Ontology in critical realism takes priority over epistemology as implied by the identification of the ‘epistemic fallacy’. That is, it is the nature of the scientific object that should determine its proper epistemology, rather than vice versa.

The establishment of a philosophical ontology requires that we “follow the Kantian road” while at the same time rejecting “the idealist and individualistic cast into which Kant pressed his own inquiries” (PON3, 5). This is done through transcendental argument, which asks what must be the case for some (human) activity to be possible. Thus, while Bhaskar employs transcendental argument, he categorically rejects Kant’s transcendental idealism.

It should be emphasized that what Bhaskar is developing here is a philosophical ontology as opposed to a scientific one. As Bhaskar (RTS2, 39) puts it, a philosophical ontology “is developed by reflection upon what must be the case for science to be possible; and this is independent of any actual scientific knowledge.” A scientific ontology, on the other hand, can be said to be “the particular entities and processes postulated by some substantive scientific theory” (RTS2, 29-30).

The premises of Bhaskar’s transcendental argument are the generally recognized social activities of science, specifically the scientific experiment. Hence, he asks, what must the world be like in order for science to be possible and its activities intelligible? Bhaskar’s conclusion is that the world must be structured (distinguishing between the level of the empirical, the actual, and the real) and differentiated (distinguishing between open and closed systems).

Bhaskar establishes that reality is stratified requiring that we disambiguate the levels of the empirical, the actual, and the real. At the level of the real lie generative mechanisms defined as the powers and liabilities of things. “Such mechanisms endure even when not acting; and act in their normal way even when the consequents of the law-like statements they ground are, owing to the operation of intervening mechanisms or countervailing causes, unrealized” (RTS2, 46). Since generative mechanisms generally operate in open systems, they are rarely, if ever, actualized as a regular sequence or constant-conjunction of events, which may or may not be empirically experienced by humans.

Domain of the Real

Domain of Actual

Domain of Empirical










(Table from RTS2, 13)

Bhaskar’s analysis of experimental activity pinpoints the necessity of scientists being causal agents in the construction of a regularity of events, rather than being merely passive observers and recorders. That is, the objects of scientific investigation exist independently of the investigators. They are ‘intransitive’. There are two functions that scientists must carry out in any experiment. First, in ‘experimental production’, the scientist(s) “must trigger the mechanism under study to ensure that it is active” (RTS2, 53). Second, scientists must achieve ‘experimental closure’ in order to “prevent any interference with the operation of the mechanism” (RTS2, 53). Thus, it is only in the limiting case of a carefully constructed, closed scientific experiment that the level of the real becomes aligned with the level of events and human experience. These conclusions refer to the 'intransitive' objects (or dimension) of science.

Furthermore, in order for science to be possible, there must exist a 'transitive' dimension of science or 'philosophical sociology' (cf. RTS2, 195). These are 'Aristotelian material causes', the 'raw materials of science': "They include the antecendently established facts and theories, paradigms and models, methods and techniques of inquiry available to a particular scientific school or worker" (RTS2, 21). The transitive dimension refers to the social aspect of knowledge. It implies that that science never constructs its theories ex nihilo, but, rather, must work with the theoretical tools that are given at a certain point of scientific development. Thus, science must be “produced by the imaginative and disciplined work of men on what is given to them. But the instruments of the imagination are themselves provided by knowledge. Thus, knowledge is produced by means of knowledge. The objects from, and by which, knowledge is generated are thus always themselves social product (as is the knowledge generated)” (RTS2, 185).

By showing that the world is stratified and differentiated, as a condition for the possibility and intelligibility of scientific activity, Bhaskar is in the position to critique the empirical realism of positivism. In Humean empiricism, causal laws are identified as being a constant conjunction of events that are apprehended in sense experience. However, this assumes that the world consists of a closed system, the only conditions under which such a regularity could be actualized. Thus, the empirical realist is left with a dilemma. Either there are no laws of nature ('strong actualism') since a constant conjunction of events is never actualized outside (generally artificially constructed) closed systems, or the laws of nature only operate in closed systems ('weak actualism'), leaving one to wonder what governs phenomena in open systems (RTS2, 65, 92).

In contrast to empirical realism, transcendental realism identifies causal laws as operating at the level of the real, as the causal powers and liabilities of things. Thus, outside the conditions of closure, causal laws must be viewed as tendencies, “potentialities which may be exercised or as it were ‘in play’ without being realized or manifest in any particular outcome” (RTS2, 50). This means that law-like statements must be understood 'transfactually' (or 'normically'), in that the powers of the generative mechanism "may be possessed unexercised, exercised unrealized, and realized unperceived (or undetected) by [humans]" (RTS2, 184).