Realism

Realism is a philosophy that understands the world as relational, causality as contingent, and knowledge as socially produced. It gained prominence in geography during the 1980s as an analytical framework for thinking about and “practicing” both social science (critical realism) and natural science (transcendental realism). This distinction between the social and natural sciences [Page 2370]is inherent to a realist separation between natural and social “kinds”: Society is recognized as an open system that can never replicate the closed conditions of a laboratory, and institutions and actors are understood to be in a constant state of flux as compared with the more intransitive physical phenomena of the natural world. This is consonant with a realist ontology, which posits the existence of a reality outside human consciousness that is independent of the social construction of nature. As an epistemology, realism asserts that we can actually know something about this world beyond our cognition of it.

To “know” something about the world in the realist sense is to explain phenomena—both physical and social—in terms of their causality. Realism, however, rejects the conflation of explanation with prediction under logical positivism, which explains phenomena in terms of the repeatability and replicability of experimental results. In contrast, realism maintains that explanation warrants an understanding of the structures that allow things—events and phenomena—to happen. Structures are the elements internal to a phenomenon that demonstrate, or have inherent susceptibilities toward, particular kinds of “behavior” or action. Causal powers are enacted and expressed through the mechanisms of a structure. The realization or activation of this autonomous force, however, depends on the circumstances or the context in which things occur.

Realism explains how things happen by analytically separating out these mechanisms of causality. There are two kinds of relations between phenomena and their generative causes: (1) necessary and (2) contingent. Necessary conditions are the internal properties of a relation, wherein the event or phenomena observed cannot “be” or occur as essentially that which “it is” in the absence of a particular relation. For instance, a “slave” can only be a slave in relation to a “master.” For a mechanism to be causal, it must be a necessary condition for a phenomenon. Contingent conditions are the external properties of a relation or coincidences of coexistence. They describe the possibility for things to happen in terms of both the properties of phenomena and the circumstances of context. Andrew Sayer provides the example of gunpowder: Gunpowder has certain properties, such as the ability to explode; explosion, however, depends on a spark to cause that explosion.

Mechanisms and structures are not always empirically discernible and must therefore be elucidated via rational abstraction. This holds in particular for causality in the social world, where the mechanisms that operate are both myriad and often unobserv-able. Abstraction involves the theoretical identification and conceptual isolation of the essential features or salient attributes of an event or phenomenon on the basis of its relations with, as opposed to similarities to, other entities. Abstraction is the basis for concrete research, which involves the “testing” of abstraction in the real world via empirical study to assess the “fit” or explanatory power of those abstractions. Empirical evidence may then lead to an adjustment of the abstraction such that the conceptualization yields a consistent picture of the world. This is an iterative process as it is through repeated abstraction—and grounding in empirics—that the difference between necessary and contingent conditions for causation can be elicited.

...