Critical realism is a philosophy of science that says the world exists independently of our ability to observe or understand it, and that the job of research is to uncover the hidden mechanisms that cause the events we see. Developed by philosopher Roy Bhaskar in his 1975 book A Realist Theory of Science, it offers a middle path between two dominant camps in research: one that treats only measurable data as real, and another that treats reality as entirely shaped by human perspective. Critical realism says both camps get something right, but both miss something important.
The Core Idea: Reality and Knowledge Are Separate
The foundation of critical realism rests on a single, firm distinction: what exists in the world (ontology) is not the same thing as what we can know about the world (epistemology). Bhaskar argued that collapsing these two things together is a mistake he called the “epistemic fallacy,” where we reduce what’s real to what we can observe or measure. A virus existed before anyone had a microscope to see it. Gravity operated before Newton described it. Critical realism insists that reality doesn’t wait for human awareness.
This leads to two types of knowledge. “Intransitive” knowledge refers to the objects and processes that exist regardless of whether anyone studies them. The structure of a cell, the dynamics of an economy, the pull of a magnetic field: these carry on whether or not a scientist is in the room. “Transitive” knowledge, on the other hand, is the theories, models, and data we produce to try to understand those things. Transitive knowledge is provisional and changeable. Our theories improve, get revised, sometimes get thrown out. But the underlying reality they point toward doesn’t change just because our understanding does.
Three Layers of Reality
Critical realism divides reality into three nested domains, and understanding these layers is key to grasping the whole framework.
The empirical domain is the smallest layer. It contains everything we directly experience or observe: the data we collect, the events we witness, the patterns we can measure. This is where most conventional science focuses its attention.
The actual domain is broader. It includes everything that exists or happens, whether or not anyone is around to observe it. A tree falls in a forest with no one nearby. A chemical reaction occurs in the deep ocean. These events are real and actual, but they sit outside the empirical domain because no one recorded them.
The real domain is the broadest and deepest layer. It encompasses the other two domains and adds something else: the underlying structures and mechanisms that produce events in the first place. These mechanisms are often invisible. You can’t directly see market forces, systemic racism, or the molecular interactions that cause a disease. But you can observe their effects and infer that they exist. Critical realists consider these mechanisms just as real as anything you can touch.
A common analogy is an iceberg. The tip above the waterline is the empirical domain: what you can see. The full iceberg, including everything below the surface, is the actual domain. And the ocean currents, temperature gradients, and geological processes that shaped the iceberg in the first place represent the real domain.
Generative Mechanisms Over Simple Cause and Effect
Traditional scientific thinking often looks for straightforward cause-and-effect relationships. If A happens, then B follows. Critical realism pushes deeper. Instead of asking “does X cause Y,” it asks “what is the underlying mechanism that produces this outcome, and under what conditions does it activate?”
These are called generative mechanisms. A generative mechanism is something that exists beyond what we can directly observe but explains why observable events happen. Think of it this way: poverty doesn’t cause poor health through a single, clean causal chain. Multiple mechanisms are at work simultaneously, including limited access to nutrition, chronic stress responses, reduced access to healthcare, and environmental exposures. Each of these mechanisms is real, even though none of them shows up as a single variable in a spreadsheet. They interact with each other and with local conditions to produce outcomes that vary from place to place and person to person.
This is why critical realists are skeptical of universal laws in the social sciences. The same mechanism can produce different outcomes depending on context, and the same outcome can be produced by different mechanisms. Research, from this perspective, is about identifying which mechanisms are at work and how the surrounding conditions shape what actually happens.
How It Differs From Positivism and Interpretivism
Critical realism is often taught alongside two other major research philosophies, and it helps to see where they diverge.
Positivism treats reality as objective and directly measurable. It prizes quantitative methods like experiments and surveys, seeks universal causal relationships, and aims for results that can be generalized across settings. If you can’t measure it, positivism has little use for it. The limitation, according to critical realists, is that positivism stays trapped in the empirical domain. It only counts what’s observable and misses the deeper mechanisms generating those observations.
Interpretivism (sometimes called social constructivism) swings the other way. It holds that reality is subjective, built through human interaction and meaning-making. It favors qualitative methods like interviews and ethnography, and it values rich, contextual understanding over broad generalizations. The limitation here, from a critical realist view, is that it can deny the existence of real structures that operate whether or not people are aware of them. Systemic inequality, for instance, doesn’t stop being real just because individuals interpret their experiences differently.
Critical realism borrows from both. It agrees with positivism that an external reality exists independent of human thought. It agrees with interpretivism that our knowledge of that reality is always filtered through perception, culture, and context. But it insists that acknowledging the limits of our knowledge doesn’t mean reality itself is unknowable or that it’s merely a social construction. Research can use both quantitative and qualitative methods, mixing approaches to get closer to the mechanisms that drive outcomes.
Structure, Agency, and Social Science
One of critical realism’s most influential contributions is how it handles the relationship between individuals and the social structures they live within. This is a classic tension in sociology and related fields: do people shape society, or does society shape people?
Critical realism says both, but not at the same time and not in the same way. Social structures like economic systems, legal frameworks, and cultural norms exist before any individual encounters them. They constrain and enable what people can do. But people also act on, reproduce, and sometimes transform those structures through their choices and actions. The philosopher Margaret Archer developed this idea extensively, arguing that structure and agency operate on different timescales and should be analyzed separately rather than blended together. Structures pre-exist the actions that eventually reshape them.
This matters for research because it means you can’t fully explain social outcomes by looking only at individual behavior, and you can’t fully explain them by looking only at structural conditions. You need both levels, plus an account of how they interact over time.
Retroduction as a Research Strategy
If the goal of research is to identify hidden mechanisms, you need a way of reasoning that goes beyond standard approaches. Critical realism uses a form of inference called retroduction. Where deduction moves from general principles to specific predictions, and induction moves from specific observations to general patterns, retroduction asks a different kind of question: “What must be true for this event to be possible?”
In practice, a researcher observes a pattern or outcome and then works backward, proposing candidate mechanisms that could plausibly have produced it. These proposals are then tested, refined, or eliminated through further investigation. It’s a process of creative reasoning followed by empirical checking, and it often requires combining different types of evidence, including statistical data, interviews, historical records, and case studies.
Where Critical Realism Gets Used
Critical realism has gained traction across a wide range of disciplines, particularly where researchers deal with complex, layered problems that resist simple measurement.
In health research, it provides a framework for evaluating complex interventions. A public health program, for example, rarely works the same way in every community. Critical realism helps researchers understand why an intervention succeeds in one setting and fails in another by identifying the mechanisms the intervention is supposed to activate and the contextual factors that help or hinder that activation. Researchers in nursing and public health have argued that this approach is especially valuable for frontline services trying to apply evidence-based interventions in the messy reality of clinical practice.
In ecological economics, scholars like Clive Spash have promoted critical realism as a philosophical foundation for addressing environmental sustainability and social justice. His 2024 book Foundations of Social Ecological Economics argues for critical realism as a basis for moving beyond conventional economic thinking. This application is not without controversy. Critics point out that critical realism developed entirely within the social sciences and may not adequately account for natural realities and technological systems. The debate is active and ongoing.
The framework also appears in education, international relations, management studies, and the study of religion. A 2024 special issue of Method and Theory in the Study of Religion debated how critical realist and discourse-based approaches relate to each other when analyzing social phenomena like religion, reflecting the kind of cross-paradigm conversation critical realism frequently generates.
Common Criticisms
Critical realism is not universally accepted, and several objections come up regularly. The most common is that its central claim, that unobservable mechanisms are real, is difficult to verify by definition. If a mechanism can’t be directly observed, how do you confirm it exists rather than simply fitting a convenient explanation to the data after the fact?
Others argue that the framework is better at describing what research should do than at providing concrete tools for doing it. Retroduction, for instance, gives researchers a logical structure but doesn’t prescribe specific methods, which can leave newcomers unsure how to apply it in practice. Some scholars in the natural sciences and interdisciplinary fields question whether a philosophy built primarily around social structures translates well to domains governed by physical and biological processes.
Supporters counter that critical realism was never meant to replace specific methods. It operates as what Bhaskar called an “underlabourer” for science: a philosophical foundation that clarifies what researchers are actually doing when they investigate the world, helping them choose appropriate methods rather than dictating which ones to use.