Tuesday, April 19, 2022

Realism about social entities

Critical realism depends on the key notion that sociologists are justified in construing their statements about social entities as real, objective features of the social world — “intransitive” objects, in Bhaskar’s somewhat idiosyncratic vocabulary. But does a realist ontology actually require this assumption? Or are there realist interpretations of sociological theory that do not “reify” entities like social structures, ideologies, or normative systems? Could we be realist about the social analogues of atoms and forces but nominalistic about larger ensembles like proteins? In chemistry there would be no reason to ask this question; but in the composition of the social world, it is possible that the linkages between "micro" and "macro" are sufficiently loose as to make it plausible that the ensembles have less permanence and fixity than their components. It is possible that the social world is more akin to a jigsaw puzzle in which all the pieces are square and can be fitted together in countless different ways; so there is no reason to attribute "existence" to the various combinations that occur.

This probably sounds needlessly paradoxical. But I think there are problems in asserting the independent, objective reality of a social structure like “the US system of academic tenure” that do not arise with respect to the building blocks of social structures and institutions like incentives, group priorities, property relations, and so on.

The ontological problem about large social entities arises from the open boundaries, multiple dimensions, and heterogeneity characteristic of the great preponderance of social “structures”. Take “tenure”: there are some important features commonly associated with tenure, like “protection of academic freedom,” “peer review,” and “self-governance”. But there is a great range of institutional arrangements, institutional priorities, and processes that make multiple instances substantially different from each other. Harvard’s tenure process looks very different from that practiced at the University of Massachusetts Lowell. And this is not a statement about elitism and status, but rather a reflection of different institutional priorities, needs, and the path-dependency relating to the historical composition of the respective arrangements.

We might be more confident in “realist” reference to a sociological entity when we refer to the particular instances of structures and ideologies, in a time and place, rather than the class of entities. So we might say that both Harvard and Lowell have particular tenure arrangements that are stable and well-defined and that can be selected as social entities. The class of all tenure systems, on the other hand, should be understood nominalistically as referring to a group of substantially different individual systems.

A more fundamental approach a realist might take is to abandon realist interpretations of large social things altogether and choose instead to refer realistically to the mechanisms, modular organizations, and interests and actions that make up the large macro structures and systems. This would represent a social realism about underlying processes, forces, and constraints, while adopting a nominalistic view of the higher-level entities. This orientation seems to point back to the view that I call "methodological localism": the view that "the 'molecule' of all social life is the socially constructed and socially situated individual, who lives, acts, and develops within a set of local social relationships, institutions, norms, and rules" (link, link). And it suggests that sociological realism is most clearly justified when applied to social circumstances at the micro- and meso-levels -- not the grand macro level.

This modest version of social realism preserves the key ideas of the objectivity, independence from the observer, and persistence and recurrence in the social world of the object of investigation. And this sustains the primary features of a realist understanding of social research and theorizing. On this more limited interpretation of realism, the social world exists independently from the researcher; causes, actions, incentives, and constraints exist, and social actors interact with each other in a variety of ways. Large structures, however, are too indefinite to count as "real" social entities. They have the shape-shifting status of the forest rather than the trees.

And what about the large structures that many of the critical realists care about the most -- capitalism, mode of production, economic structure, forces of production, bourgeois ideology, ...? It seems reasonable to construe these social things as conceptual constructs or ideal types rather than ontological entities. The "capitalist mode of production" is an intellectual construct, not an independent social reality. It is composed of real social actions, institutions, arrangements, and practices, all of which can be empirically investigated and which are independent from the observer. But when we think of the CMP along the lines of Marx, Engels, Lenin, Althusser, Deng Xiaoping, or Hayek, we are thinking of substantially different ways of conceptualizing the social and economic world. Further, the capitalisms in Britain, Germany, Japan, and post-Communist Poland all have important differences when it comes to their fundamental institutions and dynamics. The CMP is not a single abiding and theory-independent social reality.

Does the "economic structure of the United States" exist? It does, in that there are specific and empirically accessible institutions, relations, actions, and organizations that hang together and persist, and which we intellectually organize as the "economic structure". And yet it does not exist, if by "exist" we mean something solid, unchanging, and specific. As Marx once wrote, "all that is solid melts into air" (link).

(An earlier post considers a different kind of retreat from ambitious realism in the natural sciences, structural realism (link).)


jim said...

I think it’s a common sense take, but at different time scales even atoms and molecules start to look like forests rather than trees. If you can account for some of the dynamics of capitalism independent of location or time, why not say it’s real. I realize that this may be a structural realist argument. To be is to be real pattern, but something is worrisome about the intellectual construct position because in a very real sense all of science is an oversimplification.

Yufan Sun said...

It seems that your comment on realism has changed from your early book, Understanding Peasant China, to your recent books, New Directions in the Philosophy of Social Science and New Contributions to the Philosophy of History. Could you describe your intellectual trajectory? Thank you.

Paul D. Van Pelt said...

Real objective features of the social world change. This is a foundation of my thesis on contextual reality---a work-in-progress. The text of a current draft discusses the indistinguishability of reality and truth while pointing out the emergence/subsidence of notions about what is real and true under changing circumstances and contingencies.
Our penchant for mass/popular culture and following trends are examples. The word 'critical' has reached critical mass in our modern lexicon, seems to me. The resurgent political upheaval over critical race theory is an example of this---all in consideration of the origins of the phrase: activists for racial equality, Sojourner Truth; W.E.B. DuBois; and Frederick Douglass. Real people who defend what they deem a racial status quo do not want to hear of critical race theory. I am not, uh, critical of your fine writing. These are only one person's observations.

Dan Little said...

Jim, thanks for your reflections about this topic. I'm not sure I agree about the portability of this "agnosticism" to the natural sciences. A protein model is a complex chemical structure that can be investigated through a variety of research techniques in physics and chemistry; and it is reasonably stable over medium periods of time. There is the interesting complication that a protein can sometimes "fold" in multiple ways, giving it different chemical/physical properties. But even these variants are stable and explicable. I suppose that a critical realist might say that the position I'm trying out here gives up too much "social ontology" and places the ontological action at the level of actors and their relations; so it is too close to methodological individualism for their comfort. But I feel this position preserves "social realism" in a very specific sense: social facts exist independent of the mental frameworks of the investigator (which is the criterion of Bhaskar's distinction between transitive and intransitive entities).

Dan Little said...

Yufan, thank you for your thoughtful comment and question. The immediate impulse towards the thinking in this post came out of the seminar I'm teaching now, along with some puzzles about Bhaskar's assumptions. The basic question in my mind is this: what are the minimal commitments that a "social realist" must make about the social world? Does realism require commitment to a wide range of social entities at a variety of levels, or is it enough to be realist about the circumstances, mechanisms, actions, and social relations of the social actors who make up the social world? What is the line between "transitive" and "intransitive" things, in Bhaskar's terms? If X's are real, and Y's are composed of X's, does that imply that Y's are real existing entities as well? Consider the ontology of evolutionary biology. We have the mechanisms of natural selection, genotype and phenotypes of organisms, competition for reproduction within an environment, ... But are we justified in asserting the reality of the local ecology in which evolutionary processes unfold? Or is the idea of an ecology simply an umbrella term for encompassing an open-ended set of environmental conditions and resources in an open-ended region of space? Is the idea of the "ecology" in which the finches of the Galapagos just a convenient way for biologists to refer to more concrete items of interest in the evolutionary process? When we theorize about local ecologies, we have fairly specific ideas in mind about the systemic interconnections that exist among the species, food sources, climate conditions, vegetation, disease vectors, ... , that are co-present and that jointly influence the reproductive success of variant individuals; but notice how open-ended that list is. So I would be inclined to argue that we shouldn't reify the idea of an ecology, but rather keep in mind always the component processes and mechanisms that it encompasses.