The case for realism in the social realm

Image. Orderly chaos in flight path of ocean-foraging albatross

The case for scientific realism in the case of physics, microbiology, and chemistry is a strong one. The theories of physics, biology, and chemistry postulate unobservable entities, forces, and properties. These hypotheses are specified in a fair degree of precision. They are not individually testable, because we cannot directly observe or measure the properties of the hypothetical entities. But the theories as wholes have a great deal of predictive and descriptive power, and they permit us to explain and predict a wide range of physical phenomena. And the best explanation of the success of these theories is that they are true: that the world consists of entities and forces approximately similar to those hypothesized in physical theory. So realism is an inference to the best explanation, based on the engineering and observational successes of physics, chemistry, and biology. (In the diagram above we might hypothesize that the foraging strategies of the albatross have evolved towards a combination of random walk and orderly search pattern through a process of natural selection; this hypothesis can be empirically investigated in a variety of ways.)

If we lived in a more chaotic physical world, with a larger number of more variable forces at work, our physical theories would be greatly less successful at representing the behavior of observable physical systems, and we would have much less confidence in the idea that various snippets of our physical theories are "true" of the world. If space were more like a pudding with abrupt variations in curvature and gravitational force, and were in addition subject to a numerous other factors and forces, our confidence in the science of mechanics would be greatly undermined. We would never know even approximately where the fly ball will go.

The situation in political science and sociology is quite different from astronomy, atomic theory, and mechanics. First, there are no theories in the social sciences that have the predictive and explanatory success of the physical sciences. Second, the social world is more like the fantastic and chaotic scenario just mentioned than it is an ice rink with frictionless surfaces and predictable mechanics. The social world embodies multiple heterogeneous causal and structural influences that aggregate in contingent and surprising ways. Third, sociologists and political scientists sometimes make hypotheses about unobservable or hypothetical social entities. But these hypotheses do not assume the logical role of that played by hypotheses in the natural sciences. Hypothetical social entities may be unobservable in a fairly ordinary sense -- no one can directly observe or measure a social class. But in fact, these concepts do not depend on holistic confirmation in the way that hypotheses in the natural sciences do. Rather, it is perfectly possible to further refine our ideas about "social class", "prisoners' dilemma", or "bipolar security field" and then investigate the manifold aspects of these concepts through direct social research. Sociology and political science do not consist of unified deductive systems whose empirical success depends upon a derivation of distant observational consequences; instead, it is possible to investigate essentially every sociological or political concept through various direct methods of research and inquiry. (This ability is not unique to the social sciences. The study of animal behavior likewise admits of a variety of hypotheses at various levels that can be independently studied.)

In short, the social sciences do not possess the remarkable coherence and predictive accuracy of physics, so confidence in realism is not grounded in the high level of success of the enterprise. Sociology is not like physics.

But equally, the concepts of the social sciences are not "hypothetical constructs" that depend upon their role in a developed theoretical system for application. It is therefore possible to be piecemeal realists. Again, sociology is not like physics.

So it seems that two specific ideas follow. First, the inference to the best explanation argument for realism doesn't work at all in sociology or political science. We simply don't have the extraordinary predictive successes of a theoretical system that would constitute the ground of such an argument. Social science theories and models remain heuristic and suggestive, but rarely strongly indicative of the reality of the social factors they highlight.

But second, there is a very different kind of argument for social realism that is not available in the natural sciences: the piecemeal investigation of claims and theories about social entities, properties, and forces. If we believe that class conflict is a key factor in explaining political outcomes, we can do sociological research to further articulate what we mean by class and class conflict, and we can investigate specific social and political processes to piece together the presence or absence of these kinds of factors.

So it seems that we can justify being realists about class, field, habitus, market, coalition, ideology, organization, value system, ethnic identity, institution, and charisma, without relying at all on the hypothetico-deductive model of scientific knowledge upon which the "inference to the best explanation" argument depends. We can look at sociology and political science as loose ensembles of empirically informed theories and models of meso-level social processes and mechanisms, each of which is to a large degree independently verifiable. And this implies that social realism should be focused on mid-level social mechanisms and processes that can be identified in the domains of social phenomena that we have studied rather than sweeping concepts of social structures and entities.

This perspective converges unexpectedly with some of the thinking that Peter Manicas put forward in his book on social-science realism, A Realist Philosophy of Social Science: Explanation and Understanding. What is realism, in the natural sciences, he asks? It is not a general claim to have discovered the universal laws of everything.

Rather, more modestly, theory (at least in one of its clear senses) aims to provide an understanding of the processes which jointly produce the contingent outcomes of experience. We understand why the planets move in ellipses, why materials burn, and why salt dissolves in water (if and when it does) when we have a physical theory that provides a causal mechanism. By providing the principles detailing the nature of molecules, the atomic structure of salt and water, the principles of their action, and so on, we can understand combustion and solubility – and other chemical processes. (1)

So what are the generative mechanisms in the social world? Manicas argues that these mechanisms proceed from the actions and relations of social agents:

The foregoing has also argued that persons are the dominant causal agents in society – even while, of course, they work with materials at hand. It follows, accordingly, that in the social sciences, the generative mechanisms of social outcomes are the actions of persons and no further reduction is either plausible or demanded. (75)

So his most general idea about the social world is "social mechanisms as agent-generated causal mechanisms" (2).

If this is the approach we take, then our claims about what is "real" in the social realm will be more modest that some have thought. We will understand that there are real social processes, mechanisms, and powers; that they derive from the actions and agency of actors; and that these processes can be traced out through fairly direct sociological and historical research. And we will understand too that claims about the reality of "capitalism", the world financial system, or fascism are to be understood less weightily than they first appear. Capitalism exists in a time and place; but it is understood to be an ensemble of relations and actions by the people of the time. It is not a "thing" in the way that deoxyribonucleic acid is a thing.

These thoughts should perhaps lead us to consider that the topic of realism is less important in sociology, political science, and economics than it might appear to be. Social scientists have every reason to be realist about the actions, relations, and interactions of individuals. They are justified in thinking that the practices of education and socialization that bring children to adulthood are "real" and can be empirically investigated. And they are justified in observing that there are higher-order configurations of action, power, and social relationship that are "real", insofar as they are present in the activities of the individuals who constitute them and they possess some stable characteristics over time. In other words, social scientists are justified in postulating the social reality of the social processes and institutions that they postulate and investigate. But this is a very weak and qualified conception of realism, and it suggests a fairly weak social ontology.

It will be noted that this conclusion is somewhat in tension with the argument I offered in the prior post on "flat social ontology". That's the virtue and the challenge of open-source philosophy: conclusions and arguments shift over time.

Guest post by Doug Porpora on social structures

Here is a response to my earlier post on social ontology and structure from Doug Porpora, professor of sociology, Drexel University. Doug is the author of Reconstructing Sociology: The Critical Realist Approach. Thanks, Doug, for this thoughtful and considered reflection!

FROM DOUG PORPORA:

I have four comments in response. First, while I am happy to stand somewhere alongside John Levi Martin, one important difference between us is that I attach reality – and even an emergent mind-independence – to the relations connecting social positions whereas, I believe, he would not.

Second, sticking close to an ordinary language sense of the word, I would confine the word structure to those connecting relations and not to the higher level things you speak of. I would rather call the higher level entities institutions.

Third, as you suggest, I certainly do not deny the existence of families, social movements, clubs and states. But they seem to me to be a nexus of connections among social positions.

My fourth reflection is more uncertain. Do those things – states, etc. – represent some kind of emergent entities and as such a higher level of thing? You are not alone in pushing me in this direction. Most fellow critical realists would do so, and Ruth Groff has recently been trying to get me to do the same.

I certainly believe in emergence rather than reductionism. I believe life is a level emergent from non-life and consciousness a yet higher emergent level, and self-consciousness an even further emergent level.

I believe emergence is more general than just these levels, but let us stay there. In the case of each of these levels, a new kind of causality emerges not found on the level below – replication and natural selection in the case of life, speech acts in the case of linguistic consciousness. The emergence of these new causal powers can be explained by the level below but not their functioning.

Is there anything like that going on with the putative emergent entities of families and states and such? I suppose I would say that something like Durkheimian social cohesion is an example of a new causal property not associated with individual people. So, as you suggest, do new causal properties emerge from people in social life? Yes.

Do these wholes and their properties constitute a level in the same way as does life or consciousness? I would say no for two reasons. First, as new causal kinds, speech acts and natural selection act directly from whole to whole. We can explain the causal logic connecting the holistic behavior without micro-analysis of their parts.

I do not see anything like that going on with social cohesion. Any effect it has on other wholes cannot be explained without manifesting through individual behavior. There is no new causal logic here. I think you would agree, no? Second, any new, putative causal logic is always too penetrated by acts of what Hegel called “world historical individuals” to constitute an autonomous level.

So socially emergent entities, okay. An autonomous level of them as per sociological holism? In my opinion, no. Thanks again for the reflection.

Doug

A flat social reality?

I've been inclined to talk about the social world in terms of levels or layers, with a few provisos -- multiple layers, causation across layers, fuzzy boundaries (link, link). But is this perhaps a misleading ontology? Would we be better served by thinking of the social world as "flat" -- involving processes and relations all at the same level? It sometimes appears that John Levi Martin has such an ontology in mind in Social Structures (link), and Doug Porpora envisions such a possibility in "Four Concepts of Structure" (link). So this is the idea I'd like to explore here.

What would that flat world look like? Here is one effort at formulating a flat social ontology.

The social world exists as the embodiment of sets of individual persons with powers, capacities, and actions and interactions, and who stand in a vast range of concrete social relationships with each other.

Here is a snippet from Porpora's article about structure mentioned above that seems to have this view in mind:

In contrast with the previous conception of social structure, this one is not a version of sociological holism. It does not portray social structure as something that operates over the heads of human actors. Instead, social structure is a nexus of connections among them, causally affecting their actions and in turn causally affected by them. The causal affects of the structure on individuals are manifested in certain structured interests, resources, powers, constraints and predicaments that are built into each position by the web of relationships. These comprise the material circumstances in which people must act and which motivate them to act in certain ways. As they do so, they alter the relationships that bind them in both intended and unintended ways. (200)

What does this description leave out? For starters, it leaves out things we would have said were higher levels of the social reality: families, organizations, social movements, institutions, economies, clubs, and states. And of course these are legitimate social constructs. But are they inherently "higher level"? Or are they compounds and extended aggregates of the lower-level stuff just mentioned -- individuals with powers, actions, and relations?

Playing this idea out, we might consider that a social movement is a partially ordered group of individuals in association with each other. The organizations that call them forward are other groups of individuals, including their deliberative bodies and executives. The repressive organs of the state? -- yet other organized groups of individuals with powers and agency. And in fact the theory of strategic action fields seems to lean in this direction (Fligstein and McAdam, A Theory of Fields; link).

One important consideration that might come forward for rejecting the flat ontology is the idea that there are causal properties at a higher level that don't attach to entities at the base level.

This view corresponds to the idea some sociologists have of emergence. It is sometimes maintained that social structures have properties at the structural level that cannot be reduced to the properties of the components of the structure. These are emergent properties. If this is so, then we will miss important explanations if we decline to recognize the reality of social structures. And yet a social structure is plainly a higher-level social entity than a group of coordinated individuals. Its higher-level standing is a result of this fact: it is composed of objects at level 1; but it has properties that cannot be explained or derived from objects at level 1.

A related reason for rejecting the flat ontology is the idea that structures, institutions, or value systems -- higher level social things -- may have legitimate causal properties that can be adequately discovered through study of these social things without more information about the base level (individuals in relations). This possibility doesn't necessarily imply that these are emergent properties, only that they are relatively autonomous from the base level. Here again, it seems reasonable to call these higher-level social entities -- and therefore the flat ontology isn't quite enough.

Another important consideration is the evident fact that social compounds have compositional structure. A fish is more than a collection of living cells; it has a stable structure and an internal organization that serves the needs of the fish organism. So it is entirely appropriate to refer to fish as well as living cells. And it seems correct to observe that something like this is true of some social entities as well -- government agencies, worship organizations, corporations.

Finally, it is hard to dispute that social things like kinship systems, business firms, and armies have stable and knowable characteristics that can be studied empirically. We shouldn't adopt an ontology that excludes legitimate topics for empirical research.

So it seems that the parsimonious social ontology doesn't work. It forces us to overlook explanatory factors that are important for explaining social outcomes. And it unreasonably asks us to ignore important features of the social world of which we have reasonably good understanding. In fact, the flat ontology is not far removed from the ontology associated with spare versions of methodological individualism.

So how might a bounded conception of higher-level social entities look? A formulation of a minimal multi-layer alternative to the flat ontology might go along these lines:

• The social world consists of individuals and relations at the base level PLUS stable compounds of items at this level which have quasi-permanent properties and non-reducible causal powers that have effects on items at the base level.

Here the criterion of higher-level standing in use is --

• possession of causal properties not reducible to [or needing reduction to] properties at the base level.

By analogous reasoning, we might consider whether there are more complex configurations of base and level 1 entities which themselves have properties that are emergent from or autonomous from base and level 1. And so forth iteratively.

Are there level 2 entities by this criterion? For example, might the state be a level-2 entity, in that it encompasses organizations and individuals and and it possesses new causal properties not present at level 1? In principle this seems possible. The state is a complex network of organizations and individuals. And it is logically possible that new causal powers emerge that depend on both base and level 1, but that do not require reduction to those lower-level properties.

So the language of levels of the social appears to be legitimate after all. It gives us a conceptual vocabulary that captures composition and complexity, and it allows us to identify important social causal powers that would not be accessible to us on the flat ontology.

Niiniluoto on scientific realism

Debates over realism have been at the center of the philosophy of science for at least seventy-five years. The fundamental question is this: what exists in the world? And how do we best gain knowledge about the nature and properties of these real things? The first question is metaphysical, while the second is epistemic.

Scientific realism is the view that “mature” areas of science offer theories of the nature of real things and their properties, and that the theories of well-confirmed areas of science are most likely approximately true. So science provides knowledge about reality independent from our ideas; and the methods of science justify our belief in these representations of the real world. Scientific methods are superior to other forms of belief acquisition when it comes to successful discovery of the entities and properties of the world in which we live.

But this statement conceals a number of difficult issues. What is involved in asserting that a theory is true? We have the correspondence theory of truth on the one hand — the idea that the key concepts of the theory succeed in referring to real entities in the world independent of the theory. And on the other hand, we have the pragmatist theory of truth — the idea that “truth” means “well confirmed”. A further difficulty arises from the indisputable fallibility of science; we know that many well confirmed scientific theories have turned out to be false. Finally, the idea of “approximate truth” is problematic, since it seems to imply “not exactly true,” which in turn implies “false”. Hilary Putnam distinguished two kinds of realism based on the polarity of correspondence and justification, metaphysical realism and internal realism; and it seems plain enough that “internal realism” is not a variety of realism at all.

Another central issue in the metatheory of realism is the question, what kinds of considerations are available to permit us to justify or refute various claims of realism? Why should we believe that the contents of current scientific theories succeed in accurately describing unobservable but fundamental features of an independent world? And the strongest argument the literature has produced is that offered by Putnam and Boyd in the 1970s: the best explanation of the practical and predictive successes of the sciences is the truth of the theoretical assumptions on which they rest.

Ilkka Niiniluoto’s 1999 Critical Scientific Realism proceeds from the general orientation of Roy Bhaskar’s critical realism. But it is not a synthesis of the philosophy of critical realism as much as it is an analytical dissection of the logic and plausibility of various claims of scientific realism. As such it is an excellent and rigorous introduction to the topic of scientific realism for current discussions. Niiniluoto analyzes the metatheory of realism into six areas of questions: ontological, semantical, epistemological, axiological, methodological, and ethical (2). And he provides careful and extensive discussions of the issues that arise under each topic. Here is a useful taxonomy that he provides for the many variants of realism (11):

Here is how Niiniluoto distinguishes “critical scientific realism” from other varieties of realism:

• R0: At least part of reality is ontologically independent of human minds.
• R1: Truth is a semantical relation between language and reality (correspondence theory).
• R2: Truth and falsity are in principle applicable to all linguistic products of scientific enquiry.
• R3: Truth is an essential aim of science.
• R4: Truth is not easily accessible or recognizable, and even our best theories can fail to be true.
• R5: The best explanation for the practical success of science is the assumption that scientific theories in fact are approximately true.

These are credible and appealing premises. And they serve to distinguish this version of realism from other important alternatives -- for example, Putnam's internal realism. But it is evident that Niiniluoto's "critical scientific realism" is not simply a further expression of "critical realism" in the system of Bhaskar. It is a distinctive and plausible version of scientific realism; but its premises equally capture the realisms of other philosophers of science whose work is not within the paradigm of standard critical realism. As the diagram indicates, other philosophers who embrace R0-R5 include Popper, Sellars, Bunge, Boyd, and Nowak, as well as Niiniluoto himself. (It is noteworthy that Bhaskar's name does not appear on this list!)

So how much of a contribution does Critical Scientific Realism represent in the evolving theory of scientific realism within philosophy of science? In my reading this is an important step in the evolution of the arguments for and against realism. Niiniluoto's contribution is a synthetic one. He does an excellent job of tracing down the various assumptions and disagreements that exist within the field of realism and anti-realism debates, and the route that he traces through these debates under the banner of "critical scientific realism" represents (for me, anyway) a particularly plausible combination of answers to these various questions. So one might say that the position that Niiniluoto endorses is a high point in the theory of scientific realism -- the most intellectually and practically compelling combination of positions from metaphysics, epistemology, semantics, and methodology that are available in the assessment of the truthiness of science.

What it is not, however, is the apotheosis of "critical realism" in the sense intended by the literature extending from Bhaskar to the current generation of critical realist thinkers. Niiniluoto's approach is appealingly eclectic; he follows the logic of the arguments he entertains, rather than seeking to validate or extend a particular view within this complicated field of realist arguments. And this is a good thing if our interest is in making the most sense possible of the idea of scientific realism as an interpretation of the significance of science in face of the challenges of constructivism, conceptual and theoretical underdetermination, and relativism.

Critical realism and social heterogeneity

Is the metaphysics of critical realism compatible with the idea of a highly heterogeneous social world?

Here is what I mean by heterogeneity in this context. First social causation is inherently multiple, with many kinds and tempos of social causation at work. It is therefore crucial that we avoid the impulse to reduce social change to a single set of underlying causal factors. The occurrence of a race riot at a time and place is partly caused by the instigating incident, partly caused by the long-simmering background conditions, partly caused by the physical geography of the city in question and partly caused by a legal and political context far from the site of rioting. We sometimes describe this fact as the conjunctural nature of social causation. Second, social events, changes, and forms of stability depend on contingent alignments of forces and causes, which do not recur in regular sequences of Humean causation. Third, social causes are generally historically conditioned, with the result that we do not have a general statement of, same cause, same effect. I characterize these points by saying that social causation is contingent, contextual, and conjunctural.

Another important aspect of heterogeneity in the social world has to do with the status of social kinds or social types. I take the view that social entities do not constitute social kinds, in that there is substantial and deep variation across the instances of items which we classify under riot, revolution, or state. Another way to put this point is to observe that social things do not have essential natures. Being Muslim is not an essential social or cultural or religious identity. Being a late industrial city is not an essential characteristic of a group of cities. Being a social revolution is not an essential underlying set of characteristics of the Chinese, French, and Russian episodes. Rather, in each of these examples there is broad variation across the instances that are embraced by the term.

So my question here is a simple one. Is Bhaskar's version of realism consistent with this treatment of heterogeneous social entities and heterogeneous social causes, or does Bhaskar presuppose social essences and universal causes in ways that are inconsistent with heterogeneity?

There are elements Bhaskar's theory that point in both directions on this question.

His emphasis on the logic of experimentation is key to his transcendental argument for realism. But oddly enough, this analysis cuts against the premise of heterogeneity because it emphasizes exceptionless causal factors. He emphasizes the necessity of postulating underlying causal laws, which are themselves supported by generative causal mechanisms, and the implication is that the natural world unfolds as the expression of these generative mechanisms. Here is a clear statement from The Possibility of Naturalism: A philosophical critique of the contemporary human sciences:

Once made, however, the ontological distinction between causal laws and patterns of events allows us to sustain the universality of the former in the face of the non-invariance of the latter. Moreover, the actualist analysis of laws now loses all plausibility. For the non-invariance of conjunctions is a condition of an empirical science and the non-empirical nature of laws a condition of an applied one. (PON p. 11)

And his account sometimes seems to rest upon a kind of "mechanism fundamentalism" -- the idea that there is a finite set of non-reducible mechanisms with essential properties:

On the transcendental realist system a sequence A, B is necessary if and only if there is a natural mechanism M such that when stimulated by A, B tends to be produced. (PON p. 11)

Concerns about mechanisms fundamentalism are allayed, however, because Bhaskar notes that it is always open to the scientist to ask the new question, how does this mechanism work? (PON 13) So mechanisms are not irreducible.

These are a few indications that Bhaskar's realism might be uncongenial to the idea of social heterogeneity.

More compelling considerations are to be found on the other side of the issue, however. First, his introduction of the idea of the social world as an "open" system of causation leaves space for causal heterogeneity. Here is a relevant passage from A Realist Theory of Science, deriving from an example of historical explanation:

In general as a complex event it will require a degree of what might be called 'causal analysis', i.e. the resolution of the event into its components (as in the case above). (RTS kl 2605)

For the different levels that mesh together in the generation of an event need not, and will not normally, be typologically locatable within the structures of a single theory. In general the normic statements of several distinct sciences, speaking perhaps of radically different kinds of generative mechanism, may be involved in the explanation of the event. This does not reflect any failure of science, but the complexity of things and the multiplicity of forms of determination found in the world. (RTS kl 2613)

Here is how Bhaskar conceives of social and historical things in The Possibility of Naturalism:

From this perspective, then, things are viewed as individuals possessing powers (and as agents as well as patients). And actions are the realization of their potentialities. Historical things are structured and differentiated (more or less unique) ensembles of tendencies, liabilities and powers; and historical events are their transformations. (PON 20)

The phrase "more or less unique" is crucial. It implies the kind of heterogeneity postulated here, reflecting the ideas of contingency and heterogeneity mentioned above.

Another reason for thinking Bhaskar is open to heterogeneity in the social realm is his position on reductionism.

But, it might be objected, is not the universe in the end nothing but a giant machine with inexorable laws of motion governing everything that happens within it? I want to say three things: First, that the various sciences treat the world as a network of 'machines', of various shapes and sizes and degrees of complexity, whose proper principles of explanation are not all of the same kind as, let alone reducible to, those of classical mechanics. Secondly, that the behaviour of 'machines', including classical mechanical ones, cannot be adequately described, let alone understood, in terms of the 'whenever x, then y' formula of regularity determinism. Thirdly, that even if the world were a single 'machine' this would still provide no grounds for the constant conjunction idea, or a fortiori any of the theories of science that depend upon it. Regularity determinism is a mistake, which has been disastrous for our understanding of science. (RTS kl 1590)

Here Bhaskar is explicit in referring to multiple kinds of causal processes ("machines"). And, indeed, Bhaskar affirms the conjunctural nature of social causation:

Now most social phenomena, like most natural events, are conjuncturally determined. And as such in general have to be explained in terms of a multiplicity of causes. (PON p. 54)

Similar ideas are expressed in Scientific Realism and Human Emancipation:

Social phenomena must be seen, in general, as the product of a multiplicity of causes, i.e. social events as 'conjunctures' and social things as (metaphysically) 'compounds'. (107)

Finally, his discussion of social structures in PON as the social equivalent of natural mechanisms also implies heterogeneity over time:

(3) Social structures, unlike natural structures, may be only relatively enduring (so that the tendencies they ground may not be universal in the sense of space-time invariant). (PON 49)

So on balance, I am inclined to think that Bhaskar's philosophy of social science is indeed receptive to social heterogeneity. And this in turn makes it a substantially more compelling contribution to the philosophy of social science than it would otherwise be, and superior to many of the positivist variants of philosophy of science that he criticizes.

Social construction of technical knowledge

After there was the sociology of knowledge (link), before there was a new sociology of knowledge (link), and more or less simultaneous with science and technology studies (link), there was Paul Rabinow's excellent ethnography of the invention of the key tool in recombinant DNA research -- PCR (polymerase chain reaction). Rabinow's monograph Making PCR: A Story of Biotechnology appeared in 1996, after the first fifteen years of the revolution in biotechnology, and it provides a profound narrative of the intertwinings of theoretical science, applied bench work, and material economic interests, leading to substantial but socially imprinted discoveries and the development of a powerful new technology. Here is how Rabinow frames the research:

Making PCR is an ethnographic account of the invention of PCR, the polymerase chain reaction (arguably the exemplary biotechnological invention to date), the milieu in which that invention took place (Cetus Corporation during the 1980s), and the key actors (scientists, technicians, and business people) who shaped the technology and the milieu and who were, in turn, shaped by them. (1)

This book focuses on the emergence of biotechnology, circa 1980, as a distinctive configuration of scientific, technical, cultural, social, economic, political, and legal elements, each of which had its own separate trajectory over the preceding decades. It examines the "style of life" or form of "life regulation" fashioned by the young scientists who chose to work in this new industry rather than pursue promising careers in the university world.... In sum, it shows how a contingently assembled practice emerged, composed of distinctive subjects, the site in which they worked, and the object they invented. (2)

There are several noteworthy features of these very exact descriptions of Rabinow's purposes. The work is ethnographic; it proceeds through careful observation, interaction, and documentation of the intentionality and practices of the participants in the process. It is focused on actors of different kinds -- scientists, lab technicians, lawyers, business executives, and others -- whose interests, practices, and goals are distinctly different from each others'. It is interested in accounting for how the "object" (PCR) came about, without any implication of technological or scientific inevitability. It highlights both contingency and heterogeneity in the process. The process of invention and development was a meandering one (contingency) and it involved a large group of heterogeneous influences (scientific, cultural, economic, ...).

Legal issues come into this account because the fundamental question -- what is PCR and who invented it? -- cannot be answered in narrowly technical or scientific terms. Instead, it was necessary to go through a process of practical bench-based development and patent law to finally be able to answer both questions.

A key part of Rabinow's ethnographic finding is that the social configuration and setting of the Cetus laboratory was itself a key part of the process leading to successful development of PCR. The fact of hierarchy in traditional scientific research spaces (universities) is common -- senior scientists at the top, junior technicians at the bottom. But Cetus had developed a local culture that was relatively un-hierarchical, and Rabinow believes this cultural feature was crucial to the success of the undertaking.

Cetus's organizational structure was less hierarchical and more interdisciplinary than that found in either corporate pharmaceutical or academic institutions. In a very short time younger scientists could take over major control of projects; there was neither the extended postdoc and tenure probationary period nor time-consuming academic activities such as committees, teaching, and advising to divert them from full-time research. (36)

And later:

Cetus had been run with a high degree of organizational flexibility during its first decade. The advantages of such flexibility were a generally good working environment and a large degree of autonomy for the scientists. The disadvantages were a continuing lack of overall direction that resulted in a dispersal of both financial and human resources and in continuing financial losses. (143)

A critical part of the successful development of PCR techniques in Rabinow's account was the highly skilled bench work of a group of lab technicians within the company (116 ff.). Ph.D. scientists and non-Ph.D. lab technicians collaborated well throughout the extended period during which the chemistry of PCR needed to be perfected; and Rabinow's suggestion is that neither group by itself could have succeeded.

So some key ingredients in this story are familiar from the current wisdom of tech companies like Google and FaceBook: let talented people follow their curiosity, use space (physical and social) to elicit strong positive collaboration; don't try to over-manage the process through a rigid authority structure.

But as Rabinow points out, Cetus was not an anarchic process of smart people discovering things. Priorities were established to govern research directions, and there were sustained efforts to align research productivity with revenue growth (almost always unsuccessful, it must be said). Here is Rabinow's concluding observation about the company and the knowledge environment:

Within a very short span of time some curious and wonderful reversals, orthogonal movements, began happening: the concept itself became an experimental system; the experimental system became a technique; the techniques became concepts. These rapidly developing variations and mutually referential changes of level were integrated into a research milieu, first at Cetus, then in other places, then, soon, in very many other places. These places began to resemble each other because people were building them to do so, but were often not identical. (169).

And, as other knowledge-intensive businesses from Visicalc to Xerox to H-P to Microsoft to Google have discovered, there is no magic formula for joining technical and scientific research to business success.

Microfoundations 2.0?

I 

Figure. An orderly ontological hierarchy (University of Leeds (link)

Figure. Complex non-reductionist social outcome -- blight

The idea that hypotheses about social structures and forces require microfoundations has been around for at least 40 years. Maarten Janssen’s New Palgrave essay on microfoundations documents the history of the concept in economics; link. E. Roy Weintraub was among the first to emphasize the term within economics, with his 1979 Microfoundations: The Compatibility of Microeconomics and Macroeconomics. During the early 1980s the contributors to analytical Marxism used the idea to attempt to give greater grip to some of Marx's key explanations (falling rate of profit, industrial reserve army, tendency towards crisis). Several such strategies are represented in John Roemer's Analytical Marxism. My own The Scientific Marx (1986) and Varieties of Social Explanation (1991) took up the topic in detail and relied on it as a basic tenet of social research strategy. The concept is strongly compatible with Jon Elster's approach to social explanation in Nuts and Bolts for the Social Sciences (1989), though the term itself does not appear  in this book or in the 2007 revised edition.

Here is Janssen's description in the New Palgrave of the idea of microfoundations in economics:

The quest to understand microfoundations is an effort to understand aggregate economic phenomena in terms of the behavior of individual economic entities and their interactions. These interactions can involve both market and non-market interactions.  

In The Scientific Marx the idea was formulated along these lines:

Marxist social scientists have recently argued, however, that macro-explanations stand in need of microfoundations; detailed accounts of the pathways by which macro-level social patterns come about. (1986: 127)

The requirement of microfoundations is both metaphysical -- our statements about the social world need to admit of microfoundations -- and methodological -- it suggests a research strategy along the lines of Coleman's boat (link). This is a strategy of disaggregation, a "dissecting" strategy, and a non-threatening strategy of reduction. (I am thinking here of the very sensible ideas about the scientific status of reduction advanced in William Wimsatt's "Reductive Explanation: A Functional Account"; link).

The emphasis on the need for microfoundations is a very logical implication of the position of "ontological individualism" -- the idea that social entities and powers depend upon facts about individual actors in social interactions and nothing else. (My own version of this idea is the notion of methodological localism; link.) It is unsupportable to postulate disembodied social entities, powers, or properties for which we cannot imagine an individual-level substrate. So it is natural to infer that claims about social entities need to be accompanied in some fashion by an account of how they are embodied at the individual level; and this is a call for microfoundations. (As noted in an earlier post, Brian Epstein has mounted a very challenging argument against ontological individualism; link.)

Another reason that the microfoundations idea is appealing is that it is a very natural way of formulating a core scientific question about the social world: "How does it work?" To provide microfoundations for a high-level social process or structure (for example, the falling rate of profit), we are looking for a set of mechanisms at the level of a set of actors within a set of social arrangements that result in the observed social-level fact. A call for microfoundations is a call for mechanisms at a lower level, answering the question, "How does this process work?"

In fact, the demand for microfoundations appears to be analogous to the question, why is glass transparent? We want to know what it is about the substrate at the individual level that constitutes the macro-fact of glass transmitting light. Organization type A is prone to normal accidents. What is it about the circumstances and actions of individuals in A-organizations that increases the likelihood of normal accidents?

One reason why the microfoundations concept was specifically appealing in application to Marx's social theories in the 1970s was the fact that great advances were being made in the field of collective action theory. Then-current interpretations of Marx's theories were couched at a highly structural level; but it seemed clear that it was necessary to identify the processes through which class interest, class conflict, ideologies, or states emerged in concrete terms at the individual level. (This is one reason I found E. P. Thompson's The Making of the English Working Class (1966) so enlightening.) Advances in game theory (assurance games, prisoners' dilemmas), Mancur Olson's demonstration of the gap between group interest and individual interest in The Logic of Collective Action: Public Goods and the Theory of Groups (1965), Thomas Schelling's brilliant unpacking of puzzling collective behavior onto underlying individual behavior in Micromotives and Macrobehavior (1978), Russell Hardin's further exposition of collective action problems in Collective Action (1982), and Robert Axelrod's discovery of the underlying individual behaviors that produce cooperation in The Evolution of Cooperation (1984) provided social scientists with new tools for reconstructing complex collective phenomena based on simple assumptions about individual actors. These were very concrete analytical resources that promised help further explanations of complex social behavior. They provided a degree of confidence that important sociological questions could be addressed using a microfoundations framework.

There are several important recent challenges to aspects of the microfoundations approach, however.

First, there is the idea that social properties are sometimes emergent in a strong sense: not derivable from facts about the components. This would seem to imply that microfoundations are not possible for such properties.

Second, there is the idea that some meso entities have stable causal properties that do not require explicit microfoundations in order to be scientifically useful. (An example would be Perrow's claim that certain forms of organizations are more conducive to normal accidents than others.) If we take this idea very seriously, then perhaps microfoundations are not crucial in such theories.

Third, there is the idea that meso entities may sometimes exert downward causation: they may influence events in the substrate which in turn influence other meso states, implying that there will be some meso-level outcomes for which there cannot be microfoundations exclusively located at the substrate level.

All of this implies that we need to take a fresh look at the theory of microfoundations. Is there a role for this concept in a research metaphysics in which only a very weak version of ontological individualism is postulated; where we give some degree of autonomy to meso-level causes; where we countenance either a weak or strong claim of emergence; and where we admit of full downward causation from some meso-level structures to patterns of individual behavior?

In one sense my own thinking about microfoundations has already incorporated some of these concerns; I've arrived at "microfoundations 1.1" in my own formulations. In particular, I have put aside the idea that explanations must incorporate microfoundations and instead embraced the weaker requirement of availability of microfoundations (link). Essentially I relaxed the requirement to stipulate only that we must be confident that microfoundations exist, without actually producing them. And I've relied on the idea of "relative explanatory autonomy" to excuse the sociologist from the need to reproduce the microfoundations underlying the claim he or she advances (link).

But is this enough? There are weaker positions that could serve to replace the MF thesis. For now, the question is this: does the concept of microfoundations continue to do important work in the meta-theory of the social sciences?