Should advanced practitioners consider critical realism as a distinctive epistemological position for social scientific enquiry?

What is ontology?

Ontology is concerned with the theory of being, identifying and classifying things that exist (Benton and Craib, 2010).

What is epistemology?

Epistemology is the theory of knowledge, focusing on issues such as what constitutes a knowledge claim, how knowledge can be acquired or produced, the extent to which transferability can be assessed, and the concepts of validity and generalisability (Benton and Craib, 2010).

What is science?

Before discussing social scientific enquiry, a brief consideration as to what constitutes science is required. Science has several dimensions: as a body of knowledge, as a practice of institutions, as a systematic method, as a profession. Science makes knowledge claims through a systematic process of enquiry. The SVoS relies on objectivism, positivism and empiricism to study the natural world: knowledge gained from senses and interactions that can only be generated from observable objects and tested by observations and experiments by using induction and deduction processes. Science produces and creates knowledge that constitutes reality and testable knowledge: truth that is separate to theories and values (Benton and Craib, 2010; Chalmers, 2013). The SVoS will be used to develop the argument that critical realism is a distinctive epistemological position for social scientific enquiry.

Although the SVoS has been widely critiqued by philosophers since Kuhn's (1962) initial challenge, which considered the history of science and introduced the concept of paradigms to explain how science functions, Kuhn argued that ‘normal’ science operates within a dominant paradigm. This paradigm persists until a crisis arises, prompting the development of an alternative paradigm to resolve the issue in a process of scientific revolution. Science conducted within the new paradigm, according to Kuhn, is ‘incommensurable’ with science conducted under the old paradigm (Benton and Craib, 2010). However, Kuhn was challenged and the discourse around ‘what is science’ continued with scholars such as Karl Popper, with the key argument that science should be based on hypotheses that are subject to falsification by experimentation. This approach, known as hypothetico-deductive reasoning, is exemplified in methodologies like randomised controlled trials (Benton and Craib, 2010).

This position was further developed by Imre Lakatos, who argued that the SVoS failed to account for the historical and social context of scientific research (Benton and Craib 2010; Chalmers 2013). He postulated that the focus should not be on falsification, but on evaluating scientific theories based on their ability to solve problems and make predictions. He argued that consideration should be given to research programmes rather than individual theories, with studies strengthening or weakening programmes of research. He acknowledged that observations may produce grey areas, where the findings need evaluation and lead to probabilities, rather than yes or no results, bringing in an element of interpretivism (Benton and Craib, 2010; Chalmers, 2013). This debate continues, with philosophers such as Fleck, Planck and Feyerabend, who have different perspectives on ‘what is science, how should it be done, is science rational, and is there scientific progress’ (Benton and Craib, 2010).

However, the SVoS, with its focus on objectivity and its emphasis on causal links and predictions, remains embedded in research institutions, culture and practice. Yet, it does not address the social process of scientific inquiry—particularly the evaluation of research data carried out by researchers with contrasting worldviews and theoretical assumptions.

What is social scientific enquiry?

Sociology, positivism, and social scientific inquiry stem from the work of Auguste Comte in the 19th century. Comte argued that society progresses through three stages: the theological, the metaphysical (scientific revolution) and the positive. In the positive stage, explanations are confined to verifiable and measurable correlations between phenomena, rather than seeking causal relationships. This approach focuses on identifying the rules that govern these relationships, replacing absolute concepts with relative ones (Comte, 1865). He classified sciences and argued that only maths and astronomy had achieved a positive status, while other sciences (sociology, biology, chemistry and physics) retained theological and metaphysical thinking. He postulated that sociology was the key and final science as a strategy for reorganising society and promoting all sciences (Comte, 1865).

However, scientific knowledge cannot be generated solely through traditional positivist methods and cannot be universal, objective and detached from the socio-political context. This was argued by philosophers such as Neurath, who took an anti-foundational view of science in that knowledge can shape the world, but only by gradual reconstruction (Jordi, 2023).

He argued that verification and falsification are irrational, and that physical, psychological and social factors must be considered when analysing issues (Jordi, 2023). Positivism has also been challenged by those sceptical of the universality of science, for example feminists and post-colonialists, such as Mignolo and Quijano, who argue that colonialism shaped, and continues to shape, modern science (Benton and Craib, 2010; Mignolo and Walsh, 2018).

Social phenomena are not durable and are subject to change (unless intentionally maintained), and they are influenced by the actions of individuals as sentient beings, whose decisions and behaviors are central to the study of society. This leads to the position that the study of the social world is contextual and involves a series of interpretations to create meaning and understanding, along with the expressions of those meanings (Taylor, 1971). Within interpretivism, hermeneutics, the formal study of appropriate methods of interpretation, is a process of complex interactions between the researcher and the interpreted object and has been described as intersubjective. The process of interpretation may have no end, as described by the hermeneutic circle (Benton and Craib, 2010). This indicates that interpretivism involves a degree of uncertainty, as meanings are continually subject to reinterpretation and revision. Taylor explained this as:

Given that the SVoS cannot offer a paradigm that enables investigation of society and the social sciences, alternative paradigms must be considered for social scientific enquiry.

The intransitive and transitive dimensions

From an ontological perspective, the critical realist position believes that there is a real universe out there, at least in part independent of human minds (externalism and theory-independence). Bhaskar categorised this as the intransitive dimension (Vandenberghe, 2009). Critical realism aims to identify social practices that are ontologically real and may be causal. From an epistemological perspective, critical realism takes a relativist position, arguing that the nature of reality is relative to the perspective of the researcher (Benton and Craib, 2010), categorised as the transitive dimension (internalism and socio-historical variability of scientific knowledge) by Bhaskar (Vandenberghe, 2009). The distinctions that critical realism makes between the intransitive and transitive dimensions avoids epistemic fallacy, which subsumes ontology into epistemology and defines being in terms of knowledge. The combination of ‘ontological realism, epistemological relativism and judgemental rationality is…uniquely compatible with a sociology of science’ (Vandenberghe, 2009).

In contrast to the SVoS, critical realism acknowledges that the world is real, including both objects and social structures. It recognises that while knowledge production is fallible and theory-dependent, it is not entirely determined by theory. Critical realism also emphasises that while meaning and discourse are important, there is more to reality than just these constructs (Benton, 1981).

Bhaskar argued that the laws discovered through experiments represent tendencies of underlying mechanisms, which researchers within the SVoS paradigm tend to overlook because of their focus on the empirical level. Therefore, positivism has significant limits and is eclipsed by critical realism, which offers an alternative robust ontology for the natural sciences (Vandenberghe, 2009). His view of laws as tendencies of underlying mechanisms is a key realist argument. Bhaskar went further in his analysis of the intransitive dimension (independently existing objects of knowledge); he postulated that reality is stratified into the three levels explained in Table 1 (Benton and Craib, 2010). The key distinguishing feature of critical realism is its transcendental claim that an independent reality exists, composed of underlying causal mechanisms and powers. These mechanisms and powers require research to uncover reality, using logic and retroductive reasoning. While recognising that causal mechanisms operate as tendencies rather than producing universal laws, critical realism asserts that they give rise to a variety of effects that can vary across different contexts. When two mechanisms converge and produce new social phenomena, this is referred to as emergence. As a result, knowledge should not be viewed as the absloute truth, but rather as an ongoing journey (Benton and Craib, 2010).

The position holds that there are real ontological features of the world that are relational and exist at different levels, and the world is emergent and stratified (Archer et al, 1998). There is some linguistic confusion in using the term ‘real’ to refer to a level of reality, as well as in the sense that each of the levels is considered real (Benton and Craib, 2010). However, the pursuit of a deeper understanding distinguishes critical realism as a unique ontological and epistemological philosophy, as it aims to uncover more profound ‘truths’ beyond the surface-level observations of the SVoS (Figure 1).

Open and closed systems

A further distinction is made by critical realists between open and closed systems. Systems are described as open where mechanisms interact and coexist. Systems are described as closed when mechanisms exist in isolation, competing mechanisms are balanced, or there is artificial control of interfering mechanisms (such as in an experiment with controls for confounders) (Bhaskar, 1975). However, Bhaskar argued that attributing causal power to an object of study within a closed system is valid only if, when applied to the real world, that attribute continues to operate in the same way (Bhaskar, 1975). For example, when a drug is administered to a specific cohort as part of a randomised controlled trial (a closed system) and is found to have certain effects and a safety profile, the SVoS presumes that the causal mechanism will apply transfactually—it will operate in the same way when in general use. However, that may not be the case when administered to a non-controlled population (an open system). Bhaskar referred to this aspect of the intransitive dimension as differentiation, revealed through the analysis of scientific experiments and their broader applications. He argued that the world exists independently of our beliefs and is both differentiated and stratified (Benton and Craib, 2010).

In contrast to the SVoS position, critical realism recognises the social and historical nature of science in accordance with the positions of Kuhn (1962), Feyerabend and other constructionists (Benton and Craib 2010). It acknowledges that science is a social practice with institutions and culture, which can be drawn on to support the development of scientific knowledge. Kuhn referred to the social process of knowledge production as the transitive dimension. Thus, critical realism has a distinct theory of science (Benton and Craib, 2010) (Table 2). This position contrasts with the SVoS, which overlooks the social dimension of science, offers a limited view of reality based on surface-level observations, and treats scientific knowledge as a mere record of these observable facts. It also differs from the positions of Kuhn, Feyerabend and other constructionists, who recognised the social elements of science but did not offer a developed or rational account of the independent reality of observed phenomena.

Structure and agency

From Bhaskar's perspective, society and people are real and distinct, yet interdependent and interacting. This suggests that if a theory ignores structure or agency, a basic ontological feature of the world is ignored. While structures are causally necessary to enable actions, which would otherwise not be possible, they also constrain actions. Bhaskar developed a transformational model of social action, proposing that social structures are reproduced and sustained through the actions of social agents, whether intentional or unintentional. However, he argued that individual or collective agency—whether social or political—has the potential to modify or transform these structures (Benton and Craib, 2010). Although, some theorists argue that collective action is required to effect change in social structures to reduce inequality (Archer et al, 1998). So, critical realism has an emphasis on experimental practice based on assumptions that there is a relationship between structure and agency. This can be conceptualised through a three-step process (Table 3).

This is a persuasive argument against the SVoS position that dismisses the reality of social structures, although it does not consider the impact of non-human contributions to human social life. Nor does it consider the issue that many elements of social life are habitual and do not involve conscious thought or causal mechanisms that operate independently of agency, such as the causal mechanisms of poverty, social class and occupations, chronic illness and premature death (Benton and Craib, 2010).

However, the concept of interactivity between structure and agency enables scientific work as the beliefs social actors have about their social life are a resource for social scientific study, and provides a hermeneutic dimension that is a starting point for social science (Bhaskar, 1975). Although empirical evidence and theoretical arguments regarding social structures can offer contrasting perspectives to those derived from social actors (Benton and Craib, 2010), some argue that isolating the properties of specific structures is challenging. This is because of the interplay between actors and structures, as well as the reliance on knowledge of existing structures (Kemp and Holmwood, 2003). This led Collier (1994) to conclude that truly scientific social inquiry is not feasible and that structural hypotheses cannot be effectively tested. On the other hand, Sayer (2000) argued that abstraction and qualitative questioning can identify social world structural properties. This discourse continues as scholars and philosophers explore and debate the complexities of this issue.

Bhaskar's commitment to naturalism sits uncomfortably with his position on structure and agency, as it implies that there are ontological differences between society and nature which in turn affect epistemology. Bhaskar defended this critical naturalism on an ontological level. He argued that ontological limits related to differences between natural and social structures (Benton and Craib, 2010) (Table 4).

This challenges the distinction Bhaskar made between the intransitive and transitive dimensions in relation to social sciences. Collier (1994) argued that the epistemological limit within naturalism was the absence of the possibility of experimental closure in social sciences, in line with the SVoS. Collier (1994) and Benton (1981) suggested that Bhaskar had made too much of a contrast between social and natural reality. Benton (1981) further suggested that Bhaskar's ontological position on structure and agency did not align with his stance on the activity-dependence of social structures. Benton also argued that space-time dependency applied to some natural sciences, while concept-dependence could be misleading because many aspects of social life were mundane and routine. Benton's position was that Bhaskar's contrast between natural and social ontology was partly because of consideration of hard sciences such as chemistry and physics (the SVoS), rather than natural sciences that share features with the social sciences, such as meteorology, evolutionary biology and ecology. Benton pointed out that collaboration across the social and natural spheres is necessary for research programmes to address problems, such as the relationship between socioeconomic progress and ecological change.

However, Bhaskar (1975) argued that social science was possible because of the differences between the natural and the social spheres. He suggested that the fact that social science can study itself did not negate the transitive-intransitive dimension; it is possible to differentiate between what is being studied from the process of study. This leads to the argument that methodological reflexivity is present in social sciences and is a strength, whereas reflexivity is absent from the SVoS given its objective nature.

More challenging, however, is the issue of the epistemological limit. Some critical realists argue that this represents an insurmountable weakness, as highlighted by Bhaskar when he suggested that there are parallels between social phenomena and experiments in the natural sciences—such as when a crisis in social order exposes hidden structures. However, in general, socio-political crises tend to polarise interpretations of the social world, rather than leading to consensus. While transcendental arguments also offer an alternative to experiments through the analysis of social practices and accounts of social structures (Bhaskar, 1975), this appears to offer a more limited role for empiricism in social science. On the other hand, there is strong support for the use of statistical analysis in the social sciences (Benton and Craib, 2010).