![]()
Part One
The Development of
Critical Realism
![]()
Chapter One
From a Philosophy of Science
to a Philosophy of
Universal Self-realisation*
Part I: From Science to Freedom
It is a real pleasure to be here with you in Mumbai and to be the guest of this institute. What I want to do is to talk this morning and tomorrow about the progressive development of critical realism starting out from a concern with science, through various stages, to a concern with questions of human realisation and ultimately universal self-realisation. There are five stages, as I understand it, in the development of critical realism. It started out as a philosophy of science, a critique of positivism but also of neo-Kantianism and radical philosophers of science like Kuhn and Feyerabend who said many shocking things which have also been resumed in postmodernist discourse today. That I call transcendental realism. Then it moved on to critical naturalism and concerned itself with the dispute between naturalists and anti-naturalists, between positivism and hermeneutics, and it tried to resolve this dispute. Basically it was oriented against the dualisms that beset social theory in the mid- to late 1970s and still to a large extent do today. Those are the two things that I will concentrate on today, that is problems in the philosophy and methodology of science, but particularly leading on to social science. Tomorrow I will go through the next three phases of critical realism, which I will briefly mention today. The third stage in critical realism broke down one particular dichotomy, characteristic of the dualisms of social science, which was very popular and insistent particularly in western thought: that one could not move from a factual statement or any set of factual statements to a value judgement. This prohibition was called Hume's law, and I argued that one could move from facts to values. And I did this through what I called the theory of explanatory critique. This then provided the lynchpin by means of which I moved from a concern with science to a concern with questions of values and human freedom and emancipation. The fourth stage of the development was a dialectical one in which I developed a system which I called dialectical critical realism, which went into dialectical notions such as absence, totality, negativity and so on. In the latest stage of my work I have gone on to questions of the convergence of east and west liberatory thought around what could be loosely called a spiritual dimension. This I have called transcendental dialectical critical realism. Those later stages, particularly the dialectical and the spiritual turns within critical realism, I will be dealing with later. Here I will mainly be talking about science and social science.
So how did critical realism start off? It started off really as an account of science which critiqued the seeming incapacity of philosophies of science to really say anything about the world. As Professor Singh has mentioned I was initially a practical economist and I wanted to do a thesis on the relevance of economic theory to underdeveloped countries. And it seemed to me to be very difficult to do this in a rational manner because most philosophy as we have it today embodied presuppositions about the nature of the world which were so obviously false that we have no guidelines for any sort of assessment of the relevance of economic theory to the problems of developing countries. So I had to go back to what was probably my first love, intellectually, that was philosophy. What I found was that existing philosophy of science, in the mid- and late 60s, when I started work on the movement of thought which eventually became critical realism, really pivoted on a central theory, that is the Humean theory of causal laws. This was a theory that the natural world was governed by causal laws which were understood as empirical regularities, that was as constant conjunctions of events. This theory underpinned the very fashionable Popper/Hempel model of explanation. I think you are probably all familiar with this, this is still the orthodoxy. It is called the deductive nomological model of explanation and it says that to explain an event you subsume it under a set of initial conditions and a set of covering laws; that is how you effect explanation. Now those covering laws are empirical regularities, constant conjunctions of events, Humean empirical regularities. This pivotal Humean theory of causal laws finds its way into very radical approaches to social science. For example Habermas, in describing in his theory of our three constitutive interests in knowledge, our knowledge-constitutive interest in prediction and control, presupposes a world which can be described and explained in a Popperian/Hempelian way. That account of reality informs his understanding of nature. Even without going to Habermasâ standard anti-naturalist account of social science the dominant hermeneutical accounts of the 70s and 80s all took a very broad contrast and argued for the categorial differentiation of social science, the absolute differentiation of social science from natural science, because they believed that the natural world was described by constant conjunctions of events.
We can see this if we take what was a very influential text book in the 70s which was Winch's Idea of a Social Science. His argument was that social phenomena could not be understood in the same way as natural phenomena but instead had to be understood in terms of the rule-governed linguistic paradigm that Wittgenstein had enunciated in his Philosophical Investigations. This led to two things, and there are essentially only two main arguments in his book. One is that constant conjunctions of events, that is empirical regularities, were neither necessary, as Weber had argued, nor sufficient for understanding social life. This was achieved in an entirely different way, namely by the discovery of intelligible links in its subject matter. You could certainly say that that is what we are concerned to do in social science, namely to discover intelligible links. But of course the idea that natural science is just concerned with the search for constant conjunctions is completely false, as I will go on to show in a moment. But it was tacitly presupposed and it provided the grounds for Winch's contrast. His second argument for the specificity of the social and the categorial differentiation of the social and natural sciences turned on the fact that social things, as distinct from physical things, have no existence apart from our conceptualisation of them. That of course is quite true, and that marks, at least in the first instance, a major difference between social and natural scientific methodology. But when he came to give his positive account of how social science was done we find there the displacement of themes from positivist philosophy of science. Thus rules function as a normative displacement of empirical regularities and the assumption is that conceptualisations which are ordered in a rule-like way completely exhaust the subject mater of social science and that they are more or less incorrigible. So there is nothing for social science other than investigating the way that agents understand their own subject matter as intelligible, and furthermore, that social science must accept the way that the social agents interpret their own subject matter as incorrigible. Those were assumptions or theses that I want to take issue with. So that is really why I felt that it was very important to come to terms with the Humean theory of causality.
So what we do in philosophy is that we start from phenomena, as in any other subject matter. You cannot just say something is false, you cannot just juxtapose in philosophy your idea of the world with something like the world. What you have to do is engage in an immanent critique of some existing conceptualisation or theory of the world. You have to engage in a reassessment of something which your protagonist thinks is pivotal and which you can show he has given a wrong account of.
Now there was one thing that positivism and positivists, empiricists and neo-Kantians all agreed was crucial to science, and that was experience. Experience, and more particularly experimental activity, is the heart of the empiricistsâ account of science. Actually I did not disagree with this, but whether I did or did not, on the method of immanent critique that had to be my starting point. What I did was try to show that experimental activity actually presupposes a different ontology or account of the world. So let us see how this would work out. What the experimental scientist does in the laboratory is to artificially generate a closure of what is essentially an open system, in order to identify the working of a single generative mechanism or causal complex, a single process,1 or complex, totality or field, whatever the object of study, in isolation from the influence of other factors. Only when he has effected that closure experimentally can he identify an empirical invariance. This means that what he is looking for, the object of scientific investigation, is not a surface pattern of events, because the surface pattern of events in natural life is chaotic. We do not actually find empirical invariances anywhere in the world happening spontaneously. The whole point of experimental activity is to generate an empirical invariance. Now if it is artificially produced, then we have to ask: is it that all man is doing in the experimental laboratory is discovering something he himself produced? The answer is of course not. Because what he or she is trying to do is isolate a structure, a causal complex, a process which will work independently of that closure, independently of whether he artificially closes the system or not. In virtue of that he can then apply it to the open systemic world where it will act in co-determination with a multiplicity of other factors, a multiplicity of other mechanisms, a multiplicity of other agencies. To give you an example, the law of gravity is operating on me now but I am not falling to the ground, it is operating tendentially, it is one of the factors that go to explain why I am sitting relatively stationary. It is actually not operating empirically, it is operating at a supra-empirical level, what I call transfactually, as one of the tendencies which are operating on me now. If you wanted to do a test of the law of gravity we would have to set up a laboratory situation and then measure the rate at which a heavy object falls to the ground. That we can only do in a few special contexts.
This had a very important implication because it does mean that what the scientist is doing is searching for something behind the pattern of events which generates them, whether or not we have a closure of that system of eventsâirrespective of the human activity which discovers it. So what I did was to argue that the field or the domain of the real is greater, more encompassing than the field of the actual, which describes the pattern of events; and that in turn is greater than the field of the empirical, which describes the pattern of events that we actually apprehend. Because clearly there would be things, and there would be events, actual phenomena, even if humanity was not here to observe it, even if one particular agent was not here to observe it, or there was no agent to observe itâ there would still be patterns in the world. The differentiation between the real, the actual and the empirical was a big break with empiricist ontology. As was the view of causal laws as the workings of generative structures, causal mechanisms and processes; that is, as being tendencies, tendencies which could be possessed without being exercised, which could be exercised without being realised, which could be realised without being manifest in a one to one way with any set of phenomena, and which also could be manifest without being detected or identified by men, that is by humanity. This was a radically new account of causality and it led to a completely different ontology.
It led to a view of the world as stratified, differentiated and changing as distinct from the implicit, Humean empirical realist ontology which informed almost all philosophy of science since at least the days of Hume. It was there in Kant, also in Hegel in his account of nature, but probably goes back far longer than Hume, perhaps back to Aristotle in the west, and it was certainly there as the general assumption of philosophers of science in the 60s and 70s. So we hadâthrough this analysis of experimental activity and a similar analysis of the conditions of possibility of applied activity, that is, of what happens when we apply science in the open systemic world where we find that exactly the same categorial distinctions are presumedâa radically different account of the world.
However, at a meta-level it was very difficult to say this, because Hume and more especially Kant had declared a taboo on ontology. To use Wittgenstein's words, all we could do in philosophy was talk about what he called the net, the net of language, the network, the frameworkânot what the network described. In other words we could not talk directly about the world. Now it seemed to me pretty obvious that any claim about knowledge must tacitly presuppose something about the objects of knowledge. But that is not in itself a very strong argument so what our transcendental argument from experimental activity had done was to actually ask the Kantian question: what must be the case for sense experience, in its specific form of experimental activity, to be possible? But it had come to a radically non-Kantian conclusion that the world must be structured. Kant, you will remember, had talked about the structure within us and the way in which we, as scientists, imposed the categories and schematisms on the empirical manifold. So there was structure inside, but of what there was outside us, we could not say anything at all. The best we could say was that there was a thing in itself.
This taboo on ontology is still very strong. I call that taboo the epistemic fallacy. Let us look at some of its manifestations today. In postmodernism, in discourse theory, there is a general assumption that all you can do is talk about talk. It is most clearly explicit in the work of discourse theorists like Ernesto Laclau, but it is there in Derrida in a slightly different form, it is there in others associated with poststructuralism and postmodernism. So discourse becomes a kind of intertextuality, a kind of relating of one text to another text or talk. But what you have to do is ask what is the status of that talk: is the talk real or not? If the talk is not real then it can have no causal effect, then you have to ask what is the point of the talk at all. If the talk is admitted to be real, then you can re-refer to it; if the talk is to be intelligible it must be possible for that talk, that statement, that discourse, to become an object of a subsequent act of reference. We must be able to refer to what we have said again, and if we can refer to what we have said again, then there is at least one real object in the world. That is talk. So that is the thin edge of the wedge: because once you accept the reality of discourse, of talk, then you can accept the reality of a referent for any discourse, just a straight first-order referent. So when I am talking about a table or when I am talking about culture, there is a referent in the real world which is that which I talk about. In the language of semiotics, what we have is not only the signifier, the talk, and the signified, the meaning or the concept, but we have the referent, the thing or object talked about by the word or words, and so we have the semiotic triangle. What the postmodernists and many structuralists and post-structuralists did was to leave out the referent. What realists tried to do was to put the referent back in.
This mistake of discourse theory is sometimes corrected by saying, well of course we admit there is the thing in itself, but there is no way to describe or get at it except by talk. So what is the point of having an ontology, because after all anything that you know you know under a particular description. That is fine. There are two ways to take this line of thought further. The first is thisâit has often been said to me, by postmodernists and discourse theorists: of course we do not deny that there are things in themselves, and nor would I deny that there are things like stones, because you do not need to conceptualise them. But then I said to them, well, if you accept the reality of stones then why will you not accept the reality of the molecules and atoms that constitute them, why wonât you accept the reality of the universe that englobes them. The thin edge of the wedge has already been made, once you can say anything about the world then you are into ontology. The question then is not whether or not to do ontology. The question is whether your ontology is correct or not, which means whether it is adequate to your subject matter.
Another line that postmodernists and discourse theorists, contemporary deniers of ontology, will take is the more classical Kantian response, which is to say well, this is all just anthropic, something which is happening subjectively within myself. So let us look at Kant from that standpoint. Kant as you know had a view, contrary to Hume, that our knowledge was structured, and actually the neo-Kantians of the late 60s and 70s like Popper, Hanson, Harré, Hesse and so on (as did the more radical post- Kantians, Feyerabend and Kuhn) also had the same view. Now in its classical, Kantian form the structuration of knowledge occurs through the imposition of the categories. Now what was it that applied the categories to the empirical manifold sensed in sense experience, what was it that did the schematisms that made empirical science possible as a structured totality? It was the transcendental ego. And what was the status of the transcendental ego? For Kant it was unknowable: it was a thing in itself. But surely, if knowledge was real then that which made the knowledge had to be real. So that is the great aporia, that was the great weakness, the great problem with Kantianism is the fact that it hangs on an unknowable thing in itself inside. Not only is there an unknowable thing in itself outside, the whole system is posited on a transcendental ego or self which is the unknowable synthesiser of the empirical manifold. So ultimately the structure collapses in on itself. And something like that tendency of thought happens within all forms of neo-Kantianism and all systems of thought which take their lead from Kant.
Now let us revert to the main thread of my argument. There were many philosophers of science around that time who were having problems with orthodox philosophy of science. They were saying it leaves something out, it leaves out the way in which reality is pictured in the scientific imagination, the creative imagination of the scientists. So that certainly empirical invariances cannot be sufficient for a scientific theory. But what I was arguing was not only was it not sufficient, but it was not even necessary, because it was in fact false. It was false that the world was constituted by constant conjunctions of eventsâfalse, because they did not actually occur. This was a much more radical break and it gave us a structured world out there and a structured world in here, in the scientific community, knowledge was structured and that reflected in some way the stratification of the world.
Thus, one of the first and most fundamental theorems of transcendental realism, after the categorial distinctiveness of the real, the actual and the empirical, and the analysis of laws as tendencies, was the distinction between what I called then the intransitive dimension and the transitive dimension. More simply between ontology and epistemology, between the world and knowledge. And what I argued for was that ontological realism was quite compatible with epistemological relativism, pluralism, diversity and indeed fallibilism. This is what neo-Kantians have got right: knowledge was a social process, but it was a social process which was designed to capture the ever deeper stratification of the world.
To show the sort of problem that philosophers of science in those days got into because they did not have a concept of the world, let us take Kuhn and Feyerabend. They formulated what was known as the problem of incommensurability. What was this problem? What they noted was that Newtonian theory and Einsteinian theory were so radically different in their conceptual structures that they shared no meanings in common. If they shared no meanings in common there was no way to compare them. So that is why they said they were âincommensurableâ. This seemed to make science a process of irrational breaks or splitsâthat was actually how Kuhn formulated it, though he did it in a sociologically sophisticated way. Now let us see how a critical realist might approach this phenomenon of incommensurability. The critical realist does not deny the phenomenon at all, but by adding to that level of epistemological relativism, the levels of ontological realism and judgmental rationalism, the second and third levels of what we could call the holy trinity of transcendental realism as an account of science, it resolves the problem. First of all, the very formulation of the problem of incommensurability presupposes common referentiality, that there is an object world in common. No one bothers to say that physics and cricket are incommensurable or that Newtonian theory is incommensurable with classical music. Why? Because they are not describing the same world; intuitively you know that Newton and Einstein are describing the same world, describing the same phenomena. The problem howeve...
