Social Ontology De-dramatized

Richard Lauer deserves our thanks for bringing the question of social ontology to the center of debate in the philosophy of social science. Philosophers of social science are currently asking good questions about the nature of the social world. What kind of entities, powers, forces, and relations exist in the social realm? What kinds of relations tie them together? Are there real social groups? What are some of the mechanisms and causal powers that constitute the workings of these social entities? Are there distinctive levels of social organization and structure that can be identified? These are constructive questions that genuinely have the capacity to improve the effectiveness of social-science research and theory. Lauer’s article moves forward this discussion by suggesting that we need to do more “meta-ontological” thinking about the topic, and with this I agree.

I am happy to put myself in the “Ontology Matters!” (OM!) camp. I do believe that ontology matters for the conduct of scientific inquiry and theorizing, and especially so in the social sciences (Little 2016). I am very aware of the errors that scientists have made in their scientific work by adopting a misleading or wrong ontology for their work (radical behaviorism in psychology, vitalism in biology, operationalism in physics). However, I do not find myself in either of the camps that Lauer sketches on the map of the possible when it comes to locating the field of social ontology. I do not regard a social ontology as merely a pragmatic aid for inference, and I do not regard it as a pure field of apriori reasoning within philosophy that admits of certainty. I therefore propose a different way of approaching the field of ontology, an approach that fits well with the actual practices of empirical social scientists, sociological theorists, and philosophers when it comes to thinking concretely and specifically about the nature of the social world.

There appears to be something of a disconnect between the title of Lauer’s article and the arguments he offers. The title asks whether social ontology is prior to social-scientific methodology, but this is a question Lauer does not really address in the article. Instead, he considers two different meta-ontological views of the truth status of ontological theories—apriori realism and inference-supporting pragmatism. Fundamentally, he paraphrases the title’s question to a more modest question, “can answering ontological questions contribute to empirical success?” However, in my view, an affirmative answer to this latter question does not imply that ontology is prior to empirical science, since (as argued below) I believe that ontological theorizing is part of the extended scientific enterprise of understanding the world, and that efforts to grapple with empirical puzzles in the world are themselves helpful to refining and specifying our ontological ideas. It is not usefully approached from the point of view of pure apriori philosophical theorizing.

The approach to social ontology that I advocate rejects the idea that a theory of social ontology is a pure effort of apriori philosophy that can be justified on non-empirical grounds. If it is stipulated (as Lauer (2009), Searle (2010), and Epstein (2015) appear to do) that social ontology is that area of pure philosophical reasoning that concerns itself with the nature of social entities, and we then ask if social ontology is prior to empirical reasoning or social-scientific methodology, then my response is that there is no intellectual or practical content to the field, and it has no relationship to social-science methodology whatsoever. Kant demonstrated the sterility of pure speculative metaphysics, and his program for synthetic-apriori knowledge of the world failed. So there is no scope for purely philosophical discovery that is relevant for knowledge of the empirical world. Only if there is a creative back-and-forth between empirical investigation, theory formation, and ontological reflection will we gain knowledge and insight about the nature of the social world.

It is important to think carefully about the question, what is a theory of social ontology about? When we engage in ontological thinking, we are attempting to use what we currently know about the social world to arrive at some basic ideas about how the social world works and what its fundamental “constituents” are. The word “fundamental” is important here, and it is also somewhat misleading. It is important, because it gives a clue to the insight that there are many questions about “social entities” that are not ontological. “Is there such a thing as the Internal Revenue Service?” is a question about existence, but it is not really an ontological question. It is rather a question for mid-range theory in the sociology and politics of institutions. But “fundamental” is also misleading, if it is understood to imply a kind of “stuff-fundamentalism” for the social world, a view of social existence that postulates that there is a short list of social “things” of which all concrete social entities are composed.

The theory of social ontology I wish to defend is not foundationalist in either a metaphysical or an epistemic sense. It does not, like Democritus, search for the smallest or simplest elements of the social realm and assert that all social phenomena are molecules composed of these elements. The goal of a theory of social ontology, on my view, is not like Euclid’s theory of geometry—to arrive at a minimal conception of geometrical objects and a set of axioms defining their interactions, permitting derivation of all possible facts about space. I am confident that there is no finality or completeness in the field of social ontology. There is no hope for a final, conclusive, and complete theory of social reality that could serve as a basis for all social-science reasoning.

This anti-foundationalism in the area of metaphysics leads to skepticism about a phrase that Lauer uses frequently: “carving the social world at the joints.” In my view, this is a conception that we are better off abandoning. The social realm does not have “joints.” Is there a “joint” between the soccer team as group of intelligent actors (ontology of individual actors) and the same team as a fluid figuration (ontology of doings and relations)? Or are these rather just good alternative ways of analyzing the same set of social occurrences on the field? I am inclined to believe the latter. But if this situation is common—that there are alternative conceptions of the workings of the social world that can be seen to be different but reasonable descriptions of the same kinds of stuff that can be inter-translated (Davidson 1973-1974)—then that there is no such thing as “carving the social world at the joints.” The phrase suggests the possibility of a final, fundamental, and certain specification of the constituents of the social world and their relationships, and this is not possible. In this respect, I admire Nelson Goodman’s (1978) careful analysis of the relativity of ontological statements to conceptual frameworks.

Furthermore, the theory of social ontology I want to defend is not foundationalist in the other familiar sense either, the epistemological sense. I do not believe there is a field of pure philosophical reasoning within which philosophers can answer the question, “What exists in the social realm?” and where the answers have the status of substantive non-empirical truths. (Bhaskar [1975, 1989] believed something like this, but he was over-reaching.) Rather, like Quine, I believe that ontological theorizing is of a piece with other kinds of scientific theorizing about an unobservable world, albeit at a fairly abstract extreme of that activity. So Lauer’s realist version of social ontology fails, in my view, because it depends upon the idea that social ontology is a branch of apriori philosophy. Moreover, it assumes this is the only way we can make sense of realist ontological theorizing.

Does this mean that ontological theories are not veridical—that is, not truth-bearing? Not at all. Rather, ontological theories are advanced as substantive and true statements of some aspects of the social world, but they are put forward as being fundamentally a posteriori and corrigible. Like all scientific assertions, the world may turn out differently than our theories expect, and these differences may force changes in our scientific theories and our ontological theories. (Think of the irreparable damage done to the classical-physics ontology of the material world by the empirical discoveries of quantum mechanics.) James Coleman’s (1989) Foundations of Social Theory has had a great deal of influence in the social sciences since its appearance, and it is a good example of an intellectual construction that ranges from empirical investigation to social theory to social ontology and back. Coleman considers concrete instances of social behavior (e.g., the workings of bureaucracies; Coleman 1989, 79); he offers an explanation of how these patterns of behavior emerge from the interactions of rational actors in the setting of specific constraints and incentives; and he tries to account for the main areas of sociological interest in terms of this construction (norms, institutions, structures, groups). All of this is ranged under an ontological theory of the social world: the social world consists of elementary actors whose behaviors constitute more abiding social arrangements such as normative systems like a house of cards (Coleman 1989, 43-44, 503). Coleman’s book is important because of the clarity of its assumptions at each of these levels, from empirical problems to theoretical explanations to ontological assumptions about the nature of the social world. This very clarity permits useful and constructive debate about social ontology. It is open to sociologists and philosophers of social science to challenge Coleman’s ontological assumptions based on a different body of sociological and historical data and theorizing. The “new pragmatism” offered by Neil Gross embodies a different way of thinking about the social world, the social actor, and the social environment in which actors make their choices (Gross 2008), an environment in which career, field, and a “thick” theory of the actor play fundamental roles in the explanations offered. Coleman and Gross disagree about empirical matters, and they disagree in a less explicit way about the presuppositions they make about the way the social world works. These are substantive disagreements about social ontology, and they are corrigible and debatable.

So how do we gain epistemic warrant for ontological theories? Here too Quine’s philosophy of knowledge is helpful, in his formulation of coherence epistemology. Consider an example. If we assert that the social realm consists of social actions in Weber’s (1991) sense, this is an ontological claim. It motivates and explicates other more specific sociological ideas that we might put forward in our theories—capitalism, social knowledge, value system. It also corresponds to some “considered judgments” or observational convictions that we currently have about the social world—for example, we see individuals in work settings, we interpret their interactions with their colleagues, we observe unintended outcomes, and we make sense of the “office social system” in terms of the oriented actions of a number of social individuals engaged in a series of social actions. In other words, there is a fit between this ontological premise and both existing sociological theory and ordinary social observation. Nonetheless, it is possible that future experience will lead us to adjust, refine, or abandon our ontological theory of social action—for example, if later experience and theory suggest that individuals are less purposive and intentional than the theory implies, or that there are overriding forces of norm and heuristic that fundamentally drive individual behavior. In order to reconcile those future observations and theoretical developments, some changes in our system of beliefs must occur. And adjustments may occur within sociological theory, social observation, or social ontology. Our representation and understandings of the social world are a “web of belief” in which anomalies may force changes in any part of the system.

The philosophical position I am invoking here is a key part of W.V.O. Quine’s (1960) approach to empirical knowledge in Word and Object. His phrase, the “web of belief,” captures the idea well. All knowledge falls within that web, and it is held together only by observation (when statements have implications for outcomes that can be observed) and logic (Quine and Ullian 1970). Here is a concise statement of Quine’s (1960, 4-5) view of theory:

Analyze theory-building how we will, we all must start in the middle. Our conceptual firsts are middle-sized, middle-distanced objects, and our introduction to them and to everything comes midway in the cultural evolution of the race . . . . We cannot strip away the conceptual trappings sentence by sentence and leave a description of the objective world; but we can investigate the world, and man as a part of it, and thus find out what cues he could have of what goes on around him.

Fundamentally, the core of this position seems unassailable: there is no epistemic foundation possible outside the loose constraints of empirical observation and logic that can justify a set of beliefs about the fundamental structure of the world. There is no secret recipe for arriving at metaphysical knowledge through purely philosophical pathways.

There is another idea from Quine (1960) that is relevant for meta-theories of social ontology—the idea of underdetermination of theory by experience. In the social realm, this theory works out somewhat differently than in the realm of physics. When we consider the nature of the social world, we must begin somewhere, with some set of social phenomena in the center of our attention. And different starting points lead us as well to rather different ontological views of the social. If we begin with examples drawn from microeconomics—price behavior, labor migration, investment patterns—we are most likely to focus our ontological eyes on individual rationality and the sets of constraints in which individuals make choices. Seen from this perspective, the social world is a composite of individual rational actors responding to specific environments of choice. If we begin with examples from the field of the sociology and politics of ethnic conflict, we are likely to highlight groups, processes of mobilization, the prerequisites of cooperation, the workings of conflicting sets of norms, and the variety of motivations that individual Hindu and Muslim villagers have as they interact with various leaders and political organizations. Rational self-interest does not disappear in this ontology, but we are drawn to a broader palette of social motivations and a richer conception of the ontology of a “social group” (along the lines of Gilbert (1989), Tuomela (2013), or List and Pettit (2011)).

Or take a more fundamental kind of ontological disagreement, this time between those who think that “actors doing things” represent rock-bottom social stuff, and those who argue for processes, relationships, and doings are more fundamental than discrete actors in decision settings. Andrew Abbott (2016) and Charles Tilly (1995, 1998) represent versions of this latter social ontology; and each of these eminent social theorists argues against an individualist social ontology. But here is a key point: both the process ontology and the individual-actor ontology can be used to interpret the same sets of social facts. (Is this the social equivalent of “wave-particle duality”?) Seen from one perspective, the soccer players on attack instantiate a relational process (Elias and Dunning 1986); seen on a more granular level, the soccer players can be treated as individual sensory-cognitive-kinesthetic actors responding to the fluidly changing situation on the pitch, but acting individually. The two ontological perspectives are complementary rather than contradictory.

And how about what some would regard as the ultimate ontological disagreement about the social world, between individualists and holists, or between agent-centered and structure-centered theories? Do the holists and structuralists live in a different world from the agent-centered ontologists, economists, and game theorists? No. Both ontologies are descriptions of the same realities, but they proceed from different judgments about foreground and background. There are aspects of the social world where it is most instructive to highlight the properties of groups, structures, and normative systems, which suggests a holist ontology. But equally there are aspects of the social world where the most interesting puzzles lie at the level of individuals making sense of the world around them, making decisions about what to do, and setting out on a plan of action. This suggests an individualist ontology. Which ontological framework is true? My view is that they are not incompatible, and both theories are approximately true. As has long been understood in the agent-structure debate, social phenomena consist of agents-in-structures, and there is no fact of the matter ultimately about which part of the phrase is more fundamental. Structures are embodied in individual actions and mental frameworks; agents are enmeshed in structures that shape their choices and their beliefs. This is part of what Margaret Archer (1995) hopes to transcend through her theory of morphogenesis. She puts forward a social ontology that takes both agent and structure as part of an analytical framework.

Some readers might say that the “ontological disagreements” described here are really substantive disagreement about sociological and economic theory, empirical disagreements rather than philosophical disagreements. But that is exactly my point: ontological disagreements are just high-level scientific disagreements. The difference between ontological statements and hypothetical theoretical statements is merely a matter of degrees—degree of abstractness, degree of generality, degree of distance from experimental and observational evidence. “Economies are constituted by specific modes of production” is a theoretical statement within a particular strand of theory within political economy; “social life is constituted by actors in meaningful relations with each other” is an ontological statement common to a great variety of social-scientific theories. Both statements stand or fall according to the degree that they play important roles within empirical theories of the social world.

It should be noted as well that there is an important descriptive feature of ontological theorizing, in the sense offered by Peter Strawson (1963): careful efforts to uncover the conceptual presuppositions of a way of talking about a domain of experience. Strawson (1963, 15) describes his efforts in these terms: “Part of my aim is to exhibit some general and structural features of the conceptual scheme in terms of which we think about particular things.” Part of the task of philosophical theories of social ontology is to uncover the generally unspoken presuppositions made by social scientists about the nature of the social world as they formulate theories and explanations. It is plain that economists, sociologists, anthropologists, and historians bring a wide range of contrasting intellectual frameworks about the nature of the social world to their research, and it is valuable to make those frameworks explicit. Consider theories of modern economic development. Immanuel Wallerstein’s (1976) conception of the social world is very different from that of Douglas North (1981). Wallerstein is committed to a conceptual scheme in which causal primacy is attributed to vast global entities (world systems), while North focuses his attention on the intermediate-level historical patterns that result from purposive individuals acting within specified institutional settings. Their conceptions of the world are different but not incompatible, and the ontological insights we gain from each are valuable whether we favor the Wallerstein framework or the North framework.

Other readers might argue that my treatment of ontological disagreement shows that none of the parties can claim “truth” for their frameworks, which might be thought to push us in the direction of Lauer’s pragmatist interpretation. On the pragmatist interpretation, the ontological theory is valued simply as an aid to inference, not as an approximately true description of the world. Individualist social ontology is more highly valued to the extent that it suggests more fecund theoretical developments than structuralist ontology. My philosophical approach to interpreting ontology is different from that. I want to say that two ontological theories can disagree, not because they assert inconsistent claims about the world, but because they are partial representations of aspects of the world, and both perspectives may potentially be valuable for theory and partially true. If we reject the hope that there should be a complete and final ontological theory, then the fact that two ontological theories are different and highlight different aspects of the world does not entail that one of the theories must be false. It is possible that the disagreements between the two ontologies might suggest more ontological work to show how they can be reconciled into a single coherent representation of the full range of phenomena. This approach would resemble the search for a unified field theory in physics. But on its face, both ontological frameworks may be construed as being approximately true. We should not expect closure of ontological theory; we should not put our eggs in the “unity of scientific ontology” basket and hope for a definitive and final list of “what there is, socially speaking” in terms of which all social-science concepts can be reconstructed. The social world has enough heterogeneity, contingency, and plasticity to support a wide range of ontological ideas.

In short, I argue for a third position on ontology that is different from either of the approaches described by Lauer—ontology without apriori reasoning, ontology as simply the outer limit of scientific and empirical theorizing, and ontology construed realistically as making claims that are approximately true. Against Epstein and Searle, I reject the possibility or utility of pure apriori reasoning in ontology. It is therefore unfortunate for Lauer to pin the whole of realist social ontology on Searle and Epstein.

On this approach, social ontology is corrigible, provisional, and advisory, rather than certain and apriori—even though ontological statements are intended referentially. Tuukka Kaidesoja (2013) presents arguments along these lines addressed to Bhaskar and the critical realists, where he proposes to interpret realism “naturalistically” (as an extension of empirical scientific research and theorizing). As noted above, it fits Quine’s theory of empirical knowledge as well. And this is the model I have in mind when I say that social ontology is crucial to social science. By having some clearly formulated ideas about how the social world may work, future empirical and theoretical research can be more confident of moving in a positive direction.

It is perhaps surprising to invoke Goodman and Quine in the context of reflections on a generally realist interpretation of the social sciences, since their philosophies are anti-realistic (or at least agnostic between realism and anti-realism), and the logical-positivist background of much their thinking is anathema to many scientific realists. Moreover, both lend support to a certain kind of conceptual relativism: Quine through his arguments about the indeterminacy of translation and ontological relativity, and Goodman (1978) through his view of “many worlds” in Ways of Worldmaking. This perspective does not necessarily commit one to anti-realism; in fact, Hilary Putnam’s effort to create a defensible formulation of “internal realism” indicates one possible direction of argument toward realism from these premises (Baghramian 2008).

The statement of realism in which I have the greatest confidence is this: we are justified in acknowledging the reality in the world of the things, processes, structures, and forces that are postulated or implied by the best scientific theories we have to date. And we acknowledge that these beliefs, like all scientific and empirical beliefs, are fallible and corrigible. This approach has resonance with pragmatist philosophy of science and epistemology associated with Quine’s “web of belief” model of knowledge. But it also admits of a realist interpretation: when the social ontologist maintains that the theories of political science are grounded in a specific set of “ontological” assumptions about the nature of the individual actor, he or she is making assertions which are amenable to being approximately true.