specific cases only in accordance with properties of the general system (Hempel & Oppenheim, 1948). There may be one true knowledge system, or there may be many such systems, but in either case, coherency exists between specific facts and general propositions. Like correspondence theory, coherence theory is a traditional model of truth and knowledge that is closely associated with scientific positivism.
For text mining research, coherence theory implies that while there may be multiple ways to analyze online groups and communities, to be considered true the specific facts uncovered by text mining and analysis methods must fit within a coherent system of theoretical and empirical propositions.
Pragmatism
In pragmatist approaches to the philosophy of social science, truth is defined as those tenets that prove useful to the believer or user, and truth and knowledge are verified through experience and practice (James, 1907/1975, 1909/1975). In a pragmatist ontology, objective truth cannot exist because truth needs to relate to groups’ practices, and truth, therefore, always has both subjective and objective dimensions (Howell, 2013, p. 132).
A pragmatist approach to text mining treats as true and valid those specific facts uncovered by text mining methods that relate not only to objective reality or to a theoretical system but to the practices and interests of researchers themselves, be they members of academic communities, government agencies, or for-profit companies.
Constructionism
The constructionist position in philosophy of social science is generally traced to Berger and Luckmann’s 1966 sociological classic The Social Construction of Reality. Constructionism considers that no external objective reality or system exists; therefore, truth is not a matter of concepts’ correspondence to objective reality or coherence between specific cases and a system of propositions. Instead, truth and knowledge are produced when communities and groups reach consensus about what is or is not true. The core assumption of the constructionist paradigm is thus that socially constructed realities are not independent of the observer but are constructed by people under the influence of a variety of social and cultural factors that lead to intersubjective agreement about what is real (Howell, 2013, p. 90).
As in pragmatism, constructionism implies that text mining methods cannot eliminate subjectivity. But constructionism goes further in implying that objectivity is not a realistic goal of social research. Instead, researchers should aim to provide new interpretations of social phenomena that are edifying or enlightening. But they are not justified in claiming that their methods give them privileged access to an objective reality that cannot be otherwise apprehended. Accordingly, constructionism is a foundational philosophical position in postpositivist social science.
Critical Realism
Critical realism, as pioneered by Bhaskar (1975/2008), combines the realism of correspondence theory with the sociocultural reflexivity required by constructionism. Critical realism involves a distinction between the production of knowledge by human beings and knowledge that is of things or grounded within things. In critical realism, some objects are understood to be more socially constructed (or more “transitive”) than others. The critical realist ontology perceives reality as external to humanity but “considers our intellectual capacities to be unable to fully understand opaque and confounding truth” (Howell, 2013, pp. 50–51).
For text mining research, critical realism implies that because documents produced by social groups and communities are forms of knowledge produced by human beings, they are socially constructed and our understanding of them is always partial and limited at best. Yet unlike constructionism, critical realism also implies that there are elements of texts that are “intransitive,” meaning they can be treated as objective facts that are measurable and amenable to scientific research.
Metatheory
Along with epistemology and ontology, metatheory is an area of philosophy that has special relevance to the practice of social science research. Metatheory involves searching for answers to questions about the role of theory within social science research projects. Different approaches to social science research are generally based on one or another of the epistemological and ontological positions outlined previously. Partly stemming from their particular philosophical premises, these approaches to research end up using theory in strikingly different ways. Put another way, we can say that different approaches to text mining and text analysis are based on different metatheoretical premises, which are premises about the uses, capacities, and limitations of theory within empirical research. Howell (2013) provided the following list of metatheoretical positions commonly found in social science methodologies, ordered from the most abstract to the most concrete types of theorizing.
Grand Theory and Philosophical Positions
Grand theory refers to sweeping abstract explanations of social phenomena and human existence (e.g., Marxian historical materialism). Regardless of the method used, individual research projects can rarely directly test the claims of grand theories and philosophical positions. But over time, the findings from large numbers of empirical research projects can generally support or weaken the standing of such theories and positions.
Meso Theory
Meso theory is less sweeping and abstract than grand theory and is more closely connected to the practice of empirical research. The term refers to theories of the “middle range” (Merton, 1949) that draw on empirically supported substantive theories and models. Meso theory is common in psychology and sociology, where researchers develop, test, and refine relatively narrow theories related to specific psychological and sociological phenomena, such as theories of cognitive biases or of gender discrimination in hiring.
Models
Theoretical models are simplified, often schematic representations of complex social phenomena. They are used in almost all empirical research, particularly in research that is done in a positivist mode of inquiry.
Substantive Theory
Substantive theory is theory derived from data analysis that involves rich conceptualizations of specific social and historical situations. While theorists and empirical researchers understand that the more parsimonious the theory, the easier it is to identify relations of cause and effect between independent and dependent variables, greater simplification often makes it more difficult to reach a full understanding or explanation of a phenomenon. Where theoretical models may at times be oversimplified, substantive theories are more complex and arrived at inductively through interpretive methods (e.g., archival research and ethnography).
Making Inferences
Social science research involves making inferences (drawing conclusions) about theories; about patterns in data and about the individuals and communities that are, ultimately, the source of the data used in research. Inferential logic involves thinking about how and why it is warranted to make inferences from data. Based on their analysis of collected data, researchers use specific forms of logic to make inferences about relationships among social phenomena and between social phenomena and theoretical propositions and generalizations. In the early stages of a project, a researcher may not know the sorts of inferences they will make or conclusions they will draw. But in the end, they inevitably use one or more of the following forms of inferential logic, and it is beneficial for you as a researcher
