by invoking some or other scientific ontol- sophical Review July 1977, p. 115. It is, how- ogy. Now it is a fact unfortunate though it may ever, going too far when this reviewer takes me to seem to some romantically-inclined philosophers, task for distinguishing a tendency from a power that there is at present no scientific ontology that in two different and incompatible ways within includes the three principles (salt, sulphur and the space of two pages (ibid, p. 118n.8) when on mercury), phlogiston and electrons. Each new those very pages (pp. 230–1) I explicitly distin- scientific breakthrough redraws, or situates the guish two concepts of tendency (tendency1 and possibility of redrawing, the contours of its ter- tendency2), devoting considerable attention to rain of reality. So that from each new stand- an elucidation precisely of their differences! Car- point in theoretical time the history of the sci- oline Whitbeck’s captiousness is surpassed only ence (its positive and negative contributions to by that of John Krige who, writing in Radical the present) looks different, is rewritten, a fact Philosophy 12 (Winter 1975), p 39, represents that Bachelard attempted to register with his me as attempting to deduce the development of concept of ‘recurrence’ (see e.g. G. Bachelard, Le chemistry (presumably on pp. 168–9) from one Rationalisme Applique, P.U.F., Paris 1949, p. 2 inorganic reaction, when it is clear that the reac- & passim. Cf. also, of course, T. S. Kuhn, whom tion cited, that of hydrochloric acid and sodium, Krige cites approvingly, The Structure of Scien- is given only as an example of the kind of reac- tific Revolutions, p. 137 ff.). Now it follows from tion explained by the theory of atomic number the syncategorematic character of philosophical 362 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION discourse, stressed in the book, that I cannot as distinct from possibilites. Moreover events, give a concrete illustration of the philosophical when they are explained, are typically explained ontology of trancendental realism, without mak- as instances of the possible. Real possibilities ing use of some or other scientific ontology. But may or may not have an actual basis: fragility, that does not imply a philosophical commitment for us, does; gravity, for Newton, did not (cf. p. to it. A history of chemical theories represent- 180 above). ing the development of thought (a concept in the Krige argues that if it is discovered that some transitive dimension) not the stratification of the naturally occurring pattern of events repeatedly world (a concept in the intransitive dimension), reveals the presence of a specific tendency ‘the would of course include alchemy, phlogiston and claims made by orthodoxy philosophy of science much else besides. gain respectability’ (op. cit., p. 38). Now while As I have cited Whitbeck and Krige as cap- it is clear that some systems, such as biological tious critics it is perhaps only fair to respond to ones, are more nearly closed (reveal a greater de- their more serious points. Whitbeck finds a diffi- gree of regularity of behaviour, or recurrence of culty in the concept of the non-actual real, con- syndromes) than others, such as social ones, if tending that possibilities and mechanisms ‘must the law-like statements still have to be analysed be actual in order to explain the behaviour which as tendencies and not as invariant conjunctions may occur’ (op. cit., p. 117). Now the non-actual of events, as orthodox philosophy of science re- real includes both real possibilities (and latent quires, then it is not a whit better off. And if, on mechanisms), and their unfulfilled exercise. It is the other hand, they do not, then we are deal- not the case that a tendency has to be actualised ing with a naturally closed system, which tran- in an event to help to explain it (see above pp. scendental realism can of course allow (see p. 91 99–100). But neither is it the case that science above). But orthodox philosophy of science now is concerned only with the explanation of events, faces the problem of what governs phenomena, 5.1. FOOTNOTES FOR THE POSTSCRIPT 363 and would license the application of laws, out- pressive ideology’. Allez les rouges! Now in this side such systems (see p. 65 above). book I have attempted to sketch some of the Whitbeck also claims that I do ‘not suffi- ways in which traditional philosophical concep- ciently appreciate how little can be said . . . tions of science can function as ideologies. But about the general character of the real world’ A Realist Theory of Science does not purport to (op. cit., p. 117), declaring the task I have un- be a political economy, or sociology of science. dertaken to be beyond human capabilities (ibid. Obviously, such work is both legitimate and nec- p. 118). Now, as in 1.4, I argued that every the- essary. ory of science, or epistemology, whether it likes 2 R. A. Sharpe, ‘Review of R.T.S.’, Philosoph- it, admits it or not, presupposes an ontology, the ical Quarterly July 1976, p. 284. onus is surely on any critic to show how a theory 3 See e.g. D. Davidson, ‘Psychology as Philos- of knowledge can do without one. It is that en- ophy’, Philosophy of Psychology, ed. S. Brown, deavour, though not without effects (see e.g. pp. Macmillan, London 1974, reprinted in The Phi- 41–4 above), that I have tried to show is truly be- losophy of Mind, ed. J. Glover, Oxford Univer- yond human capabilities. Knowledge cannot be sity Press, Oxford 1976 pp. 102–3. (For a cri- prized apart from its form, and cut loose from tique of the Davidsonian defence of Humean the- assumptions about what the world must be like ory see pp. 140–1 above). for it to have the form that in some theory of 4 Perhaps it is this which accounts for knowledge it is claimed to possess. From now on Sharpe’s failure to appreciate why ‘the connec- the boot is on the other foot: there is no escape tion between the underlying mechanism and the from ontological commitment. observed phenomena’ cannot, as the Humean Finally, Krige reckons (op. cit. pp. 38–9) that supposes, ‘succumb to an analysis of cause in I fail to describe the mechanisms that would terms of sequence’ (loc. cit). As most readers make science ‘a liberating force and not an op- of the book have fully appreciated, it is precisely 364 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION because of the existence in nature of two kinds that two objects have the same essential nature; of systems, viz open and closed ones, and the just as it is contingent how many other objects fact that human activity is in general necessary do (cf. p. 226 above). for the latter, that laws, if they are to be uni- On the realist theory of universals developed versal and operate independently of human ac- here, there is no puzzle about generality. Invo- tivity, cannot be analysed as sequences at all. cation of the latter as a criterion for the exis- (See e.g. the reviews by R. Harre, Mind Octo- tence of normic behaviour derives from the as- ber 1976, pp. 627–30; S. K¨orner, T.L.S. 11.4.75, sumption of the identity of powerful particulars p. 397; W. Outhwaite, Social Studies of Science, classified together as members of a natural kind, February 1976, pp. 123–7). Incidentally Sharpe which in turn expresses a science’s commitment quite erroneously attributes acceptance of a ‘like to a certain line of enquiry (cf p. 210 above). cause like effect’ principle to me (op. cit. p. 285). Sharpe is also worried about how ‘the suppo- (For an explicit rejection see p. 77 above.) As I sition that there is an underlying structure in repeatedly emphasise I am committed to (rela- virtue of which substances have the features they tive) invariance only at the level of structures, do . . . [turns] the relation from one of analyt- not events (see e.g. p. 219). Still this does leave icity into physical necessity,’ (loc. cit.). But if open the fair question of ‘what entitles us to use my transcendental argument for the intransitive . . . generality as a criterion of the existence of and structured nature of the objects of scientific a causal (nomic) connection?’ (loc. cit.). Now if enquiry (in 1.3) is correct then Sharpe has in- two objects have the same essential nature they verted the problem of the relationship between must behave the same way (see 3.5 above). But physical and logical necessity. For whether or this condition derives from the identity of their not a relationship of natural necessity obtains in natures. Each object must behave as it does be- any given case is quite independent of men, and cause of its essential nature. But it is contingent hence prior to the question of the logical status 5.1. FOOTNOTES FOR THE POSTSCRIPT 365 of the propositions we use to express it (changes New York, 1972, pp. 249–58. in which can, indeed, be shown to have a certain 15 Thus note the overlapping domain marked rationale) (cf. pp. 200–1 above). I cannot give a D in the second diagram on p. 178 of his Against sense to the notion of logical necessity in nature, Method, New Left Books, London, 1975. unless it be taken as referring to that aspect of (For the record I should add that the par- nature which consists in the spatio-temporally ticular argument that Ruben cites (on pp. 31–2 situated actions of men. above) I take only as establishing the intransitiv- 5 S. K¨orner, loc. cit. ity of the objects of experience. The intransitiv- 6 See my forthcoming ‘Philosophies as Ide- ity of the objects of knowledge is established by ologies of Science’, Essays in Marxist-Philosophy the analysis of experimental activity, not simple (Provisional title) eds. J. Mepham & D. Ruben, sense-perception. For me, the objects of knowl- Harvester Press, Hassocks, 1978. edge are disjoint from the objects of experience; 7 D. H. Ruben, Marxism and Materialism, and the conditions of possibility of knowledge Harvester Press, Hassocks, 1977. and experience are not the same.) 8 ibid, p. 131. 16 A paper published in the same year ‘Forms of Realism’, Philosophica 15 (1), 1975 pp. 99–127 9 See my ‘Feyerabend and Bachelard: Two merely touches on the subject. Philosophies of Science’, New Left Review 94 17 cf. W. H. Walsh, Kant’s Criticism of Meta- (Nov.-Dec. 1975). p. 46. physics, Edinburgh University Press, Edinburgh, 10 ibid, p. 54. 1975, pp. 102–6. 11 D. H. Ruben, op. cit., p. 101. 18 S. K¨orner, Categorical Frameworks, Black- 12 ibid, p. 100. well, Oxford, 1970, p. 72. 13 loc. cit. 19 Cf. e.g. S. B. Barnes, Interests and the 14 cf. I. Kant, Critique of Judgement, Hafner, Growth of Knowledge, Routledge and Kegan 366 CHAPTER 5. POSTSCRIPT TO THE SECOND EDITION

Paul, London, 1977, chap. 1. 20 See my ‘On the Possibility of Social Scien- tific Knowledge and the Limits of Naturalism’, Journal for the Theory of Social Behaviour 8 (1) March 1978 21 See my forthcoming The Possibility of Nat- uralism, Harvester Press, Hassocks, 1978. Bhaskar’s Bibliography

263:1 Nb. This bibliography contains only Ross, Oxford 1924. books and articles referred to in the text and 263:5 D. Armstrong, ‘Dispositional Prop- postscript plus a few others. erties’, Analysis 22 (1961–2). 263:2 P. Achinstein, Laws and Explana- 263:6 — A Materialist Theory of Mind, tions, Clarendon Press, Oxford 1971. Routledge and Kegan Paul, London 1969. 263:3 G. E. M. Anscombe, Causality and 263:7 — Belief, Truth and Knowledge, Determination, Cambridge University Press, Cambridge University Press, 1973. 1971. 263:8 J. R. Aronson, ‘Explanation With- 263:4 Aristotle, Metaphysics, ed. W. D. out Laws’, The Journal of Philosophy, Vol.

367 368 BHASKAR’S BIBLIOGRAPHY

66 (1969). Paul, London 1977. 263:9 J. L. Austin, Sense and Sensibilia, 263:17 J. Bennet, ‘Real’, Mind 1966. Clarendon Press, Oxford 1962. 263:18 R. Bhaskar, ‘Forms of Realism’, 263:10 A. J. Ayer, Language, Truth and Philosophica 15(1), 1975. Logic, 2nd ed., Victor Gollancz, London 263:19 — Feyerabend and Bachelard: 1962. Two Philosophies of Science’, New Left Re- 263:11 — Foundations of Empirical view 94 (Nov.-Dec. 1975). Knowledge, Macmillan, London 1963. 263:20 — On the Possibility of Social Sci- entific Knowledge and the Limits of Natu- 263:12 — The Fundamental Questions of ralism’, Journal for the Theory of Social Be- Philosophy, Weidenfeld and Nicolson, Lon- haviour 8 (1), March 1978. don 1973. 263:21 — ‘Philosophies as Ideologies of 263:13 M. R. Ayers, The Refutation of De- Science’, Essays in Marxist Philosophy (Pro- terminism, Methuen, London 1968. visional title), ed. J. Mepham & D. H. 263:14 G. Bachelard, Le Rationalisme Ap- Ruben, Harvester Press, Hassocks 1978. plique, P.U.F., Paris 1949 264:1 — The Possibility of Naturalism, 263:15 F. Bacon, Novum Organum, ed. T. Harvester Press, Hassocks 1978. Fowler, Oxford 1889. 264:1 B. Blanshard, Reason and Analysis, 263:16 S. B. Barnes, Interests and the Open Court, La Salle, Illinois 1962. Growth of Knowledge, Routledge and Kegan 264:1 R. J. Boscovitch, A Theory of Nat- 369 ural Philosophy, M.I.T. Press, Cambridge, Hall, New Jersey 1963. Mass. 1966. 264:9 — ‘What are Physical Theories 264:2 R. Boyle, Collected Works, ed. T. About?’, A.P.Q. Studies in the Philosophy Birch, London 1672. of Science, ed. N. Rescher, Blackwell, Oxford 264:3 R. B. Braithwaite, Scientific Expla- 1969. nation, Cambridge University Press, 1953. 264:10 E. A. Burtt, The Metaphysical Foundations of Modern Physical Science, 264:4 P. W. Bridgman, The Logic of Mod- 2nd ed., Routledge and Kegan Paul, London ern Physics, Macmillan, New York 1932. 264:5 M. Brodbeck, ‘Methodological Indi- 264:11 K. Campbell, ‘One form of scep- vidualism: Definition and Reduction’, Read- ticism about induction’, Analysis Vol. 23. ings in the Philosophy of the Social Sci- Reprinted in The Justification of Induction, ences, ed. M. Brodbeck, Macmillan, New ed. R. Swinburne, Oxford University Press, York 1969. 1974. 264:6 G. Buchdahl, Metaphysics and the 264:12 N. R. Campbell, The Foundations Philosophy of Science, Basil Blackwell, Ox- of Science, Dover 1957 (Physics, The Ele- ford 1969. ments, Cambridge 1919). 264:7 M. Bunge, Causality, Harvard Uni- 264:13 — What is Science?, Dover 1953. versity Press, 1959 (Meridian 1963). 264:14 M. Capek, The Philosophical 264:8 — The Myth of Simplicity, Prentice Impact of Contemporary Physics, van 370 BHASKAR’S BIBLIOGRAPHY

Norstrand, New York 1961. Cioffi, Cambridge University Press, 1970. 264:15 R. Carnap, ‘Testability and Mean- 265:1 A. W. Collins, ‘Explanation and ing’, Readings in the Philosophy of Science, Causality’, Mind 1966. eds. H. Feigl and M. Brodbeck, New York 265:2 J. B. Conant, The Overthrow of the 1953. Phlogiston Theory, Cambridge, Mass. 1950. 264:16 — The Logical Foundations of 265:3 N. Copernicus, De Revolutionibus, Probability, University of Chicago Press, trans. C. G. Wallis, Great Books of the West- 1950. ern World, Vol. 16, Encyclopaedia Britan- nica, Chicago 1952. 264:17 R. Chisholm, ‘The Contrary-to- 265:4 I. Copi, ‘Essence and Accident’, Fact Conditional’, Mind 1946. Reprinted in Journal of Philosophy 1954. Reprinted in Readings in Philosophical Analysis, eds. H. Universals and Particulars, ed. M. J. Loux, Feigl and G. Maxwell, New York 1949. Anchor Books, New York 1970. 264:18 N. Chomsky, ‘Objectivity and Lib- 265:5 W. Craig, ‘The Replacement of eral Scholarship’, American Power and the Auxiliary Expressions’, Philosophical Re- New Mandarins, Penguin, Harmondsworth, view Vol. 65 (1956). 1969. 265:6 M. Crosland, Historical Studies in 264:19 N Chomsky, ‘Problems of Explana- the Language of Chemistry, Heinneman, tion in Linguistics’, Explanation in the Be- London 1962. havioural Sciences, eds. R. Borger and F. 265:7 A. C. Danto, ‘Basic Actions’, A.P.Q. 371

1965. Reprinted in The Philosophy of Ac- Theory Reconsidered’, Philosophical Analy- tion, ed. A. White, Oxford University Press, sis and History, ed. W. H. Dray, New York 1968. 1966. 265:8 D. Davidson, ‘Actions, Reasons 265:14 W. H. Dray, Laws and Explanation and Causes’, Journal of Philosophy 1963. in History, Oxford University Press, 1957. Reprinted in The Philosophy of Action, ed. 265:15 F. Dretske, Seeing and Knowing, A. White, Oxford University Press,’1968. Routledge and Kegan Paul, London 265:9 — ‘Causal Relations’, Journal of 265:16 P. Duhem, The Aim and Structure Philosophy 1967. of Physical Theory, trans. P. Weiner, New 265:10 — ‘Psychology as Philosophy’, York 1962. Philosophy of Psychology, ed. S. Brown, 265:17 A. S. Eddington, The Nature of the Macmillan, London 1974. Reprinted in The Physical World, London 1929. Philosophy of Mind, ed. J. Glover, Oxford 265:18 A. S. Eve, Rutherford, Cambridge University Press, l976. 1934. 265:11 E. J. Dijksterhuis, The Mechani- 265:19 A. C. Ewing, The Fundamen- sation of the World Picture, trans. C. Dik- tal Questions of Philosophy, Routledge and shoorn, Clarendon Press, Oxford 1961. Kegan Paul, London 1931. 265:12 I. Dirac, ‘Is there an aether?’ Na- 265:20 B. A. Farrel, ‘On the Design of a ture 168. Conscious Device’, Mind 1970. 265:13 A. Donagan, ‘The Popper-Hempel 265:21 P. K. Feyerabend, ‘Explana- 372 BHASKAR’S BIBLIOGRAPHY tion, Reduction and Empiricism’, Minnesota Vol. 37 (1969). Studies in the Philosophy of Science Vol III, 266:5 A. Flew, An Introduction to West- eds. H. Feigl and G. Maxwell, University of ern Philosophy, Thames and Hudson, Lon- Minnesota Press, Minneapolis don 1971. 265:22/o — ‘Problems of Empiricism’, 266:6 M. Foucault, The Order of Things, Beyond the Edge of Certainty, ed. R. G. Tavistock, London 1970. Colodny, Prentice Hall, New Jersey, 1965. 266:7 D. Gasking, ‘Causation and 266:1 — ‘Problem of Empiricism II’, The Recipes’, Mind 1955. Nature and Function of Scientific Theories, 266:8 P. Gardiner, The Nature of Histor- ed. R. G. Colodny, University of Pittsburgh ical Explanation, Oxford University Press, Press, 1970. 1952. 266:2 — ‘Against Method’, Minnesota 266:9 P. T: Geach, ‘Aquinas’, Three Studies in the Philosophy of Science Vol IV, Philosophers, G. E. M. Anscombe and P. T. eds. M. Radner and S. Winokur, University Geach, Blackwell, Oxford 1961. of Minnesota Press, 1970. 266:10 J. Gibbs and W. Martin, Status, 266:3 — Against Method, New Left Integration and Suicide, University of Ore- Books, London 1975. gon Press, Eugene, Oregon 1964. 266:4 M. Fisk, ‘Are there necessary con- 266:11 J. J. Gibson, The Senses Consid- nections in nature?’, Philosophy of Science ered in Perceptual Systems, George Allen 373 and Unwin, London 1966. ery, Cambridge University Press, 1963. 266:12 N. Goodman, Fact, Fiction and 266:18 — ‘A Picture Theory of Theory Forecast, 2nd ed., Bobbs-Merill, New York Meaning’, The Nature and Function of Sci- 1959. entific Theories, ed. R. G. Colodny, Univer- 266:13 — The Structure of Appearance, sity of Pittsburgh Press 1970. 2nd ed., Bobbs-Merill, New York 1966. 266:19 — Observation and Explanation, 266:14 A. Grunbaum, ‘Temporally Asym- Allen and Unwin, London 1972. metric Principles, Parity between Explana- 266:20 G. Harman, ‘Enumerative Induc- tion and Prediction and Mechanism versus tion as Inference to the Best Explanation’, Teleology’, Philosophy of Science Delaware Journal of Philosophy 1968. Seminar Vol 1, ed. B. Baumrin, New York 266:21 R. Harre, ‘Powers’, B.J.P.S. Vol. 21 1963. (1970). 266:15 E. Grunberg, ‘The Meaning of 267:1 — Principles of Scientific Thinking, Scope and External Boundaries of Eco- Macmillan, London 1970. nomics’, The Structure of Economic Science, 267:2 — Philosophies of Science, Oxford ed. S. R. Krupp, Prentice Hall, New Jersey, University Press, 1972. 1966. 267:3 — ‘Surrogates for Necessity’, Mind 266:16 S. Hampshire, Thought and Ac- 1973. tion, Chatto and Windus, London 1959. 267:4 — ‘Review of A Realist Theory of 266:17 N. R. Hanson, Patterns of Discov- Science’, Mind, October 1976. R. Harre and 374 BHASKAR’S BIBLIOGRAPHY

E. H. Madden, ‘Natural Powers and Powerful Indianapolis 1966. Natures’, Philosophy Vol. 45 (1973). 267:13 — In Defence of Objectivity, Ox- 267:5 — Causal Powers, Blackwell, Ox- ford University Press, 1973. ford, 1975. 267:14 T. Hobbes, ‘On Human Nature’. 267:6 R. Harre and P. Secord, The Expla- Reprinted in Body, Man and Citizen, ed. R. nation of Social Behaviour, Blackwell, Ox- S. Peters, Macmillan, New York 1967. ford 1972. 267:15 M. Hollis and E. J. Nell, Rational 267:7 R. F. Harrod, The Life of John May- Economic Man, Cambridge University Press, nard Keynes, Macmillan, London 1951 1975. 267:8 H. L. A. Hart and A. M. Honore, Causation in the Law, Oxford University 267:16 M. Hollis and E. J. Nell, Rational Press 1959. Economic Man, C.U.P., 1974. 267:9 G. F. W. Hegel, Philosophy of Right, 267:17 D. Hume, A Treatise on Human trans, and ed. T. H. Knox, Oxford 1942. Nature, ed. L. A. Selby-Biggs, Clarendon 267:10 W. Heisenberg, Physics and Phi- Press, Oxford 1967. losophy, Allen and Unwin, London 267:18 An Enquiry Concerning Human 267:11 C. G Hempel, Aspects of Scientific Understanding, ed. A. Selby-Biggs, Claren- Explanation, Free Press, New don Press, Oxford 1962. 267:12 M. B. Hesse, Models and Analogies 267:19 T. W. Hutchinson, The Signifi- in Science, University of Notre Dame Press, cance and Basic Postulates of Economic The- 375 ory, 2nd ed. New York 1968. Press, Chicago 1962. 267:20 W. F. Johnson, Logic, Cambridge 268:3 W. Kneale, Probability and Induc- 1921. tion, Clarendon Press, Oxford 1949 267:21 W. D. Joske, Material Objects, 268:4 L. Kolakowski, Positivist Philoso- Macmillan, London 1967. phy, Penguin, Harmondsworth; 1972. 267:22 C. B. Joynt and N. Rescher, ‘The 268:5 S. K¨orner, ‘Substance’, P.A.S. Problem of Uniqueness in History, History Suppl. Vol. 35 (1964). and Theory Vol. 4. 268:6 — Experience and Theory, Rout- 267:23 I. Kant. On the Distinctiveness ledge and Kegan Paul, London of the Principles of Natural Theology and 268:7 — Categorical Frameworks, Black- Morals, 1764. well, Oxford 1970. 267:24 — Critique of Pure Reason, trans. 268:8 — ‘Review of A Realist The- N. Kemp Smith, Macmillan, London 1970. ory of Science’, Times Literary Supplement 267:25 — Critique of Judgement, trans. J. 11.4.1975. Bernard, Hafner, New York 1972. 268:9 A. Koyr´e, ‘The Influence of Philo- 268:1 — Metaphysical Foundations of sophical Trends in the Formulation of Sci- Natural Science, trans. J. Ellington Bobbs- entific Theories’, The Validation of Scien- Merill, New York 1970. tific Theories, ed. P. G. Frank, Beacon Press, 268:2 J. J. Katz, The Problem of Induc- Boston 1959. tion and its Solution, University of Chicago 268:10 — Metaphysics and Measurement, 376 BHASKAR’S BIBLIOGRAPHY

Chapman and Hall, London 1968 Order, New York 1929 268:11 J. Krige, ‘Review of A Realist The- 268:18 J. Locke, An Essay Concerning Hu- ory of Science’, Radical Philosophy 12 (Win- man Understanding; ed. A. C. Fraser, Dover ter 1975). 1959. 268:12 T. S. Kuhn, The Structure of Sci- 268:19 J. Losee, An Historical Introduc- entific Revolutions, 2nd ed, University of tion to the Philosophy of Science Oxford Chicago Press, 1970. University Press, 1972 268:13 I. Lakatos, ‘Criticism and the 268:20 E. Mach, Popular Scientific Lec- Methodology of Scientific Research Pro- tures, Chicago 1894. grammes’, Criticism and the Growth of 268:21 J. L. Mackie, Truth, Probability Knowledge, eds. I. Lakatos and A. Musgrave, and Paradox, Clarendon Press, Cambridge University Press, 1970 268:22 E. H. Madden, ‘Hume and the 268:14 P. S. de Laplace, A Philosophical Fiery Furnace’, Philosophy of Science Essay on Probabilities, New York 1951. 268:23 — ‘Scientific Explanations’, Re- 268:15 — The System of the World, Lon- view of Metaphysics 1973 don 1809. 268:24 E. H. Madden and P. Hare, ‘The 268:16 D. Lecourt, Marxism and Episte- Powers That Be’, Dialogue 1971 mology, New Left Books, London 268:25 M. Mandelbaum, ‘Historical Ex- 268:17 C. I. Lewis, Mind and the World planation: The Problem of Covering Laws’, 377

History and Theory Vol. I. Routledge and Kegan Paul, London 1961. 269:1 N. Maxwell, ‘Can there be necessary 269:8 O. Neurath, ‘Universal Jargon and connections between successive events?’, Terminology’, P.A.S. Vol. 41. B.J.P.S. Vol. 19 (1967). Reprinted in The 269:9 I. Newton, Principia Mathemat- Justification of Induction, ed. R. Swinburne, ica, 2nd ed., University of California Press, Oxford University Press, Berkeley 1947. 269:2 D. H Mellor, ‘Physics and Furni- 269:10 J. Nicod, Foundations of Geome- ture’, A.P.Q. Studies in the Philosophy of try and Induction, trans. P. Weiner, London Science, ed. N. Rescher, Blackwell, Oxford 1934. 1969. 269:11 W. Outhwaite, ‘Review of A Real- 269:3 A. Michotte, The Perception of ist Theory of Science’, Social Studies of Sci- Causality, New York 1963. ence, February 1976. 269:12 J. Passmore, ‘Explanations in Sci- 269:4 J. S. Mill, A System of Logic, 8th ence, History and Everyday Life’, History ed. London 1961. and Theory, Vol. II. 269:5 — Auguste Comte and Positivism, 269:13 — Man’s Responsibility to Nature, 2nd ed., London 1866. Duckworth, London 1974. 269:6 M. J. Mulkay, The Social Process of 269:14 W. Pelling, Modern Britain 1885– Innovation, Macmillan, London 1955, Nelson, Edinburgh 1960. 269:7 E. Nagel; The Structure of Science, 269:15 J. Piaget, The Child’s Perception 378 BHASKAR’S BIBLIOGRAPHY of Physical Causality, N.J. 1960. emies, Vol. I, Routledge and Kegan Paul, 269:16 — Structuralism, Routledge and London 1962. Kegan Paul, London 1971. 269:24 — The Logic of Scientific Discov- 269:17 C. S. Pierce, Collected papers Vol. ery, Hutchinson, London 1959. II, eds. Hartshorne and Weiss, Cambridge, 269:25 — Conjectures and Refutations, Mass. 1932. Routledge and Kegan Paul, London 269:18 H. Poincare, Science and Hypoth- 269:26/o — ‘Normal Science and Its Dan- esis, trans. G. B. Halsted, New York 1903. gers’, Criticism and the Growth of Knowl- 269:19 M. Polanyi, Personal Knowledge, edge, eds. I. Lakatos and A. Musgrave, Cam- Routledge and Kegan Paul, London bridge University Press, 1970. 269:20 — The Tacit Dimension, Routledge 270:1 — Objective Knowledge, Clarendon and Kegan Paul, London 1967. Press, Oxford, 1972. 269:21 — ‘The Structure of Conscious- 270:2 H. Putnam, ‘The Analytic and the ness’, The Anatomy of Knowledge, ed. M. Synthetic’, Minnesota Studies in the Philos- Grene, Routledge and Kegan Paul, London ophy of Science Vol. III, eds. H. Feigl and G. 1969. Maxwell, Minneapolis 1962. 269:22 K. R. Popper, The Poverty of His- 270:3 R. A. Putnam, ‘Seeing and Observ- toricism, Routledge and Kegan Paul, Lon- ing’, Mind 1969. don 1957. 270:4 W. V. O. Quine, ‘Designation and 269:23 — The Open Society and Its En- Existence’, Readings in Philosophical Anal- 379 ysis, eds. H. Feigl and W. Sellars, New York New York 1963. 1949. 270:12 D. H. Ruben, Marxism and Mate- 270:5 — Methods of Logic, 2nd ed., Lon- rialism, Harvester Press, Hassocks 1977. don 1962. 270:13 R. Rudner, The Philosophy of So- 270:6 — From a Logical Point of View, cial Science, Prentice Hall, New Jersey, 1966. Cambridge, Mass. 1953. 270:14 B. Russell, Mysticism and Logic, 270:7 — Ontological Relativity and Other Allen and Unwin, London 1962. Essays, New York 1969. 270:15 A. Ryan, The Philosophy of John 270:8 A. Quinton, The Nature of Things, Stuart Mill, Macmillan, London 1970. Routledge and Kegan Paul, London 1973. 270:16 G. Ryle, The Concept of Mind, 270:9 F. P. Ramsey, The Foundations of Penguin, Harmondsworth, 1963. Mathematics, Routledge and Kegan Paul, London 1931. 270:17 — Dilemmas, Cambridge Univer- sity Press, 1964. 270:10 J. R. Ravetz, Scientific Knowledge and Its Social Problems, Clarendon Press, 270:18 W. Salmon, ‘Foundations of Scien- Oxford 1971. tific Inference’, Mind and Cosmos, ed. R. G. Colodny, Pittsburgh 1966. 270:11 V. Ronchi, ‘Complexities, advances and misconceptions in the development of 270:19 K. Schaffner, ‘Correspondence the science of vision: what is being discov- Rules’, Philosophy of Science Vol. 36 ered?’, Scientific Change, ed. A. Crombie, 270:20 M. Schlesinger, ‘The Prejudice of 380 BHASKAR’S BIBLIOGRAPHY

Micro-Reduction’, B.J.P.S. Vol. 12. Routledge and Kegan Paul, London 1963. 270:21 H. Schwyzer, ‘Thought and Real- 271:4 D. Shapere, ‘Meaning and Scien- ity: The Metaphysics of Kant and Wittgen- tific Change’, Mind and Cosmos, ed. R. G. stein’, Philosophical Quarterly 1973. Colodny, Pittsburgh 1966. 270:22 M. Scriven, ‘Truisms as the 271:5 R. A. Sharpe, ‘Review of A Realist Grounds for Historical Explanation’, Theo- Theory of Science’, Philosphical Quarterly, ries of History, ed . P. Gardiner, Free Press, July 1976. New York 1959. 271:6 B. F. Skinner, Beyond Freedom and 270:23 — ‘Explanation, Prediction and Dignity, Jonathan Cape, London 1972. Laws’, Minnesota Studies in the Philosophy 271:7 J. J. C. Smart, Philosophy and Sci- of Science Vol. III, eds. H. Feigl and G. entific Realism, Routledge and Kegan Paul, Maxwell, Minneapolis 1962. London 1963. 271:1 — ‘Causes, Connections and Condi- 271:8 L. S. Stebbing, Philosophy and the tions in History’, Philosophical Analysis and Physicists, London 1938. History, ed. W. H. Dray, New York 1966. 271:2 W. Sellars, ‘The Language of The- 271:9 P. Strawson, An Introduction to ories’, Current Issues in the Philosophy of Logical Theory, Methuen, London 1952. Science, eds. H. Feigl and G. Maxwell, New 271:10 — Individuals, Methuen, London York 1961. 1959. 271:3 — Science, Perception and Reality, 271:11 R. Taylor, Action and Purpose, 381

Prentice Hall, New Jersey, 1966. ed. R. P. Wolff. 271:12 W. Thomson (Lord Kelvin), Notes 271:20 W. H. Walsh, Kant’s Criticism of of Lectures on Molecular Dynamics, 1884. Metaphysics, Edinburgh University Press, 271:13 S. Toulmin, The Philosophy of Sci- 1975. ence, Hutchinson, London 1953. 271:21 M. Weber, The Methodology of the 271:14 — ‘Crucial Experiments: Priestley Social Sciences, Free Press 1949. and Lavoisier’, The Journal of the History of 271:22 C. R. Weld, History of the Royal Ideas, Vol. XVIII (1957). Society Vol. I, London 1848. 271:15 — Human Understanding Vol. I, 271:23 C. Whitbeck, ‘Review of A Real- Clarendon Press, Oxford 1973. ist Theory of Science’, Philosophical Review, 271:16 J. Tyndall, ‘Scientific Uses of Imag- July 1977. ination’, Fragments of Science for Unscien- 271:24 P. Winch, The Idea of a Social Sci- tific People, New York 1872. ence, Routledge and Kegan Paul, London 271:17 M. Vernon, The Psychology of Per- 1959. ception, Penguin, Harmondsworth, 1962. 271:25/o L. Wittgenstein, Tractatus 271:18 W. Wallace, Causality and Scien- Logico-Philosophicus, trans. D. Pears and tific Explanation Vol I, University of Michi- B. Mc.Guiness, Routledge and Kegan Paul, gan Press, 1972. London 1961. 271:19 W. H. Walsh, ‘Categories’, Kant- 272:1 — Philosophical Investigations, Studien Band 45 (1954). Reprinted in Kant trans. G. E. M. Anscombe, Blackwell, Ox- 382 BHASKAR’S BIBLIOGRAPHY ford 1953. 272:2 G. H. von Wright, Explanation and Understanding, Routledge and Kegan Paul, London 1971. Bibliography

[ABC+98] Margaret Archer, Roy Bhaskar, [Bha78] Roy Bhaskar. A Realist Theory of Andrew Collier, Tony Lawson, Science. Harvester Wheatsheaf, and Alan Norrie, editors. Crit- London and New York, second ical Realism: Essential Read- edition, 1978. ings. Routledge, London and New York, 1998. [Bha89] Roy Bhaskar. The Possibility of Naturalism: A Philosophical Cri- [Bha75] Roy Bhaskar. A Realist Theory tique of Contemporary Human of Science. Leeds Books Ltd., Sciences. Harvester-Wheatsheaf, 84 Woodhouse Lane, Leeds, LS2 London and New York, second 8AB, U.K., 1975. edition, 1989.

383 384 BIBLIOGRAPHY

[Bha97] Roy Bhaskar. A Realist Theory of Science. Verso, London and New York, reprint of the 1977 second edition, 1997. [Col94] Andrew Collier. Critical Re- alism: An Introduction to Roy Bhaskar’s Philosophy. Verso, London, New York, 1994.