Critical Realism

Critical realism is a way of understanding the nature of both natural and human social sciences. Particularly in the case of the latter, it also advocates some and argues against other ways of practicing them. As a version of realism, it is committed to the view that the objects of scientific knowledge both exist and act independently of our beliefs about them. However, the qualification critical serves to distinguish this form of realism from others, sometimes called direct or naive realisms, which assume simple “one-to-one” links between beliefs and reality. In the human social sciences, the qualifier critical also often signals a normatively critical orientation to existing social forms.

Critical realists make a clear distinction between the independently existing real beings, relations, processes, and so on (the “intransitive dimension”), which are the objects of scientific knowledge, and the socioculturally produced concepts, knowledge claims, and methods through which we attempt to understand them. The latter (sometimes referred to as the transitive dimension) (Bhaskar, 1979/1998) are always provisional and fallible. Critical realists thus fully acknowledge the socially and historically “constructed” character of scientific theoretical frameworks and knowledge claims, but they argue that practices such as experiment, scientific education, and the application of scientific ideas presuppose the independent existence of the objects of scientific knowledge. We cannot make sense of these practices other than on the basis that they are about something, even though we may doubt the truth of specific assertions about that “something.”

In most cases, critical realists reject the relativism often associated with approaches (such as their own), which emphasize the socially constructed character of knowledge. Without commitment to unsustainable notions of “ultimate” or “absolute” truth, it is usually possible to evaluate rival knowledge claims as (provisionally) more or less rationally defensible in light of existing evidence.

Critical realists have provided powerful arguments against both empiricist and relativist accounts of the nature of natural scientific activity. Bhaskar (1975/1997), in particular, used transcendental arguments from practices such as scientific experiment to show the incoherence of empiricist restrictions on scientific ONTOLOGY and accounts of scientific laws. Experimental practice is designed to isolate causal mechanisms, so producing regular event sequences, or “constant conjunctions.” But because these constant conjunctions are the outcome of experimental manipulation, they cannot be identical with scientific laws (otherwise, we would have to say that laws of nature were products of human manipulation).

This analysis results in a view of the world as “stratified,” distinguishing between the “level” of observed events (the “empirical”) and that of the “actual” flow of events (most of which pass unobserved). A third level, misleadingly referred to as the “real” (because all levels are in some sense real), is that of the “causal mechanisms,” whose tendencies are captured by statements of scientific laws. Critical realism might reasonably be described as a form of “depth” realism, implying the existence of multiple layers of reality behind or below the flow of sense-experience and subject to discovery by scientific experiment and theoretical analysis. Science is, in this way, seen as a process of discovery of ever deeper levels of reality (genes, molecules, atoms, subatomic particles, fields of force, quasars, etc.) rather than as one of accumulating experiential generalizations. The analysis of practices of applying scientific knowledge in technology yields a further claim about the nature of the world as an object of scientific knowledge. This is its “differentiated” character. Experimental practice involves isolating causal mechanisms from extraneous influences (i.e., the creation of “closed [Page 222]systems”). However, technologies involve application of scientific knowledge in “open systems,” involving the interaction of multiple causal mechanisms. In open systems, interaction between mechanisms results in the overriding or modification of their effects, such that regular or predictable flows of events and observations may not occur.

...