Meso powers and causal mechanisms

I've argued in earlier posts for several ideas:

1. Meso-level social structures have causal powers.
2. Meso-level social structures must have microfoundations.
3. Causal relations are carried out by causal mechanisms.

Are these claims consistent with each other?  An earlier post argues for the consistency of (1) and (2).  Here I will address the relation between (1) and (3): what kinds of causal mechanisms are available to mediate the causal powers of meso-level structures? Are there meso-level mechanisms as well as meso-level causes?  Or are the mechanisms that mediate meso causal powers themselves necessarily located at the micro level?

It is evident that there are micro-level mechanisms for meso-level causal powers. The mechanisms of a run on the bank or a self-fulfilling prophecy are both micro-level; they show how a meso-level event disaggregates onto the beliefs and actions of the individuals involved, leading to collective behavior that results in the meso-level effect. These are This is exactly the kinds of mechanism that analytical sociologists, including Thomas Schelling, want to be the exclusive kind of causation in the social realm. So the hard question is this: Are there meso-level causal mechanisms?

I addressed one aspect of this question in an earlier post, where I considered whether causal mechanisms can be complex and multi-staged.  There I focused on the idea that most theories of causal mechanisms require that there be a reasonably strong regularity between input and output.  This implies that a multi-step causal process will probably not qualify as a mechanism, because the intervening causal steps will imply a low association between initial condition and output condition.  (See the post for more details.)  And this in turn implies that a social mechanism needs to consist of one step: the instigating condition occurs and then the output condition occurs. The diagram illustrates the point. Scenario I represents a single-step causal mechanism; scenario II represents a three-step causal process composed of three component mechanisms.  Scenario II does not qualify as a mechanism itself, because the compound probability of D given A is only 50%.

 

So this would seem to answer the question for the case in which the putative meso-level causal mechanism involves multiple steps of causation. For example: {[Distracting work environment] => [low productivity] => [low profits] => [downsizing]} is a multiple-step causal process entirely at the meso level; but it doesn't establish a meso-level mechanism {[Distracting work environment] => [downsizing]} because the intervening steps imply new contingencies and a subsequent low conditional probability between antecedent and consequent.

There is another possible case to consider as well, however: the case in which a given meso-level structure or occurrence itself gives rise with high probability to a given meso-level result.  Are there such examples?

It would appear that there are.  Social isolation (referred to in an earlier post) is a meso-level characteristic of a population.  And Alford Young argues that decreasing social isolation causes rising inter-group hostility, another meso-level characteristic of a population.  This seems like a clear and uncontroversial example of a meso-meso causal link that is also a meso-level causal mechanism. It looks like Scenario I above.  Moreover, this mechanism can be employed in explanations of more complex social events like race rebellions.  It would seem that there are similar examples in the work of Robert Sampson on neighborhood effects (link).

And here is a third possibility. A meso-level structure S's causal powers may result from the causal powers of its component systems Ti, and the causal mechanisms that embody the causal powers of S may be found at the level of mechanisms at the level of Ti.  For example, a certain type of police organization S may have a high level of corruption C among its street officers.  So S has the causal power to bring about C. The mechanisms that establish this power may reside at a lower level of organization: the processes for giving assignments and the processes for checking incident reports. The diagram might look something like this.

The dashed line indicates the causal power. The solid arc lines on the left indicate "composition". The solid arrows indicate meso-meso causal mechanisms.  And the dotted lines indicate the aggregate consequences of "poor orientation to goals" and "weak incentives for conformance", which is an elevated level of officer corruption. So the meso-level power depends on sub-meso level mechanisms and actor-level mechanisms.

So the conclusion I'd like to tentatively offer is this: Meso-level causal powers are created as a consequence of specific causal mechanisms triggered by the properties and changes of the meso-level structure. Those mechanisms may be at the actor level; they may be at a sub-meso level; or they may be at the meso level.

Organizational failure as a meso cause

A recurring issue in the past few months here has been the validity of meso-level causal explanations of social phenomena. It is self-evident that we attribute causal powers to meso entities in ordinary interactions with the social world. We assent to statements like these; they make sense to us.

• Reorganization of the city's service departments led to less corruption in Chicago.
• Poor oversight and a culture of hyper-masculinity encourages sexual harassment in the workplace.
• Divided command and control of military forces leads to ineffective response to attack.
• Mining culture is conducive to social resistance.

We can gain some clarity on the role played by appeals to meso-level factors by considering a concrete example in detail. Military failure is a particularly interesting example to consider. Warfare proceeds through complex interlocking social organizations; it depends on information collection and interpretation; it requires the coordination of sometimes independent decision-makers; it involves deliberate antagonists striving deliberately to interfere with the strategies of the other; and it often leads to striking and almost incomprehensible failures. Eliot Cohen and John Gooch's Military Misfortunes: The Anatomy of Failure in War is a highly interesting study of military failure that makes substantial use of organizational sociology and the sociology of failure more broadly, so it provides a valuable example to consider.

Here are a few framing ideas that guide Cohen and Gooch in their analysis and selection of cases.

True military "misfortunes" -- as we define them -- can never be justly laid at the door of any one commander. They are failures of the organization, not of the individual. The other thing the failures we shall examine have in common is their apparently puzzling nature. Although something has clearly gone wrong, it is hard to see what; rather, it seems that fortune -- evenly balanced between both sides at the outset -- has turned against one side and favored the other. These are the occasions when it seems that the outcome of the battle depended at least as much on one side's mishandling of the situation as on the other's skill in exploiting a position of superiority ... The causes of organizational failure in the military world are not easy to discern. (3)

From the start, then, Cohen and Gooch are setting their focus on a meso-level factor -- features of organizations and their interrelations within a complex temporally extended period of activity.  They note that historians often start with the commander -- the familiar explanation of failures based on "operator error" -- as the culprit.  But as they argue in the cases they consider, this effort is as misguided in the case of military disaster as it is in other kinds of failure.  Much more fundamental are the organizational failures and system interactions that led to the misfortune. Take Field Marshal Douglas Haig, whose obstinate commitment to aggressive offense in the situation of trench warfare has been bitterly criticized as block-headed:

Not only was the high command confronted by a novel environment; it was also imprisoned in a system that made it well-nigh impossible to meet the challenges of trench warfare. The submissive obedience of Haig's subordinates, which Forester took for blinkered ignorance and whole-hearted support, was in reality the unavoidable consequence of the way in which the army high command functioned as an organization under its commander in chief. (13)

So why are organizations so central to the explanation of military failure?

Wherever people come together to carry out purposeful activity, organizations spring into being. The more complex and demanding the task, the more ordered and integrated the organization. ... Men form organizations, but they also work with systems. Whenever technological components are linked together in order to carry out a particular scientific or technological activity, the possibility exists that the normal sequence of events the system has been designed to carry out may go awry when failures in two or more components interact in an unexpected way. (21, 22)

And here is the crucial point: organizations and complexes of organizations (systems) have characteristics that sometimes produce features of coordinated action that are both unexpected and undesired.  A certain way of training officers may have been created in order to facilitate unity in combat; but it may also create a mentality of subordinacy that makes it difficult for officers to take appropriate independent action.  A certain system for managing the flow of materiel to the fighting forces may work fine in the leisurely circumstances of peace but quickly overload under the exigencies of war.  Weapon systems designed for central Europe may prove unreliable in North Africa.

Eliot and Gooch place organizational learning and information processing at the core of their theories of military failure.  They identify three kinds of failure: "failure to learn, failure to anticipate, and failure to adapt" (26). As a failure to learn, they cite the US Army's failure to learn from the French experience in Vietnam before designing its own strategies in the Vietnam War.  And they emphasize the unexpected interactions that can occur between different components of a complex organization like the military.  They recommend a "layered" analysis: "We look for the interactions between these organizations, as well as assess how well they performed their proper tasks and missions" (52).

The cases they consider correspond to this classification of failure.  They examine failure to learn in the case of American antisubmarine warfare in 1942; failure to anticipate in the case of the Israel Defense Forces' failure on the Golan Heights, 1973; and failure to adapt in the British disaster at Gallipoli, 1915.  Their example of aggregate failure, involving all three failures, is the defeat of the American Eighth Army in Korea, 1950.  And they reserve the grand prize, catastrophic failure, for the collapse of the French army and air force in 1940.

Each of these cases illustrates the authors' central thesis: that organizational failures are at the heart of many or most large military failures.  The example of the failure of the American antisubmarine campaign in 1942 off the east coast of the United States is particularly clear.  German submarines preyed at will on American shipping, placing a large question mark over the ability of the Americans to continue to resupply Allied forces in Europe.  The failure of American antisubmarine warfare was perplexing because the British navy had already developed proven and effective methods of ASW, and the American navy was aware of those methods.  Unhappily, Eliot and Gooch argue, the US Navy did not succeed in learning from this detailed wartime experience.

The factors that Eliot and Gooch cite include: insufficient patrol boats early in the campaign: insufficient training for pilots and patrol vessel crews; crucial failures on operational intelligence ("collection, organization, interpretation and dissemination of many different kinds of data"; 75); and, most crucially, organizational failures.

A prompt and accurate intelligence assessment would mean nothing if the analysts could not communicate that assessment directly to commanders on the scene, if those commanders did not have operational control over the various air and naval assets they required to protect shipping and sink U-boats, if they saw no reason to heed that intelligence, or if they had no firm notion of what to do about it. The working out of correct standard tactics ... could have no impact if destroyer skippers did not know or would not apply them. Moreover, as the U-boats changed their tactics and equipment ..., the antisubmarine forces needed to adopt compensating tactical changes and technological innovation. (76)

This contrasts with the British case:

The British system worked because it had developed an organizational structure that enabled the Royal Navy and RAF to make use of all of the intelligence at their disposal, to analyze it swiftly and accurately, and to disseminate it immediately to those who needed to have it. (77)

So why did the US Navy not adopt the British system of organization?  Here too the authors find organizational causes:

If the United States Navy had thought seriously about adapting its organization to the challenges of ASW in a fashion similar to that chosen by the British, it would have required major changes in how existing organizations operated, and in no case would this have been more true than that of intelligence. (89)

So in this case, Eliot and Gooch find that the failure of the US Navy to conduct effective antisubmarine warfare off the coast of the United States resides centrally in features of the organization of the Navy itself, and in collateral organizations like the intelligence service.

This is a good example of an effort to explain concrete and complex historical outcomes using the theoretical resources of organizational sociology.  And the language of causation is woven throughout the narrative.  The authors make a credible case for the view that the organizational features they identify caused (or contributed to the cause) of the large instances of military failure they identify.

 

Are there meso-level causes?

Social structures and other social "things" are ontologically peculiar in some ways. Most especially, they are abstract, distributed, and non-material. We can't put a culturally dominant food aversion or a group prejudice in a box and weigh it. And yet many of us want to say that social structures are "real", not merely theoretical constructs.

One important aspect of something's being real is that it has causal powers: the specific properties of the thing bring about differences in the world on the behavior of other things. This is a version of the interventionist theory of causation associated with Jim Woodward: change something about C and you bring about a change in E (link).

So let's consider this question: Can meso-level social structures have meso-level effects?

Of course meso-level structures have effects -- on individuals. The fact that there are laws and enforcement mechanisms governing highway speed has some effect on drivers' behavior. The question here is whether it is legitimate to postulate causal powers for structures whose effects are realized in other meso-level structures. And I want to explore the affirmative answer to the question: it is legitimate and coherent to assert meso-meso causal interactions, and we sometimes have empirical evidence to support such assertions.

(It would be possible, of course, to take the view that social structures are epiphenomenal and have no causal properties whatsoever. On that approach, what seems to be the effect of the legal system on individual behavior is really just the aggregate effect of the many individuals involved in the legal system. I don't find this view at all compelling, however.)

My question is relevant to two groups of sociological theorists, each of whom thinks the answer is trivial and obvious -- but in opposite directions. The new methodological individualists, represented by analytical sociology, think the answer is trivially "no", because social causation proceeds always and exclusively through actions and interactions of individuals (link). This is the fundamental idea underlying Coleman's Boat as a model of the relationship between macro and micro.  And a range of anti-individualists -- Giddens, Elder-Vass, Archer -- believe it is self evident from everyday experience that causal structures do have causal powers, and that it is a waste of time to defend the notion (link). It is obvious.

My position is a precarious one. On the one hand I advocate an actor-centered approach to sociology and the social sciences. I defend the idea that social claims need microfoundations in a specific (weak) sense. And on the other hand I believe that structures have a degree of stability that permits us to couch causal claims in terms of those structures directly, rather than needing to supplement those claims with disaggregated foundations at the level of the individual. So I argue for the idea that we can sometimes regard causal powers of social entities as "relatively autonomous" from individual-level facts.

By meso-level structures I mean to refer to things like these:

• National Science Foundation
• Nuclear Regulatory Commission
• IBM corporation
• AFL-CIO
• German paramilitary organizations 1930
• German ideology of cultural despair 1910
• Islamic norms of Zakat
• ...

In each case there are numerous actors assigned to roles, governed by rules defining their activities, and leading to a certain kind of functioning in the broader social environment.

Generically I would define a meso-level structure as --

A composite of individuals and roles that incorporates a set of rules and norms for internal and external actors, and that possesses procedures of inculcation and enforcement through which internal and external actors are brought to comply with the rules and norms (to some degree).

I would define a normative system as ...

a set of rules, norms, and expectations embodied in a population of actors and meeting a threshold level of success in coordinating and constraining behavior.

We have a number of sociological concepts that capture social items at this level: organization, bureaucracy, institution, normative community, social network, communications system, legal system, civil war, military coup, advocacy group.

It is evident that social entities often incorporate elements of several of these kinds of things. Organizations and structures often incorporate or depend upon normative systems, and normative systems often generate organizations and institutions that convey their impact to the young and adult actors.

What about other mid-level social nouns -- ethnic group, electorate, financial crisis, ...? These strike me as being compounds of a miscellaneous set of social things -- there are bits of organizations, normative systems, affinity groups, and social networks in each of them. The concept of assemblage seems to fit these nouns well.

The "meso" qualifier is a bit more difficult to specify. It is intended to focus our attention on mid-level social arrangements, between actors and global institutions like the US state, global Islam, and the world trading system. The intuitive idea is straightforward. These are the smaller-scale, lower-level social arrangements or units of which macro structures are composed. Bert Leuridan makes an effort to offer a more specific definition based on causal roles.

So I have a simple but important question in mind here. Is it ever legitimate to assert something like this:

• Meso structure X produced changes in meso structure Y,

without being obliged to demonstrate the individual-level pathways through which this effect is thought to have come about? Is a type 4 causal claim ever supportable (link)?

The question I am posing is related to the idea of the methodological individualism associated with James Coleman. Basically the idea propounded by Coleman and more recently by the analytical sociologists is that all social properties, including causal powers, work through the activities of individuals, and we need ideally to replace claims that appear to attribute causal powers to structures with theories that disaggregate these powers onto the patterned activities of individuals. This is a reductionist theory.

Other theorists, notably Dave Elder-Vass, want to assert that social structures have "emergent" properties and powers (link). An emergent property according to E-V, is one that is possessed by the aggregate but not by the composing units. On this account, there are causal properties of structures that cannot be represented as the aggregate effect of individual actors.

My own approach depends on a line of reasoning long familiar in the special sciences. It is anti-reductionist, in that it denies that we need to derive higher-level properties from lower-level properties. It accepts the compositional ontology: social structures are composed of individual actors. But it asserts explanatory autonomy for theoretical statements about mid-level mechanisms (link, link, link).

One particularly direct way of supporting the idea that structures have meso effects is to establish correlations at that level for a few examples. But this isn't the only way we establish causation in other areas of the sciences. We do experiments ("remove X and observe whether Y persists"), we analyze the outcomes of "natural" experiments, we do comparative studies, and we engage in process tracing of particular cases. We even engage in theoretical analysis to try to determine what causal powers a certain entity ought to be expected to have given its constitution.

So it seems that there is ample room for sociologists to assert and investigate the causal properties of social structures. And given appropriate attention to the principle of microfoundations, we have a social ontology that supports the legitimacy of such claims as well.

A related question is whether there are "mechanisms" that operate at the meso level, or whether all social mechanisms must operate at the individual level (as Hedstrom and the AS world believe).  I will return to this question.

 

Mechanisms and powers

source: William Bechtel, Discovering Cell Mechanisms: The Creation of Modern Cell Biology

The causal-powers approach to the understanding of causation is sometimes presented as an exclusive alternative to both traditional regularity theories and to more recent causal mechanism theories. In an earlier post I discussed Ruth Groff’s contributions to this topic. Here I would like to present a provocative view: that the causal mechanisms and causal powers are complementary rather than contradictory. The causal mechanisms theory benefits by being supplemented by a causal powers theory and the causal powers theory benefits by being supplemented by a causal mechanisms theory. In other words, the two theories are not exclusive alternatives to each other, but rather serve to identify different parts of the whole of causation.

The causal powers theory rests on the claim that causation is conveyed from cause to effect through the active powers and capacities that inhere in the entities making up the cause. The causal mechanisms theory comes down to the idea that cause and effect are mediated by a series of events or interactions that lead (typically) from the occurrence of the cause to the occurrence of the effect. In other words, cause and effect are linked by real underlying causal sequences (often repeatable sequences).

My thesis of the mutual compatibility of powers and mechanisms goes along these lines. If we press down on a putative mechanisms explanation, we are led eventually to postulating a set of causal powers that provide the motive force of the postulated mechanisms. But equally, if we press down on the claim that a certain kind of entity has a specified causal power or disposition, we are led to hypotheses about what mechanisms are set in play be its constituents so as to bring about this disposition.

Begin with a causal mechanism story:

• C => {x happens bringing about y, bringing about z, bringing about u, which is E} => E

How is it that the sub-links of this chain of mechanism pieces happen to work to bring about their consequent? We seem to have two choices: We can look to discover a further underlying mechanism; or we can postulate that the sub-link entity or structure has the power to bring about its consequent. So if we push downward within the terms of a mechanism explanation, one way to close the story is by postulating a causal power at some level.

Now start with a causal power claim. Suppose we assert that:

• Salt has the causal power of making H2O electrically conductive when dissolved.

Is this simply an unanalyzable fact about salt (or saline solution)? It is not; instead, we can look downward to identify the physical mechanisms that are brought into play when salt enters solution in H2O. That mechanism is well understood: the Na+ and Cl- ions created by the dissolution of salt permit free electrons to pass through the solution.

So we can explain the causal power by discovering the causal mechanism that gives rise to it; we explain links in the putative mechanism by alluding to the powers of the entities involved at that stage; and we can explain other things by referring to the causal powers that we have discovered to be associated with various kinds of things and structures.

If we take this set of possibilities seriously, then powers and mechanisms are answering different questions within the causal nexus. The reference to powers answers the question, “What does x do?”, while the reference to mechanisms answers the question, “How does x work?"

From a scientific point of view, it is always legitimate to ask how the powers of an entity or structure come to be in the natural world. What is it about the micro-structure of the thing in virtue of which the thing’s properties are established? In fact, this is one of the key intellectual challenges of the sciences. And this is a request for specification of some of the mechanisms that are at work. But likewise, it is always legitimate to ask what gives force to a given mechanism; and here we are eventually driven back to the answer, “some of the components of the mechanism have X, Y, Z powers to affect other entities” without further analysis within that particular explanation.

One might imagine that there are primitive causal powers — powers attached to primitive particles that have no underlying components or mechanisms.  We might begin to give a list of primitive causal powers: mechanical interactions among physical objects (transfer of momentum from one particle to another); electromagnetic properties inhering in one object and creating forces affecting other objects; gravitational forces among objects possessing mass; the causal interactions that occur within the central nervous system. And we might seek to demonstrate that all causal powers depend on combinations of these sorts of "primitive" causal powers -- a kind of Hobbesian materialism.  But this is needlessly strenuous from a metaphysical point of view. Better is to consider the middle-level range of powers and mechanisms where we are able to move upwards and downwards in our search for underlying causal mechanisms and supervening causal powers.

This line of thought suggests that questions about the metaphysics of causation are perhaps less pressing than they are sometimes made out to be. A thing's powers are not irreducible attributes of the thing; rather, they are the orderly consequence of the composition of the thing and the causal properties of those components and their interactions. It is hard to see that much turns on whether we think of the world as consisting of entities with powers, or as composites with system properties created by their components. The key question seems to be something like this: what is implied when we make a causal assertion? Both CP and CM agree that the core implication is the idea that one event, structure, or condition brought about the occurrence of another event, structure, or condition.  And the languages of both powers and mechanisms do a pretty good job of expressing what we mean in asserting this implication.

(John Dupré takes a similarly ecumenical view about several approaches to the theory of causation in a recent article, "Living Causes", where he advocates for what he calls "causal pluralism"; link. He writes: "I believe that causality is a complex and diverse set of phenomena, and most or all of these accounts provide valuable and complementary perspectives on the topic. Such a pluralistic view is quite a common one among contemporary philosophers; however there are significant differences in the form that such pluralisms can take" (20). On the mechanisms side within the philosophy of biology is William Bechtel's Discovering Cell Mechanisms: The Creation of Modern Cell Biology, who writes: "Beginning in the 1940s an initially small cadre of investigators who were pioneers in the modern discipline of cell biology began to figure out the biochemical mechanisms that enable cells to perform these functions. although miniaturized, the mechanisms they found to be operative in each cell are staggeringly complex" (1 ).)

Causality and metaphysics

Advocates of the causal powers approach attach a great deal of importance to the metaphysics of causation -- the sorts of properties and relations that we attribute to the kinds of things that we want to postulate. The neo-Aristotelian point of view represented by Ruth Groff and others appears to have metaphysical objections to the causal-mechanisms approach: the CM approach postulates the wrong kind of relations among entities, according to this group. So if I want to argue that mechanisms and powers are compatible, as I do, then I need to take into account the metaphysical arguments. It will be necessary to tell a story about the nature of the world that gives a place and meaning to the metaphysical premises of each theory.

The possibility of fundamental metaphysical incompatibility cannot be trivially ruled out. Consider this different kind of example: the distance between the premises of analytical Marxism and the neo-Hegelian theory of internal relations espoused by Bertell Ollman in Alienation: Marx's Conception of Man in Capitalist Society (link). Even if there were the possibility of some degree of convergence in conclusions about capitalism -- e.g. the likelihood of recurring crises -- the two schools of thought differ fundamentally on the nature of social entities and structures. They differ in terms of their social metaphysics. Analytical Marxists take the view that the structures of capitalism are the composite effect of variously motivated individuals; so there is an underlying atomism in the ontology of AM. Causes are fully distinct from the things they affect. Ollman, by contrast, believes that we need to conceive of the structures and social relations of capitalism relationally: the wage labor relation is not an atomistic relation between capitalist and worker, but rather a mutually implicating set of relations between the two that cannot be fully separated. Here is a passage in which Ollman attempts to capture the distinctive features of Marx's social metaphysics:

What is distinctive in Marx's conception of social reality is best approached through the cluster of qualities he ascribes to particular social factors. Taking capital as the example, we find Marx depicting it as "that kind of property which exploits wage-labor, and which cannot increase except on condition of getting a new supply of wage-labor for fresh exploitation" (Marx and Engels, 1945, 33). What requires emphasis is that the relation between capital and labor is treated here as a function of capital itself, and part of the meaning of "capital". This tie is extended to cover the worker as well, where Marx refers to him as "variable capital" (Marx, 1958, 209). The capitalist is incorporated into the same whole: "capital is necessarily at the same time the capitalist . . . the capitalist is contained in the concept of capital" (Marx, 1973, 512). Elsewhere, Marx asserts that "the means of production monopolized by a certain section of society", "the products of laborers turned into independent powers", "money", "commodities" and even "value that sucks up the value creating powers" are also capital (Marx, 1959b, 794-5; Marx, 1958, 153; Marx, 571). What emerges from these diverse characterizations is a conception of many tied facets, whose sense depends upon the relations Marx believes to exist between its components: property, wage-labor, worker, his product, commodities, means of production, capitalist, money, value (the list can be made longer still). (Chapter 2, section ii)

This example demonstrates the possibility of a genuine and deep incompatibility between two social theories at the level of the assumptions they make about the nature of the world -- their metaphysical theories.

So what about causal powers and causal mechanisms? The primary metaphysical commitment that the CP theorists advocate derives from their treatment of powers and essences -- two characteristic ideas from Aristotle. A power is thought to inhere in a thing in a particularly deep way; it is not an accidental expression of the empirical properties of the thing but rather an essential and active expression of the nature of the thing. The causal powers theory comes down to the idea that things and structures have an active capacity to bring about certain kinds of effects. In Groff’s terms, things are not passive but rather active.

Here is how Tuukka Kaidesoja introduces the metaphysical framework of critical realism in relation to causality in Naturalizing Critical Realist Social Ontology. Kaidesoja finds that the concept of a causal power is fundamental to critical realism (105). A thing's power is the characteristic of the world through which causal influence arises; without the concept of causal power, we would indeed be stranded in a Humean world of pure constant conjunction. Kaidesoja quotes Harre and Madden in these terms:

"X has the power to A" means "X (will)/(can) do A, in the appropriate conditions, in virtue of its intrinsic nature". (Kaidesoja, 106)

So what about the metaphysics of the causal mechanisms theory? Generally speaking, advocates of the mechanisms approach have not been very interested in the metaphysical issues. They (we) are generally realist, so we postulate that there are real causal interactions. This is indeed a metaphysical position. But this family of thinkers tends to be mid-range realists: they want to understand the necessity of causal relations at one level as deriving from the real workings of the physical or social system a bit lower down; but they generally don't seem to want to pose the ultimate question: how could any event or structure exert causal influence on another? So the causal mechanisms theorists are perhaps better described as scientific realists rather than philosophical or critical realists. They take the view that the world has the properties (approximately) that our best scientific theories attribute to things. (Could we call them "Galilean realists"?)

Curiously enough, this contrast seems to have a lot to do with the quibble I raised for Ruth Groff in the earlier post: whether powers should be thought to be "irreducible". Scientific realists would say they are not irreducible; rather, we can eventually arrive at a theory (molecular, genetic, economic, psychological, rational choice, physical) that displays the processes and mechanisms through which the ascribed power flows from the arrangement and properties of the thing.

On this standpoint, powers are attributions we make to things when we don't know quite enough about their composition to work out the physics (or sociology) of the underlying mechanisms. They do attach to the entity or structure in question, surely enough; but they do so in virtue of the physical or sociological composition of the entity, not because of some inherent metaphysical property.

We might try to reconcile these two perspectives with a few simple ideas:

1. Entities and structures at a range of levels of being have causal powers: active capacities to influence other entities and structures.
2. Whenever we identify a causal power of a thing, it is always open to us to ask how this power is embodied; what it is about the inner constitution of the entity that gives it this power.
3. When we succeed in arriving at a good scientific answer to this question, we will have shown that the power in question is not irreducible; it is rather the consequence of a set of mechanisms set in play by the constitution of the entity.

So the discovery of a given causal power of a thing is not a metaphysical fundamental; it is rather an empirical scientific discovery that invites analysis into its underlying composition.

Causal powers from a metaphysical point of view

 

A number of scholars who are interested in causation have recently expressed new interest in the concept of causal powers. This makes sense in a very straightforward and commonsensical way. But it also raises some difficult questions about metaphysics: how are we to think about the underlying nature of reality such that things, events, or conditions have "causal powers"? These questions raise issues that a number of talented philosophers are now taking on in a systematic way. Particularly interesting are recent writings by Ruth Groff, who represents a wave of contemporary thinking in metaphysics that aims to revitalize portions of Aristotle's views of causation in opposition to Hume's.

Groff's work on causal powers is sustained over a number of recent works, including especially her 2012 book Ontology Revisited: Metaphysics in Social and Political Philosophy (Ontological Explorations), her introduction and chapter in Greco and Groff, eds., Powers and Capacities in Philosophy: The New Aristotelianism, and her contribution to Illari, Russo, and Williamson, Causality in the Sciences. Groff emphasizes a broad clash of perspectives between a Humean theory of causation ("constant conjunction, no necessary relations among things or events") and a neo-Aristotelian theory ("things have powers, powers underlie causal relations among things and events"). Here is how she and Greco put the perspective of the "New Aristotelianism" in Powers and Capacities:

Humeanism is now under serious pressure within analytic metaphysics. In particular, after having been dismissed for generations as so much antiquated animism, the loosely-Aristotelian theorizing of real causal powers has now come to be a major focus of research within the specialty. (kl 204)

Moreover, Groff believes that American social sciences are still largely in the grip of the Humean metaphysics.  In "Getting Past Hume" in Causality in the Sciences, she writes:

One can't help but wonder what the outcome would actually be, were there to be a floor-fight on the question, i.e., a substantive debate within analytic philosophy and methodology of social science on the merits of Humean anti-realism about causality versus the merits of a powers-based, realist account of causality.

...

What is significant about all of this for my own argument is not so much that Humeanism continues to be the default ontology of especially American, often positivist, social science; but rather that it can be combined with the idea that it is not -- i.e., with the idea that competing versions of regularity theory somehow differ in a deep way, or that it is possible to remain neutral on what causality is, whilst engaging in causal explanation.

This poses a stark contrast; either you are positivist, anti-realist, and Humean or you are anti-positivist, realist, and neo-Aristotelian.  However, it is worth observing that there seem to be two currents of realist thought that reject Humean causation, not just one: the powers ontology that Groff (and Mumford and Anjum in Getting Causes from Powers) advocate; and the causal-mechanisms approach that has been advocated by philosophers and sociologists such as Hedstrom, Elster, and Ylikoski under the broad banner of analytical sociology. One might take the view that the causal-mechanisms approach ultimately requires something like the powers ontology -- "How else are we to account for the fact that sparks cause gasoline to explode?"; but on its face, these are two fairly independent realist responses to Hume. And certainly it is difficult to find a neo-Aristotelian predilection among the causal-mechanism advocates.

Groff takes up causal mechanisms theory in "Getting Past Hume" in  Causality in the Sciences. She concedes that this approach -- in the hands of Jon Elster and in my own writings, for example -- claims to be realist and anti-positivist, in that it rejects the notion that explanation depends on the discovery of general laws. But she doesn't think that the causal-mechanisms approach actually succeeds in presenting a substantive alternative to the Humean framework on causation: "Upon closer examination, the mainstream mechanisms model is more of the same, metaphysically."

Fleshing out her argument, she seems to be arguing that causal-mechanisms theory can either retreat to constant conjunction (at the level of the linkages of individual causal mechanisms) or it can press forward to a causal powers interpretation; there is no third possibility. "As with the other models, nothing on the mainstream mechanisms model is in a position of actually doing anything, in the sense of actively producing an effect. Thus here too, with an extra bit of ironic panache, the explanation-form functions as a delivery mechanism (no pun intended) for a Humean metaphysics." And it is true that most definitions of causal mechanisms make some kind of reference to regularities and repeatability.

My own formulation of the mechanisms theory is one of the targets of Groff's critique. And in fact I can reconstruct my reasons for thinking that mechanisms need to involve some kind of regularities; and I don't think it implies a collapse onto Humean causation. (At one point I wanted to call them "pocket regularities", to distinguish them from the grand social or psychological laws that Hempel and Mill seemed to want to discover.)  I wanted to assert that:

"M [information diffusion] is the mechanism connecting E [police beating] with O [rapid mobilization of an angry crowd]" is a description of a real underlying (perhaps unknown) causal process through which the features of E bring about the occurrence of O.

This is an ontological claim and it is a realist claim. But there is also an epistemic issue: How would we know that M is indeed such an underlying reality? It seems unavoidable that we would need to either produce empirical evidence supporting the conclusion that M frequently conveys these kinds of effects in these kinds of circumstances (the approach Tilly takes) or we need to have a theory of the mechanism which accounts for how it works to bring about the effect. The first boils down to a discovery of a limited set of regularities in a range of circumstances; the latter is a theoretical demonstration of how it works.  So this way of conceptualizing mechanisms does indeed invoke regularities of some sort.  However, it doesn't agree with the Humean idea that causation is nothing but regularities or constant conjunction. The regularities that are invoked are symptoms of the underlying causal mechanism, not criterial replacements for the mechanism.

Moreover, the powers theory seems to be subject to the same possible objection: how do we know that lightning has the causal power of starting barns on fire, unless we have repeatedly observed the chain of events leading from lightning strike to blaze?

Another thing that demands more attention is an assumption about the implications of the "realism" of powers. What follows from the idea that things have real causal powers? Groff puts the view in these terms: to assert that powers are real is to assert that they are irreducible (kl 204). But that seems questionable. We may think that feudalism was real, while at the same time thinking that its properties and dynamics derived from more fundamental social relations that compounded to create the distinctive dynamics of feudalism. So it doesn't seem that realists have to also accept the idea of irreducibility of the things about which they are realist. Or to put the point the other way around: the idea that realism implies irreducibility appears to also imply a fairly strong thesis about emergence. Groff returns to this set of ideas in Ontology Revisited, chapter three: "An emergent phenomenon (property or entity) is one that is not equivalent, ontologically, to the plurality of its parts." And here too she emphasizes irreducibility. But, as Poe Yu-ze Wan shows in "Emergence a la Systems Theory: Epistemological Totalausschluss or Ontological Novelty?" in Philosophy of the Social Sciences, philosophers have differed on the question of whether "emergence" implies "irreducibility" (link). The theory of emergence offered by Mario Bunge does not require irreducibility.

One thing I particularly like about Groff's work on causal powers is her persistence in working through the logical and conceptual implications of this field. She is painstaking in her effort to discover the implications of various parts of the several theories of causation (and freedom of the will in other essays); and this is exactly how we make progress on difficult philosophical issues like these.

Issues about microfoundations

I believe that hypotheses, theories, and explanations in the social sciences need to be subject to the requirement of microfoundationalism. This requirement can be understood in a weak and a strong version, and sometimes people understand the idea as a requirement of reductionism.  In brief, I defend the position in a weak form that does not imply a reductionist theory of social explanation. Recent discussions with Julie Zahle have led me to sharpen my understanding of the requirement of microfoundations in social theorizing and explanation. Here I would like to clarify my own thinking about the role and scope of the principle of microfoundationalism.

A microfoundation is something like this: an account of the mechanisms at the individual actor level (and perhaps at levels intermediate between actors and the current level -- e.g. institutions) that work to create the structural and causal properties that we observe at the meso or macro level. A fully specified microfoundational account of a meso-level feature consists of an account that traces out (1) the features of the actors and (2) the characteristics of the action environment (including norms and institutions) which jointly lead to (3) the social pattern or causal power we are interested in. A microfoundation specifies the individual-level mechanisms that lead to the macro- or meso-level social fact. This is the kind of account that Thomas Schelling illustrates so well in Micromotives and Macrobehavior.

My thinking about the need for microfoundations has changed over the years from a more narrow requirement ("we need to have a pretty good idea of what the individual-level mechanisms are for a macro-property") to a less restrictive requirement ("we need to have reason to believe that there are individual-level mechanisms for a macro-property"). In The Scientific Marx I liked the idea of “aggregative explanations”, which are really explanations that move from features of individual actors and their interactions, to derivations about social and collective behavior. In Varieties of Social ExplanationI relaxed the idea along these lines:

This doctrine [of microfoundationalism] may be put in both a weak and a strong version. Weakly social explanations must be compatible with the existence of microfoundations of the postulated social regularities, which may, however, be entirely unknown. More strongly social explanations must be explicitly grounded on an account of the microfoundations that produce them. I will argue for an intermediate form—that we must have at least an approximate idea of the underlying mechanisms at the individual level if we are to have a credible hypothesis about explanatory social regularities at all, A putative explanation couched in terms of high-level social factors whose underlying individual-level mechanisms are entirely unknown is no explanation at all. (kl 4746)

My adherence to microfoundationalism today is a little bit weaker still. I now advocate a version of microfoundationalism that specifies only that we must be confident (an epistemic concept) that such micro-to-macro relations exist. We must be confident there are such mechanisms but not obliged to specify them. (I also hold that the best ground for having that confidence is being able to gesture plausibly towards roughly how they might work.) Another way to put it is this requirement: "No magical thinking!" That is, we exclude explanations that would only be possible if we assumed action at a distance, blocks of wood that have complicated mental lives, or intelligent beings with infinite computational faculties. A convincing way of discrediting a meso-level assertion is to give an argument that it is unlikely that real human agents would in fact act in ways that lead to this meso-level situation. (Example: Chinese planners who created the collective farming system in the Great Leap Forward assumed that collective farms would be highly productive because a "new socialist man" would emerge. This was unlikely, and therefore the individual behavior to be expected on collective farms would lead to "easy riding" and low productivity.)

Here is an effort to simplify these issues into a series of assertions:

1. All social forces, powers, structures, processes, and laws (social features) are ultimately constituted by mechanisms at the level of individual actors. (ontological principle)
2. When we assert the reality or causal powers of a social entity, we need to be confident that there are microfoundations that cause this social entity to have the properties we attribute to it. (microfoundations principle)
   1. A description of the microfoundations of a social entity S is an account of the circumstances and individual mechanisms that bring about patterns of individual activity resulting in the properties of S.
   2. Strong version: we must provide a credible statement of the microfoundations.
   3. Intermediate version: we must have a back-of-envelope sketch of possible microfoundations.
   4. Weak version: we must have confidence that there are microfoundations, but we don’t have to have any specific ideas about what they are.
3. A "vertical" social explanation of the features of a social entity S is a derivation of S from facts about the individual level. This is equivalent to providing a specification of the microfoundations of S; a derivation of the properties of S from a model of the action situation of the individuals involved; an agent-based model. This is what JZ calls an individualist explanation.
4. A "horizontal" social explanation is one in which we explain a social entity or structure S by referring to the causal properties of other meso-level entities and conditions. This is what we call a meso-level explanation. (The diagram above illustrates these ideas.)
   1. Horizontal explanations are likewise subject to the microfoundations requirement 2: the entities and powers postulated need to be such that we have good reason to believe that there are microfoundations available for these entities and properties. (Epistemic requirement)
   2. Or slightly stronger: we need to be able to offer at least a plausible sketch of the microfoundations / individual-level mechanisms that would support the postulated entities. (Epistemic+ requirement)
5. Providing or hypothesizing about microfoundations always involves modeling the behaviors and interactions of individuals; so it requires assuming a theory of the actor. So when we try to specify or hypothesize about microfoundations for something we are obliged to make use of some theory of the actor.
6. Traditional theories of the actor are generally too abstract and too committed to a rational-choice model.
7. Social scientists will be better able to hypothesize microfoundations when they have richer theories of the actor. (heuristic principle)

So the ontological principle is simply that social entities are wholly fixed by the properties and dynamics of the actions of the actors that constitute them. The requirement of microfoundations simply reproduces the ontological principle, ruling out ontologically impossible relations among social entities. The requirement of microfoundations is not a requirement on what an explanation needs to look like; rather, it is a requirement about certain beliefs we need to be justified in accepting when we advance a claim about social entities. It is what JZ calls a “confirmation” requirement (or perhaps better, a justificatory requirement). A better theory of the actor supports the discovery of microfoundations for social assertions. Further, it provides a richer "sociological imagination" for macro- and meso-level sociologists. So the requirement of microfoundations and the recommendation that social scientists seek out better theories of the actor are also valuable as heuristics for social research: they provide intellectual resources that help social researchers decide where to look for explanatory links, and what kinds of mechanisms might turn out to be relevant.