The "realism" part of the label is fairly straightforward. Bhaskar maintains that the social sciences (sometimes, often, once in a while) succeed in discovering and describing the real properties and causal powers of social structures and systems. Social entities have real causal powers, and sociology can discover the details of these powers. The approach is anti-positivist, anti-covering-law, and anti-reductionist.
So far this is the familiar position of scientific realism, applied to the social sciences. Rom Harré laid out a version of this in his causal realism theory (Causal Powers: Theory of Natural Necessity). If there is a controversial part of the theory, it is the attribution of reality to higher-level social structures like states, modes of production, and classes; but this isn't in fact very controversial.
This realist theme about knowledge of the social world is also familiar from the "causal mechanisms" approach to social explanation, where theorists argue that there are real (though often unobservable) social causal mechanisms that constitute the motive force of social change.
The more difficult problem is to say what "critical" means in this context. And surprisingly, neither Bhaskar nor his circle is very explicit about this question. The idea of "critical" realism does not appear at all in Bhaskar's first major book, A Realist Theory of Science (1975). The idea of critical philosophy is important and prominent in his second book, The Possibility of Naturalism: A philosophical critique of the contemporary human sciences (1979). But it isn't used to qualify "realism" but rather "naturalism."
Here is how Bhaskar introduces the idea of critical naturalism in the preface to the first edition of PN:
The upshot of the analysis is a new critical naturalism, entailing a transformational model of social activity and a causal theory of mind. The transformational model necessitates a relational conception of the subject-matter of sociology and a series of ontological, epistemological and relational limits on (or conditions for) a naturalistic science of society." (kl 145)When Bhaskar comes to qualify the "realism" of RTS later in his work, he uses the phrase "transcendental realism" to describe this formulation of his theory. The idea of "transcendental realism" is derived from Kant's Critique of Pure Reason, where a transcendental argument is introduced as one that seeks the conditions of the possibility of a certain kind of knowledge. What must be true of the social world and social actors in order that they may constitute the object of empirical knowledge? Bhaskar's specific question is this:
To what extent can society be studied in the same way as nature? (kl 180)Bhaskar and Tony Lawson explain this transcendental terminology in Critical Realism: Essential Readings:
Bhaskar sustains a metaphysical realism as a way of elaborating an account of what the world 'must' be like for those scientific practices accepted ex posteriori as successful, to have been possible. (3)This all gives a strong clue to the reader that Bhaskar's intentions are philosophical and ontological from the start; he deliberately chooses to adopt the language of Kant's critical philosophy of knowledge for his own study of the social sciences.
So, again, what might be implied by attaching "critical" to "realism"?
Critical thinking as emancipatory. In the Marxist tradition the word "critical" has a fairly specific meaning. This meaning is reflected in Marx's eleventh thesis on Feuerbach. "The philosophers have sought to understand the world; the point, however, is to change it." Critical science is engaged science, committed science, emancipatory science. Critical science is committed to constructing bodies of knowledge that have substantial impact on the link long term best interests of humanity.
Critique as illusion-destroying. Another dimension of the idea of criticism in the Marxist tradition is the idea of "critique" -- focused intellectual effort to uncover the implicit (and misleading) assumptions of various schemes of thought and policy. Marx's Capital is subtitled "A Critique of Political Economy", and this phrase is found in many other of his titles as well. This brings in the idea of laying bare the implicit (often dominating) assumptions of various systems of thought. Laying bare the partisan assumptions underlying ideology and false consciousness is an exercise of critique.
Critique as self-creation. Finally, there is a third connotation of "critical" that pertains to its use in the social sciences: the constant reminder that the social world is not independent and separate from "us". This involves the feature of "reflexiveness" that obtains in the social world. We constitute the social world, for better or worse. And the forms of knowing that we gain through the social sciences also give rise to forms of creating of new social forms -- again, for better or worse. So it is crucial to pay attention to the plasticity of the social relations in which we live, and the innovations we create in those relations through our own processes of knowing and doing. Margaret Archer refers to this fundamental aspect of the relationship between actors and the social world as "morphogenesis" (link).
I think each of these elements is involved in Bhaskar's evolving conception of "criticial" philosophy. In the preface to the Second Edition of PN Bhaskar makes most of the points highlighted above. He refers to the importance of critique of "philosophical ideologies," including positivism, where critique is understood in roughly the sense mentioned above. (An intended second volume of PN was planned but not completed, which would have been called Philosophical Ideologies.) And in the Preface to the First Edition of PN he makes reference to ideas originally expressed in the eleventh thesis on Feuerbach mentioned above, but this time quoted in Capital (part iv, section 10, p. 505), in explaining why sociology is important to epistemology: "Sociology is necessary if we are to avoid 'that kind of criticism which knows how to judge and condemn the present, but not how to comprehend it'" (kl 136).
The emancipatory character of Bhaskar's conception of the social sciences emerges as well in his critiques of the fact-value dichotomy in science. He rejects the idea that the scientist must remain ethically neutral with respect to the social and historical processes he or she studies.
But none of this amounts to a systematic exposition of what "critical" philosophy is. At most it gives the reader some clues about the features of thinking, reasoning, and acting that Bhaskar seems to have in mind when he advocates for critical realism as an approach to the philosophy of sociology.
So it seems that Bhaskar has chosen to allow connotation to replace analysis when it comes to explaining "critical". He is a careful and explicit philosopher in much of his writing; but on the subject of "critical" method, he is surprisingly elliptical. And to me, this suggests that the import of Bhaskar's system is more on the side of "realism" than its "critical" methodology.