New thinking about metaphysics

It seems that there is a lot happening in metaphysics these days. There is of course the return to Aristotle that has occurred within the resurgent field of powers ontology in the theory of causation (e.g. Ruth Groff, Ontology Revisited: Metaphysics in Social and Political Philosophy, Stephen Mumford and Rani Lill Anjum, Getting Causes from Powers). But volumes like David Oderberg’s Classifying Reality and Real Essentialism make it clear that the trend is broader than this. And here too, Aristotle is making a come-back.

The fundamental question for metaphysics is this: what exists in the world in which we live? But this is actually a diverse group of questions. We might be asking what specific particulars exist. We might be asking a more general question, what kinds or species of particulars exist (living things, rocks, liquids, spacecraft)? Or we might be asking the most general question, what categories of stuff are there in the world (events, individuals, properties, relations, space, cause, …)?

This is one set of complexities raised by the field of metaphysics. Another complexity is more abstract: is metaphysics about the world or about the systems of thought that we use to make sense of the world? The three levels of questions just mentioned make it seem that metaphysics is about the world; but we might argue that metaphysics is really about the systems of categories and concepts that we use to formulate representations of the world.

The issue of realism is as relevant for metaphysics as it is for science. We might want to know whether the particulars referred to by the noun “electron” really exist in nature. Or we might want to know whether the classification system of Maxwell’s electromagnetic theory (magnetic fields, electrical charge, photons, electromagnetic force, …) draws real distinctions within nature — distinctions that divide the stuff of nature at its joints. And finally, we might want to know if reality “really” consists of individuals and properties, or of some other more complicated set of ontological categories.

In Real Essentialism Oderberg proposes a renewed development of Aristotle’s metaphysics as the basis for an answer to these questions:

The aim of Real Essentialism is to rehabilitate some of the core ideas of Aristotelian metaphysics in a contemporary context devoid of the minutiae of historical exposition and textual exegesis (though quite a bit of this will be found in the notes to each chapter). (1)

Here is his statement of essentialism within Aristotelian metaphysics:

At the heart of traditional metaphysics is the thesis that everything has a real essence – an objective metaphysical principle determining its definition and classification. Such principles are not mere creatures of language or convention; rather, they belong to the very constitution of reality. (2)

In his introduction to Classifying Reality Oderberg opens with this setting of the problem; and it might be taken as a proclamation for a new metaphysics more generally.

Is reality classifiable? In other words, does it have boundaries or ‘joints’ that enable us to assign various categories to its different constituents? (1)

As several of the contributors to Classifying Reality point out, there are two broad levels of ontological or metaphysical thinking that we can usefully engage in. First, at the level of the sciences directly there is the question of what kinds of things there are in a specific domain. The ideas of “animate thing,” “vertebrate”, “mammal”, and “horse” arise at this level in the biological sciences.

But there is also a more abstract question that can be posed: what distinct sorts of stuff do we need to allude to in analyzing many or all areas of science? This might be called general metaphysics. We might try to make do with a very spare ontology of particulars and events; or it might be argued that there are distinctions among bits of reality that warrant separate categories. (Here is one such ontology: reality consists of entities, continuents, occurrents, and qualities; (4).) Or we might say that our ontology needs to refer to individuals, properties, relations, processes, and events, and that none of these can be defined in terms of some combination of the others.

In his contribution to Classifying Reality E. J. Lowe observes that analytic philosophers since Frege and Russell have tried to answer the more general question by referring solely to particulars and properties or relations, and have generally tried to understand properties and relations extensionally (as sets of particulars possessing the property). Lowe refers to this simple ontology as a “Fantology” — there are particulars denoted by lower case letters, predicates denoted by uppercase letters, and statements of the form “Fa” (a has the property F).  Lowe favors a renewed attention to Aristotle’s metaphysical theories, arguing that reality consists of primary substance, secondary substance, property or attribute, and individual accident or mode (11). And he believes that the minimalist ontology deriving from Frege (the Fantology) cannot handle the needs we have in creating language for describing the world. Lowe makes use of this more complex understanding of the kinds of things there are in the world to formulate a new version of formal logic. He provides a formalism for expressing the different kinds of statements that can be made within the more extensive universal ontology. “The system of formal logic whose language I have been constructing is meant to be one which respects and reflects certain fundamental categorial distinctions of an ontological nature” (18). What this formulation does not provide is a set of rules of derivation.

Tuomas Tahko takes up the issue of realist metaphysics in his contribution, “Boundaries in Reality.” He believes that there is a basis for asserting that the assertions of metaphysics may be true or realistic; and to do so he addresses the conventionalist arguments that exist against this conclusion. Most generally the conventionalist line is this: there are multiple systems of conceptualization and classification, and there is no “best” system. Therefore there is no basis for concluding that one of these maximal systems is more realistic than another; and therefore there is no basis for metaphysical realism. Here is one version of this view in the words of Achille Varzi:

If all boundaries were the product of some cognitive or social fiat, if the lines along which we “splinter” the world depended entirely on our cognitive joints and on the categories that we employ in drawing up our maps, then our knowledge of the world would amount to neither more nor less than knowledge of those maps. (43)

But Tahko believes that the preponderance of evidence works against the conventionalist view:

We have seen that our system of classification is fundamentally grounded in reality. We can state this with some confidence, as otherwise this system would hardly be so reliable. It is an open question which entities are genuine, bona fide entities; we need philosophical inquiry as well as science to determine this. (60)

Gary Rosenkrantz brings the arguments into connection with biology in his contribution. He believes that there are necessary and essential truths about living things:

First, an animate being is a concrete entity capable of living a life. Second, a life, or at least any contingent being’s life, is a process consisting of a series of intrinsic changes in an animate being. Third, an animate being is not a process; such a being — at least of the most basic sort — is what was traditionally called an individual substance. (79)

And further:

I shall argue that the question “What is an animate being?” can be answered by quantifying over ontological categories and natural kinds. (82)

Here is how Rosenkrantz characterizes natural kinds:

Every such natural kind, K, is such that: (i) it is impossible that something instantiates K contingently, (ii) K is a proper object of inquiry in natural science, (iii) K figures in one or more natural laws, (iv) K is possibly instantiated in the absence of an intention or belief of a contingent being that an instances of K is for performing some goal-directed activity, (v) K supervenes on structural and compositional properties, i.e., necessarily, for any x&y, if x instantiates K and x&y have the same structural and compositional properties, then y instantiates K, and (vi) K places limits on the kinds of parts an instance of K could have. (83)

This is tough slogging. We would like to know if this series of features of “natural kinds” are thought to be definitional or substantive. Is it a discovery about my wedding ring and “gold” that “it is impossible that the ring instantiates ‘gold’ contingently”, or is this just a matter of definition? In other words, why should we accept (i)-(vi) as being true of natural kinds? And this raises a more pervasive question: what kind of philosophical or scientific reasoning is needed to establish truths of general metaphysics?

The diagram at the top is taken from Barry Smith’s contribution to Classifying Reality in which he attempts to construct a “Basic Formal Ontology” — a “top-level ontology that is serving as domain-neutral framework for the development of lower level ontologies in many specialist disciplines, above all in biology and medicine” (101). His account offers a formalized set of relations that can exist between ontological categories: “-is-a-” and “-part-of-“, for example. And he distinguishes between things in terms of their temporal qualities: continuants and occurrents, for example (107), or “things” and “events”.

What I find interesting and worthwhile about Classifying Reality is that it illustrates that there are genuine and important questions within the domain of general metaphysics — questions that are provoked by our ordinary efforts to conceptualize and understand the world around us. What is challenging is to validate the kinds of reasoning that are offered as foundation to conclusions in this field. Does pure philosophy suffice? Do we need to “naturalize” metaphysics in order to have credible theories and conclusions? Should metaphysics simply be considered the most abstract end of the scientific enterprise, ultimately dependent on empirical reasoning and logical deduction? Or is there a realm of autonomous philosophical thinking that can lead to substantive metaphysical conclusions? My own philosophical training makes me wary of that final thought; but the contributors to this new metaphysics are making their best efforts to validate it.

Conceptual schemes and social ontology

What does grammar tell us about the nature of our representations of the world? Do the linguistic categories that we use fundamentally shape the way we organize our understanding of the world? Do different cultures or different linguistic communities possess different “conceptual schemes”? Are different conceptual schemes incommensurable or can we translate from one to the other?  These questions come up in the context of any discussion of social ontology — what does the social realm consist of?  In an earlier post we noticed that “thing” and “object” are ontological categories that perhaps don’t work as well in the social realm. Perhaps more fluid categories such as process, relation, or activity work better.

First, what is a conceptual scheme?  It is an interrelated set of high-level, abstract concepts that allow us to break the empirically or historically given into a discrete set of cognitive boxes.  We might think of it as our highest-level concept vocabulary, within which more specific descriptors are arranged.  Our conceptual scheme gives us the mental resources needed to represent, describe, and explain the empirical reality we encounter.  Color, shape, mass, position, and force might be examples of components of a conceptual scheme for the realm of ordinary empirical experience.  Structure, group, ideology, and network might be components for the realm of ordinary sociological experience.  A conceptual scheme is thought in some way to be comprehensive: all the phenomena in a certain domain ought to find a place within the conceptual scheme.

Peter Strawson offered a very focused analysis of the everyday metaphysics involved in the ways we analyze and represent the world around us.  He proposes in Individuals (1959) that we can do “descriptive metaphysics” by examining the conceptual schemes we actually use.   And he argues that there are core conceptual categories that are universal.  (Paul Snowdon provides a useful discussion in his article in Stanford Encyclopedia of Philosophy).

Here is Strawson’s preliminary description of a conceptual scheme:

We think of the world as containing particular things some of which are independent of ourselves; we think of the world’s history as made up of particular episodes in which we ourselves may or may not have a part; and we think of these particular things and events as included in the topics of our common discourse, as things about which we can talk to each other.  These are remarks about the way we think of the world, our conceptual scheme.  A more recognizably philosophical, though no clearer, way of expressing them would be to say that our ontology comprises objective particulars….  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. (Individuals, 15)

How might we begin to provide a “descriptive metaphysics” for social knowledge along the lines of what Strawson describes?  It is clear, to begin, that sociological analysis generally involves a rich and intertwined set of concepts and ontological assumptions about social phenomena.  Consider the range of approaches we might take in analyzing a complex historical phenomenon such as fascism: as a social movement, as a political psychology, as an expression of psychopathology, as an ensemble of ideological currents, as a set of political institutions, as a collection of social constituents, and so on, indefinitely.

More fundamentally, what conceptual choices do we need to make when we consider the swirling, fluid complexity of politics, culture, and struggle in Europe in the 1930s?  We might place ideological and cultural change at the center; we might focus on the artistic and literary creations of the period; we might emphasize power or social class; we might give primary emphasis to economic change; or we might be drawn particularly to differences in behavior and regime across countries and regions of Europe.  Each represents a different way of conceptualizing the historical reality of the 1930s in Europe.

So how can we make some progress towards analyzing social grammar or social conceptual frameworks?  We can ask this sort of question at two levels — ordinary social cognition and language, and organized empirical social science.  At the ordinary-language level, we can ask questions like these: how do ordinary speakers represent the social world in which they live?  What is the nature of the American-English social vocabulary, the descriptive and referential terms that American people use to make statements, draw distinctions, and offer generalizations about the social world they inhabit?  At the level of theory, we can consider a given area of research and ask about the semantics and logical relationships associated with the terms that theorists use to describe and explain the phenomena of interest.

For ordinary language, we might find a list of common terms such as these:

  • Washington [the Federal government, the bureaucracy, the political system of the Congress, the major Federal agencies]
  • Lansing [state government and its bureaucracy]
  • justice / injustice [of taxation, affirmative action, executive salaries]
  • corruption [misuse of powers of office in private or public sectors]
  • major economic institutions [banks, banking system, corporations]
  • major economic facts and circumstances [unemployment, poverty, recession]
  • religious institutions
  • religious / ethnic identities
  • facts about race, racial differences, racial inequalities
  • interpretations of socially oriented behavior by others [altruism, egoism, pride, shame, rudeness]
  • judgments about unfavorable social change [“kids have no values anymore”]

Ordinary people use these concepts and other to organize and criticize their social world; and they are often articulate about what they mean by the various concepts.  But ordinary social cognition is perhaps less able to sketch out the relations that exist among the various social phenomena; this, perhaps, is one of the key tasks of social theory.  (Here is an earlier post on ordinary social cognition.)

Second, we might consider the vocabulary and conceptual resources of a given sociologist or sociological tradition.  For example, here are the main concepts Michael Mann uses in his description and analysis of European fascism in Fascists:

  • fascism [to be defined as “the pursuit of a transcendent and cleansing nation-statism through paramilitarism”]
  • socialism
  • fascist followers and activists
  • power organizations
  • social movement
  • nation-state
  • nationalism
  • ideology
  • social constituency [groups characterized by status, class, occupation]
  • individuals
  • modernity
  • paramilitarism
  • authoritarian regime
  • democratic regime
  • capitalism [as an industrial social whole]
  • social power [ideological, economic, military, political]
  • individualism, racism, ethnic purity ideology [as components of ideology]
  • major events and crises — World War I, the Great Depression
  • religious institutions and ideologies

These specific concepts could be related to a fairly short list of higher-level social concepts or what we might call social categories: individuals and their characteristics; social groups; structures; ideologies; events; influence terms [power, prestige, status].  But almost all the concepts on the list drawn from Mann’s work involve a conceptual assemblage from the higher-level categories.  Capitalism is a set of structures, a set of social movements, and a set of ideologies.  Racism depends upon both structure and ideology.  Modernity is an ideological-cultural formation, a technological-scientific stage, and a socio-economic formation.  So the relation between the higher-level category system and mid-level sociological concepts is not one of subsumption but rather one of assembly, combination, or construction.

It is sometimes thought that our conceptual systems are simultaneously contingent and deeply influential in determining how we analyze the world around us.  Different conceptual systems lead to different and incommensurable representations of the world.  Donald Davidson wrote a pivotal essay on some of these questions (“On the very idea of a conceptual scheme” (1974; link)).  Here is how Davidson summarizes the conceptual-relativist view:

Conceptual schemes, we are told, are ways of organizing experience; they are systems of categories that give form to the data of sensation; they are points of view from which individuals, cultures, or periods survey the passing scene. There may be no translating from one scheme to another, in which case the beliefs, desires, hopes and bits of knowledge that characterize one person have no true counterparts for the subscriber to another scheme. Reality itself is relative to a scheme: what counts as real in one system may not in another. (5)

But ultimately Davidson argues that conceptual relativism and incommensurability are unintelligible.  They are claims that cannot be stated coherently.  And more positively, Davidson argues that we can understand each other’s concepts and words by making use of a principle of charity: we interpret the other’s speech, vocabulary, and syntax in such a way as to maximize the truth of statements he/she utters.

We do this sort of off the cuff interpretation all the time, deciding in favor of reinterpretation of words in order to preserve a reasonable theory of belief. As philosophers we are peculiarly tolerant of systematic malapropism, and practised at interpreting the result. The process is that of constructing a viable theory of belief and meaning from sentences held true.  (18)

We get a first approximation to a finished theory by assigning to sentences of a speaker conditions of truth that actually obtain (in our own opinion) just when the speaker holds those sentences true. The guiding policy is to do this as far as possible, subject to considerations of simplicity, hunches about the effects of social conditioning, and of course our common sense, or scientific, knowledge of explicable error. (18)

(Here is a good discussion of Davidson’s view in the Stanford Encyclopedia of Philosophy.)

What is the situation in the field of social knowledge?  Are there deeply divided conceptual beginnings for the analysis of the social realm?  Or can we be confident in the mutual comprehensibility across practitioners of Marxist sociology, Durkheimian sociology, and ethnomethodology?

We can get some leverage on this question by asking whether we can provide concrete examples of candidates for alternative sets of social conceptual schemes.  For example, are individualism and holism distinct conceptual schemes for social cognition?  The first identifies human individuals as the fundamental “particular”; whereas the second identifies the social whole (structure, morality, ideology, class, way of life) as the fundamental particular.  The first requires that we define or specify higher-level social entities or conditions in terms of a compound of features of individuals; the second takes the social whole as irreducible and specifies individuals in terms of their relations to a set of social factors.

Here is another possible example — perhaps materialism and idealism are distinct conceptual schemes within which to organize social experience.  The materialist scheme identifies a set of circumstances of the human organism (needs), the natural and build environment, and the forms of social activity that transform the environment as fundamental to social analysis.  The idealist scheme takes states of consciousness — ideas, ideologies, moralities, wants, preferences, modes of reasoning — as fundamental to social analysis and undertakes to characterize social facts in these terms.

Or consider a third possible example: structure and process.  A structure is an enduring configuration of social characteristics and positions, reproducing a set of powers and constraints for individuals enmeshed in these social relations.  A process is an ensemble of things in circumstances of change over time.  Structures emphasize permanence and stability; processes emphasize change and impermanence.  So perhaps the “structure” lens leads sociologists to a very different representation of the social world than the “process” lens.

These examples make it credible that there are in fact alternative conceptual beginnings from which we can analyze the social world.  What does not seem to be true, however, is the idea that these beginnings are incommensurable.  Instead, it seems persuasive that ideas and statements that originate in an ontology of social wholes can be effectively restated in an ontology that originates in a world of individuals; likewise, materialist and ideological approaches to the social world seem compatible and mutually constructive rather than contradictory and incommensurable.  The dichotomies considered here are not exclusive or incompatible.  In fact, any adequate explanation of a social process or outcome is likely to need to refer to both sets of categories.  And this implies something very similar to the position Strawson and Davidson arrive at: the idea of inter-translatability and mutual comprehension across these large conceptual divides.

(These questions converge to some extent with several other lines of thought — the Whorf hypothesis (Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf), Quine’s theories of indeterminacy of translation and ontological relativity (Word and Object, Ontological Relativity), Kant’s view that all knowledge is structured through a set of “categories” including cause, space, time, and objects (Critique of Pure Reason), and Kuhn’s view of the incommensurability of scientific paradigms (The Structure of Scientific Revolutions).  Significantly, Strawson discusses Kant at length in The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason, and Davidson takes up the debate with Quine, Whorf, and Kuhn.)

Relations, processes, and activities

An earlier post asked what sorts of social entities exist. Posing the question this way leads us to think of persistent abstract things populating the social world — for example, structures, organizations, or institutions. But as a commentator to the earlier post pointed out, there are persistent phenomena in the social world that don’t look much like things and look more like activities and processes. Grammatically they have more in common with verbs than nouns. And when many such social phenomena are described using nouns, we are often forced to interpret them in a non-referential way.

Take the social realities of friendship, solidarity, and inflation.

The first is a characteristic of social relationships; it is a relational concept.  It doesn’t make sense to think of “friendship” as a concatenation of monadic atoms of “friend units”; rather, the concept of friend evinces a set of relational characteristics between persons.  So “friendship” doesn’t designate a continuing “thing” in the world; instead, it designates a complex relational and psychological feature of pairs of persons, widely separated across population and space.  Our theory of friendship encapsulates our interpretation of the mental and behavioral states of persons who are in the relationship of friendship with each other.

Solidarity has this feature of relationality, and it adds a feature of social motivation and psychological orientation to a group. We can ask whether solidarity exists in the social world, and we can reasonably answer that it does.  But when we affirm that “solidarity exists”, we really mean that “there are numerous instances of groups and individuals in which members of the group willingly conform their behavior to the needs and purposes of the group.”  Our explanation of “solidarity” is likely to invoke abstract ideas about individuals within consciously constituted groups rather than something analogous to a social substance.  So solidarity is not a thing, but rather a dynamic feature of consciousness shared in varying ways by individuals who orient themselves to a group.

“Inflation” is a different sort of social noun. It refers to a complex social state of affairs reflecting a set of processes in which prices of goods are determined by market forces and prices are rising across a range of commodities. We cannot define the social reality of “inflation” without specifying a set of distributed social facts and identifying a number of social processes.  So inflation too doesn’t look at all like a social thing.

Each of these social nouns corresponds to a social reality. But this reality doesn’t look much like a set of fixed entities or composites of entities. The semantics of objects and things doesn’t work well for this range of social vocabulary. And yet each of these terms identifies a domain that is perfectly well suited to empirical inquiry and discovery. The social reality of friendship practices differs across cultures; friendship practice has some degree of stability over time; and we can discover quite a bit about the culture, norms, and practices of friendship in a particular culture.  And likewise with solidarity and inflation.  Each is a legitimate object of empirical inquiry; and neither conforms to the ontology of “thing” with fixed location and properties.

What this discussion suggests is that our ontology of the social world needs to encompass not only a range and variety of entities — structures, institutions, organizations, but also ontological categories that reflect a more fluid set of social realities: processes, practices, rules, relations, and activities.  This observation converges with the styles of thought of thinkers as various as Charles Tilly and Norbert Elias; Tilly refers to “relational realism,” and Elias is an advocate of “process sociology”.

What exists in the social realm?

What sorts of social things exist?  Does the “proletariat” exist as a social entity? There are certainly workers; but is there a “working class”? What is needed in order to attribute existence to a social agglomeration?

We might want to say that things exist when they have enough persistence over time to admit of re-identification and study from one time to another.  Persistence involves some degree of stability in a core set of properties.  A cloud shaped like a cat has a set of visible characteristics at a given moment; but these characteristics disappear quickly, and this collection of water droplets quickly morphs into a different collection in a short time.  So we are inclined not to call the cat-shaped cloud an entity.  On the other hand, “the Black Forest” exists because we can locate its approximate boundaries and composition over several centuries.  The forest is an agglomeration of trees in a geographical space; but we might reasonably judge that the forest has properties that we can investigate that are not simply properties of individual trees (density and canopy temperature, for example).  The forest undergoes change over time; the mix of types of trees may shift from one decade to another, the density of plants changes, and the human uses of forest products change.  And we can ask questions like: “How has the ecology of the Black Forest changed in the twentieth century?”  So it seems reasonable enough that we can refer to the forest as a geographical or ecological entity.

We can also classify individual forests into types of forests: temperate rain forest, tropical rain forest, coniferous forest, etc.  (Here is a 26-fold classification of forests by UNEP-WCMC; the map below represents the global distribution of these types of forests.)  And we can ask ecological questions about the properties and processes that are characteristic of the various types of forests.

So what characteristics should a putative social entity possess in order to fall within the working ontology of the social sciences?  Here are a few possible candidate ontological features that might be associated with thing-hood in the social realm:

  • persistence of basic characteristics over time — spatio-temporal continuity and social analogs such as nucleated population with shared norms and identities
  • an internal structural-functional organization
  • some sort of regulative social process that maintains the thing’s identity over time, either internal or external 
  • social cohesion among the individuals who constitute the entity deriving from their social orientation to the entity (labor union, religious community, ethnic group)
  • an account of the particular material-social mechanisms through which the identity and persistence of the entity are maintained

According to these sorts of criteria, we might say that social things like these examples exist:

  • United Auto Workers
  • General Motors corporation
  • First Presbyterian Church of Dubuque
  • Missouri Synod
  • Kylie Minogue Facebook fan club 
  • 18th Street gang of Los Angeles
  • Michigan Legislature
  • Internal Revenue Service
  • University of Wisconsin
  • apprentice system for electrical workers
  • social practice of Islamic charity

Here is a slightly more abstract formulation.  We might say that these kinds of social entities exist: organizations, both formal and informal; networks of individuals oriented to each other and/or a social goal; social groups unified by features of consciousness or existential circumstance; bureaucracies of the state; enduring social practices; institutions possessing internal organization, rules, and purposes.

Social entities are composed of socially constituted individuals.  So the sinews of composition are important.  We can recognize a wide range of ways in which individuals are composed into larger social entities: agglomeration, adherence, mutual recognition, coercion, contractual relationships, marketing, recruitment, incentive systems, …  This is one place where “assemblage” theory seems to be useful (Manuel DeLanda, A New Philosophy of Society: Assemblage Theory And Social Complexity.)

Part of the confusion in this topic is the distinction between things and kinds of things. We might agree that the Chicago Police Department exists as a social entity.  But we may remain uncertain as to whether “police departments” or “state coercive apparatuses” exist as higher-order categories of social things. And perhaps this is a confusion; perhaps the issue of existence applies only to individual entities, not kinds or classes of entities. On this approach, we would stipulate the minimum characteristics of existence we would want to require of individual social entities and then be “nominalistic” about the higher-level categories or concepts into which we classify these singular individuals.

In considering the ontology of the social world it is important to be attentive to the fallacy of reification: the error of thinking that the fact that we can formulate an abstract noun (proletariat, fascism) allows us to infer that it exists as a persistent, recurring social entity.  So when we identify a given social entity as an X, we need to regard it as an open question, “What do X’s have in common?”  We can avoid the fallacy of reification by focusing on the importance of providing microfoundations for the enduring characteristics of social entities.  It is the underlying composition of the entity rather than its location within a classificatory system that provides an explanatory foundation for the behavior of the entity.

Are there "social kinds"?

Philosophers of science sometimes define the idea of a natural kind as “a group of things that share a fundamental set of causal properties.” Examples might be “gold,” “metal,” and “protein molecule.” And some philosophers assume that scientific realism means being realist about natural kinds. Do the typical concepts used in the social sciences succeed in identifying a social analog to natural kinds, which might be referred to as “social kinds”? And if not, is it possible to be realist about the social world but anti-realist with respect to “social kinds”?

First, what is involved in being “realist” in connection with the historical and human sciences? It is to assert several independent things: first, that there is the possibility of (fallibly) objective knowledge of social facts; second, that there are “social facts” to be known – that is, there are some mind- or interpretation-independent things that happen and can be the subject of knowledge; and third (questionably), that there are categories of higher-level social entities that “really” exist in the way that some philosophers say that natural kinds exist. It is entirely defensible to be a scientific realist in the social sciences, and I want to support the first two ideas but to argue against the third.

Concepts are of course essential to social knowledge. The heart of social inquiry has to do with coming up with concepts that allow us to better understand social reality: for example, racism, patterns of behavior, free market, class consciousness, ethnic identities. Theory formation in the social sciences largely consists of the task of constructing concepts and categories that capture groups of social phenomena for the purpose of analysis. But even the most successful social concepts do not identify groups of phenomena that could be called a “social kind.” High-level social concepts that serve to pick out groups of social phenomena—states, riots, property systems—generally do not refer to causally homogeneous bodies of social phenomena; instead, each of these is composed of individual social formations with their own history and circumstances. There is no uniform causal constitution that underlies all states or riots. The philosophical notions of “family resemblance” and “cluster concepts” serve better to characterize these high-level social concepts than does “natural kind”.

Examples of what might have been thought to be social kinds might include concepts such as these: proletariat, underclass resentment, revolutionary situation, racism; liberal representative states; fascism; feudalism; bureaucratic state. But I hold that these are not kinds in the strong sense that philosophers of the natural sciences have in mind. Rather, they are plastic, variable, opportunistic, individually specific instantiations across a variety of human contexts. We need to be able to identify some topics of interest, so we need language and concepts; but we must avoid reifying the concepts and thinking they refer to some underlying discoverable essence. (Think of how Chuck Tilly conceptualizes riot, rebellion, and resistance in terms of “contentious politics.” Rightly, he avoids the idea that there is one common thing going on in these instances across time, history, and place; his goal is to identify a medium-sized body of causal mechanisms that bundle together in various contexts to give rise to one signature of contention or another.)

The discovery of causal processes is essential to social explanation — not the discovery of high-level uniform categories of social events or structures. We explain social outcomes best when we can uncover the causal mechanisms that gave rise to them. However, most social ensembles are the result of multiple causal mechanisms, and their natures are not common, simple, or invariant. “States” embody mechanisms of social control. But as Tolstoy said about unhappy families, every state manages its contention in somewhat different ways. So we can’t and shouldn’t expect common causal properties across the class of “states”. And this is directly relevant to the central point here: the “state” is not a social kind, and there is no simple theory that encapsulates its causal properties.

This approach has specific implications for the conduct of the social sciences. For example, political science and the study of different types of states: we can identify common mechanisms, sub-institutions, building blocks, etc., that recur in different political systems. And we can offer causal explanations of specific states in particular historical circumstances — for example, the Brazilian state in the 1990s. But we cannot produce strong generalizations about “states” or even particular kinds of states — for example, “developing states”. Or at least, the generalizations we find are weak and exception-laden. Rather, we must build up our explanations from the component mechanisms and institutions found in the particular cases.

So here is a moderate form of scientific realism that is well suited to the nature of the social world: be realist about social mechanisms but not about social kinds. Be realist and empiricist in epistemology: we can arrive at rationally justified beliefs about social mechanisms. And be a skeptic or nominalist about social kinds. There are no macro or molar-level social kinds.

The difference ontology makes

Quite a few posts here over the past few months have been on the subject of social ontology: what can we say about the nature of the social world? I’ve focused on characteristics like heterogeneity, plasticity, and contingency, and have also given thought to some of the processes through which social phenomena are “composed” of lower-level processes and mechanisms. (The topics of methodological individualism, localism, and holism fall in this category.) Why are these questions important to the philosophy of social science? And how could they possibly contribute to better research and theory in the social sciences?

One answer is that it isn’t really possible to investigate any domain without having some idea of what sorts of things the domain consists of. So attempting to arrive at perspicuous models of what the social world is made up of is a necessary step on the way to more specific forms of empirical and causal research.

A second answer derives from recognition of the harm that has been done to the cause of knowledge by misconceived ontologies in the history of science. This has been especially true in areas of knowledge adjoining human life and activity — radical behaviorism in psychology, naturalism and positivism in sociology, and what Andrew Abbott calls the “variables paradigm” in quantitative social science. Better science will result from more propititious ontology, because we won’t be in the situation of trying to force the social world onto the wrong sorts of boxes.

Third, there are good reasons for thinking that reasoning about social ontology is possible. If ontological thinking were purely apriori, then we might be reasonably skeptical about our capacity to move from philosophy to the world. But we have a form of access to social reality that we lack for the realities of nature and mathematics: we are participants in social reality and our thoughts and actions are constitutive of that reality. So theorizing about social reality of ontology isn’t wholly apriori; rather, it is more akin to a form of intelligent observation of real social processes around us. Ontological thinking is really a form of empirically informed theorizing, at a fairly abstract level.

Let’s turn now to the question of how concretely better social ontological thinking can help create better social research.

Researchers will be encouraged to explore multiple methodologies and theories when a more satisfactory social ontology guides them. Pluralism finds strong support in the ontology explored here, given the emphasis offered to social heterogeneity and contingency.

Some specific research strategies and questions are likely to arise as a result of more adequate social ontologies. Questions about causal mechanisms and processes of composition will receive special attention.

Fields of research like comparative historical sociology and case study methods will receive theoretical support and encouragement, since these approaches are particularly well suited to the ontological ideas revolving around heterogeneity, plasticity, and composition.

So, to return to the original question: it is in fact very appropriate and potentially helpful for philosophy of social science to give more attention to social ontology than it has traditionally done. The social world of the twenty-first century is chaotic and rapidly changing, and we need better mental frameworks within which to attempt to make sense of this complexity and change.

%d bloggers like this: