
Roy Bhaskar’s A Realist Theory of Science argues that experiment presupposes a world of real structures and generative mechanisms that exist independently of our knowledge. The book is philosophy, and its evidence is conceptual: it tests accounts of science against what experimental practice assumes. The critique of empiricism and idealism is the strongest part. The positive account of natural necessity is sketched rather than worked out.
The book is written for philosophers of science and for researchers who want a philosophical basis for realist social science. It does not engage with particular scientific findings, laboratory practice or the history of any discipline. The argument is programmatic: it sets out a framework and leaves detailed application for later work.
The 2008 Routledge edition reprints the 1975 text with a new introduction that places the book within the critical realist programme. The text of the original is unchanged.
The thesis. The possibility of experiment presupposes a realist ontology. If causal mechanisms did not exist independently of the events they produce, experiments would have nothing to isolate and nothing to reveal.
The critique. Empiricism reduces causal laws to constant conjunctions of events and cannot explain why experiments are necessary. If laws were only patterns of events, scientists could observe them without intervention. The need for experiment shows that mechanisms operate independently of their actualization.
The task. The book sets out to replace both the empiricist account and the transcendental idealist account, which locates the structures of the world in the mind. Bhaskar calls the confusion of what exists with what we can know the epistemic fallacy.
Key point: The book offers no method for doing science and no empirical findings. It offers a philosophical account of what science must assume.
What kind of argument does Bhaskar use to derive the realist ontology?
A transcendental argument: he asks what must be true for experiment to be possible. (Middle chapters.)
What are the three levels of the stratified ontology?
The real, the actual and the empirical. (Later chapters.)
Two stages. The first is critical: Bhaskar shows that neither empiricism nor idealism can explain experiment. The second is constructive: he derives the realist ontology that experiment presupposes.
The method. Bhaskar asks what must be true for experiment to be possible and argues that the answer is a world of real, structured, generative mechanisms. The conclusion is presented as a condition of the possibility of experiment, not as an empirical generalization.
The order. The chapters follow this order: critique of empiricism, critique of idealism, transcendental deduction of the realist ontology, then natural necessity and the stratified ontology.
The argument is cumulative. Each critique sets up the constructive move, so the reader sees why the realist ontology is needed before it is presented.
The prose is dense and assumes familiarity with Hume and Kant. A reader without that background will need to work slowly.
What is the first stage of the argument?
The critique of empiricism and idealism. (Early chapters.)
What method does Bhaskar use to derive the realist ontology?
Transcendental deduction from the possibility of experiment. (Middle chapters.)
Conceptual evidence. Bhaskar does not present data, case studies or historical material. He argues that experiment, as practised, presupposes a realist ontology.
The key example. Scientists intervene to isolate a mechanism from conditions that would prevent it from producing its effect. If laws were just constant conjunctions, intervention would be unnecessary.
The deduction. Bhaskar claims that the realist ontology is a condition of the possibility of experiment, not an empirical generalization from experiments.
The critique of empiricism is the strongest part. The argument that constant conjunctions cannot explain the need for experiment is hard to answer.
The positive account of natural necessity is thinner. Bhaskar asserts that necessity is real but does not develop the account in detail.
What practice does the argument use as its key example?
Experiment. (Middle chapters.)
What does empiricism reduce causal laws to?
Constant conjunctions of events. (Early chapters.)
The contribution. The book offers a coherent alternative to both empiricism and idealism and gives realism a transcendental argument rather than an empirical one.
For social science. The book provides a philosophical basis for treating social structures as real. Later critical realist work in sociology, economics and organisation studies draws on this account.
The book gives realism a transcendental argument and introduces the epistemic fallacy and the stratified ontology.
The book does not offer practical guidance. Its value is indirect: it supports treating social structures as real and causally effective.
BSD’s reading, not a claim the author makes.
What name does Bhaskar give to the position the book defends?
Transcendental realism. (Throughout.)
Which movement is the book the founding text of?
Critical realism. (Introduction to the 2008 edition.)
The main gap. Bhaskar argues that necessity is a real feature of the world, not just a constant conjunction, but he does not give a detailed theory of what necessity is or how we come to know it.
The programmatic limit. The book does not apply the framework to any particular science and does not show how the stratified ontology would guide empirical research. The reader is asked to accept the framework on the strength of the transcendental argument alone.
No empirical cases. The argument is entirely conceptual. A reader who wants to see the framework at work will not find it here.
The transcendental argument assumes experiment is the paradigm of science. If experiment is not central, the argument may not apply. Bhaskar does not address this.
Which part of the argument is least developed?
The account of natural necessity. (Later chapters.)
What kind of evidence does the book not use?
Empirical cases or data. (Throughout.)
The book is a strong critique of empiricism and idealism and a clear statement of a realist alternative. Its weakness is the sketchiness of the positive account. It is a founding text, not a finished system.
Density is the cost of the book’s choices. The argument is compressed, and the reader who does not know Hume and Kant will struggle.
Its strongest part is the critique of empiricism. Its weakest is the account of natural necessity, which is asserted more than demonstrated.
Philosophers of science, who will find a rigorous alternative to empiricism.
Social scientists, who want a philosophical basis for realist research.
Students, who need to know where critical realism started.
Practising scientists: the book offers no methodology.
Readers new to philosophy of science: a more accessible introduction is needed first.
Anyone wanting empirical applications: the book provides none.
Bottom line: A foundational work of critical realism that makes a strong case for realism and a thinner one for its own positive account. Read it for the critique and the framework, not for a finished theory of necessity.
What does Bhaskar call the reduction of being to knowledge?
The epistemic fallacy. (Early chapters.)
Which edition adds a new introduction?
The 2008 Routledge edition. (Front matter.)