Simon on complexity

Herbert Simon’s The Sciences of the Artificial – 3rd Edition provided an alternative model for thinking about society. We can think of social institutions as partially designed and selected for their organizational properties; so they are different from proteins and planetary systems.  Simon is also an important contributor to the study of complexity. So his new chapter in the 1996 edition of the book, “Alternative Views of Complexity,” is worth reading carefully. Here is how he motivates this new chapter in SA:

The preceding chapters of this book have discussed several kinds of artificial systems. The examples we have examined — in particular, economic systems, the business firm, the human mind, sophisticated engineering designs, and social plans — range from the moderately to the exceedingly complex (not necessarily in the order in which I have just listed them). These final two chapters address the topic of complexity more generally, to see what light it casts on the structure and operation of these and other large systems that are prominent in our world today. (169)

It turns out that there isn’t much new in the 1996 chapter, however. In fact, most of its content is taken from his pathbreaking 1962 article, “The Architecture of Complexity” (link). The new chapter 7 and renumbered chapter 8 largely incorporate the content and sometimes the language of the 1962 article. And this is interesting, because it implies that Simon’s primary ideas about reduction, composition, and inter-level interactions were largely already formed in 1962.

There are a few ideas and themes that are new to the 1996 version. One is a more specific periodization of thinking about complexity theory in the twentieth century.  The 1996 version identifies three phases of theorizing about complexity and “whole systems”.

  1. Biological emergence theory (post World War I)
  2. Cybernetics and systems theory (post World War II)
  3. Contemporary complexity theory (post 1960s)

Simon is skeptical about the tendency towards irreducible holism that was associated with the earlier two phases of thinking in both versions; in the 1996 chapter he favors a “weak” interpretation of emergence: a commitment to …

… reductionism in principle even though it is not easy (often not even computationally feasible) to infer rigorously the properties of the whole from knowledge of the properties of the parts. In this pragmatic way, we can build nearly independent theories for each successive level of complexity, but at the same time, build bridging theories that show how each higher level can be accounted for in terms of the elements and relations of the next level down. (172)

This “pragmatic holism” is already contained in the 1962 version (link). So this doesn’t represent new ground in 1996. But Simon’s use of this idea to criticize several false starts in the field of complexity research is valuable.

Simon finds some of the central concepts of the third phase to be more promising for the study of social phenomena. The mathematics and physics of chaotic behavior (where simple low-level processes can aggregate to wildly variant higher-level outcomes), simulations of evolution through computational models (genetic algorithms), and the exploration of cellular autonoma (the game of life) all come in for favorable comments. (The Lorenz attractor illustrated here is a common example of chaotic behavior.)

One idea that is not contained in the 1962 version is that of causal non-linearity. Non-linearity is a problem for the “near decomposability” view that Simon wanted to take of complexity in the 1962 version, because it casts doubt on the ability to disentangle causal influences deriving from inter-connected subsystems. Small differences in initial conditions can lead to large differences in outcome. This is a key aspect of chaos theory and the varieties of turbulent phenomena that provide the best examples of chaotic systems. And this casts some doubt on one of the central conclusions of the 1962 paper:

The fact, then, that many complex systems have a nearly decomposable, hierarchic structure is a major facilitating factor enabling us to understand, to describe, and event to “see” such systems and their parts. Or perhaps the proposition should be put the other way round. If there are important systems in the world that are complex without being hierarchic, they may to a considerable extent escape our observation and our understanding. (477)

This is a decidedly pre-chaos understanding of the nature of complex systems. I have the impression that many contemporary complexity theorists would reject the idea that social processes are commonly the result of “nearly decomposable, hierarchic structures”. So it is a genuine change for the mathematics of chaos theory to be included in the 1996 version. Complexity research has moved forward since 1962, and Simon recognizes this in the 1996 chapter.

What we don’t find here is any discussion of whether actual social processes and systems display chaotic behavior in this well defined sense. And we don’t see Simon shifting his position on “nearly decomposable” systems.

Are there examples of social processes and phenomena that display chaotic characteristics over time? Take the occurrence of massive street demonstrations as an example; are there aspects of chaos in the technical sense involved in the outbreak of street mobilization? Do small, apparently random events have large effects on the eventual outcome?

It would appear that this is the case when we look at the cases of uprising and passivity in different cities during the Arab Spring of 2011. Some social scientists have tried to understand the likelihood of uprising as an increasing function of economic crisis, regime weakness, and regime brutality. This implies a linear assumption about the causal role of these three forces. But it seems plausible to speculate that random events like a broken phone chain, an Internet outage, or the defection of a key leader could push the process of mobilization into a different direction. Moreover, it seems that contemporary research on social complexity pays a lot of attention to non-linearity, path-dependency, and sequential processes of social mobilization — leaving a lot of room for the kinds of turbulent effects that are observed in traffic flow, storm generation, and water dripping from a leaking tap. This is the kind of work that is described in Scott Page and John Miller, Complex Adaptive Systems: An Introduction to Computational Models of Social Life.

So oddly enough, it seems that one could fairly say that Simon’s views of social complexity — as expressed in the 1996 third edition of  The Sciences of the Artificial as well as in his groundbreaking “Architecture of Complexity” in 1962 — are significantly incomplete, given the way that complexity theorists are now thinking about social processes. Simon did not incorporate the guiding assumptions of “complex adaptive systems theory” into his own thinking, and remained convinced of the adequacy of the ideas of hierarchical systems and nearly decomposable systems as late at 1996.  His own approach to social complexity remains a phase two approach, not a phase three approach.

(The graph at the top of this post is offered as an interpretation of a highly path-dependent social process. The reader is asked to consider each path as a hypothetical development from a common origin, with small stochastic variations in the situation occurring over time. Imagine the starting position is “large city, economic hardship, weak state, lots of repression”, time is the x axis, and the y axis measures civil unrest. Some of those variations push the path towards a high outcome (blue), and some towards a low outcome (magenta). The great majority of outcomes fall within a short distance of the starting position. So the most likely outcome is “not much change”, but there are unlikely but diametrically different outcomes possible as well.)

Microfoundations and meso causation

I take the view that social causation requires microfoundations. And I hold that meso causal explanations are legitimate. How are these two views compatible?

The key is the role that we expect reasoning about micro-level events to play in the explanation itself. The various versions of methodological individualism — microeconomics, analytical sociology, Elster’s theories of explanation, and the model of Coleman’s boat — presume that explanation needs to invoke the story of the micro level events as part of the explanation. The microfoundations-meso explanation perspective requires that we be confident that these micro-level events exist and work to compose the meso level; but it does not require that the causal argument incorporates a reconstruction of the pathway through the individual level in order to have a satisfactory explanation.

Putting the methodological individualist model of explanation crudely, we get something like this:

Institution {I} shapes individuals’ desires, beliefs, and passions {Ai}; individuals with these features of agency constitute the micro level; the tools of game theory, agent-based modeling, microeconomics, etc., allow us to aggregate the behavior of these individuals back to the macro level, in a new state of the macro level {O}.

The” meso-causal with microfoundations” view goes something like this.

Institution {I} has causal power P to influence future states of social aggregates; {I} occurs leading to P.
And, outside the explanation itself, we have:
[Institution I’s causal powers have micro foundations along these lines: …]

So when Michael Mann describes the effects of paramilitary organizations in Germany in the 1920s, he relies on a number of assertions about the causal powers of these organizations when introduced into the social circumstances of the 1920s. He is prepared to say how these effects work–so he satisfies the microfoundations requirement. But his explanation is a meso-level one, proceeding from the meso-level causal properties of paramilitary organizations to another set of meso-level effects (link).

When we explain why the tea kettle begins to whistle, we provide a meso-level explanation and we recognize the microfoundations that each step rests upon:

The kettle is filled with water.
The kettle is subject to a hot flame.
Water boils at 100 degrees centigrade. 
Steam is produced at higher pressure than one atmosphere.
Steam exits through the spout, creating the whistle.

This is a meso level physical explanation with very well understood microfoundations, at the level of the molecular properties of H2O, its phase transitions, etc. And, of course, that molecular explanation itself has microfoundations. But here is the key point: the explanation at the level of heat, liquid, boiling point, and steam is perfectly satisfactory for many purposes. And it is scientifically satisfactory because we are confident that we understand the micro mechanisms that underlie each step of the argument. But we are not obliged to provide a molecule-level simulation of the situation in order to be satisfied that we have explained the whistling tea kettle.

Working sociologists offer explanations like Mann’s on a regular basis. They identify what they take to be causal properties of social structures and institutions, and then draw out causal chains involving those causal properties. And often they are able to answer the follow-on question: how does that causal power work, in approximate terms, at the micro level? But answering that question is not an essential part of their argument. They do not in fact attempt to work through the agent-based simulation that would validate their general view about how the processes work at the lower level.

This account suggests an alternative diagram to Coleman’s boat. 

The diagram represents the meso-level claim that E1 and E3 jointly cause the occurrence of E2. The meso causal relation is represented as a single-directional orange arrow. The green arrows represent the microfoundations of each meso factor. These causal relations are bidirectional: individual actors are influenced by the meso factor, and their actions have influences on the meso factor. The yellow arrows, finally, represent inter-actor influences at the micro level: network relationships, alliances, communications channels, exemplary behavior, mobilization efforts, etc. 

The diagram represents each of the causal linkages represented in the Coleman boat. But it calls out the meso-meso causal connection that Coleman prohibits in his analysis. And it replaces the idea that causation proceeds through the individual level, with the idea that each meso level factor has a set of actor-level microfoundations. But this is an ontological fact, not a prescription on explanation.

And, finally, it is self-evident that we can always ask a different question that looks more like Coleman’s: how does the causal property of the meso factor work at the level of the actor? But this is a different question, and we aren’t required to incorporate the resulting story into the meso-meso explanation.

Relative explanatory autonomy

In an earlier post I indicated a degree of disagreement with the premises of analytical sociology concerning the validity of methodological individualism (link). This disagreement comes down to three things.

First, for reasons I’ve referred to several times here and elsewhere (link), I prefer to refer to methodological localism rather than methodological individualism.

This theory of social entities affirms that there are large social structures and facts that influence social outcomes. But it insists that these structures are only possible insofar as they are embodied in the actions and states of socially constructed individuals. The “molecule” of all social life is the socially constructed and socially situated individual, who lives, acts, and develops within a set of local social relationships, institutions, norms, and rules. (link)

I believe that the ideas of localism and socially constructed, socially situated actors do a better job of capturing the social molecule that underlies larger social processes than the simple idea of an “individual”. Structural individualism seems to come to a similar idea, but less intuitively.

Second, the requirement of providing microfoundations for social assertions is preferable to methodological individualism because it is not inherently reductionist (link). A microfoundation is:

a specification of the ways that properties, structural features, and causal powers of a social entity are produced and reproduced by the actions and dispositions of socially situated individuals. (link)

We need to be confident that our theories and concepts about social structures, entities, and forces appropriately supervene upon facts about individuals; but we don’t need to rehearse those links in every theory or explanation. In other words, we can make careful statements about macro-macro and macro-meso links without proceeding according to the logic of Coleman’s boat — up and down the struts. Jepperson and Meyer make this point in “Multiple Levels of Analysis and the Limitations of Methodological Individualisms” (link), and they offer an alternative to Coleman’s macro-micro boat that incorporates explanations referring to meso-level causes (66).

Third, these points leave room for a meta-theory of relative explanatory autonomy for social explanations. The key insight here is that there are good epistemic and pragmatic reasons to countenance explanations at a meso-level of organization, without needing to reduce these explanations to the level of individual actors. Here is a statement of the idea of relative explanatory autonomy, provided by a distinguished philosopher of science, Lawrence Sklar, with respect to areas of the physical sciences:

Everybody agrees that there are a multitude of scientific theories that are conceptually and explanatorily autonomous with respect to the fundamental concepts and fundamental explanations of foundational physical theories. Conceptual autonomy means that there is no plausible way to define the concepts of the autonomous theories in terms of the concepts that we use in our foundational physics. This is so even if we allow a rather liberal notion of “definition” so that concepts defined as limit cases of the applicability of the concepts of foundational physics are still considered definable. Explanatory autonomy means that there is no way of deriving the explanatory general principles, the laws, of the autonomous theory from the laws of foundational physics. Once again this is agreed to be the case even if we use a liberal notion of “derivability” for the laws so that derivations that invoke limiting procedures are still counted as derivations. (link)

The idea of relative explanatory autonomy has been invoked by cognitive scientists against the reductionist claims of neuro-scientists. Of course cognitive mechanisms must be grounded in neurophysiological processes. But this doesn’t entail that cognitive theories need to be reduced to neurophysiological statements. Sacha Bem reviews these arguments in “The Explanatory Autonomy of Psychology: Why a Mind is Not a Brain” (link). Michael Strevens summarizes some of these issues in “Explanatory Autonomy and Explanatory Irreducibility” (link). And here Geoffrey Hellman addresses the issues of reductionism and emergence in the special sciences in “Reductionism, Determination, Explanation”.

These arguments are directly relevant to the social sciences, subject to several important caveats. First is the requirement of microfoundations: we need always to be able to plausibly connect the social constructs we hypothesize to the actions and mentalities of situated agents. And second is the requirement of ontological and causal stability: if we want to explain a meso-level phenomenon on the basis of the causal properties of other meso-level structures, we need to have confidence that the latter properties are reasonably stable over different instantiations. For example, if we believe that a certain organizational structure for tax collection is prone to corruption of the ground-level tax agents and want to use that feature as a cause of something else — we need to have empirical evidence supporting the assertion of the corruption tendencies of this organizational form.

Explanatory autonomy is consistent with our principle requiring microfoundations at a lower ontological level. Here we have the sanction of the theory of supervenience to allow us to say that composition and explanation can be separated. We can settle on a level of meso or macro explanation without dropping down to the level of the actor. We need to be confident there are microfoundations, and the meso properties need to be causally robust. But if this is satisfied, we don’t need to extend the explanation down to the actors.

Woven throughout this discussion are the ideas of reduction and emergence. An area of knowledge is reducible to a lower level if it is possible to derive the statements of the higher-level science from the properties of the lower level. A level of organization is emergent if it has properties that cannot be derived from features of its components. The strong sense of emergence holds that a composite entity sometimes possesses properties that are wholly independent from the properties of the units that compose it. Vitalism and mind-body dualism were strong forms of emergentism: life and mind were thought to possess characteristics that do not derive from the properties of inanimate molecules. Physicalism maintains that all phenomena — including living systems — depend ultimately upon physical entities and structures, so strong emergentism is rejected. But physicalism does not entail reductionism, so it is scientifically acceptable to provide explanations that presuppose relative explanatory autonomy.

Once we have reason to accept something like the idea of relative explanatory autonomy in the social sciences, we also have a strong basis for rejecting the exclusive validity of one particular approach to social explanation, the reductionist approach associated with methodological individualism and Coleman’s boat. Rather, social scientists can legitimately aggregate explanations that call upon meso-level causal linkages without needing to reduce these to derivations from facts about individuals. And this implies the legitimacy of a fairly broad conception of methodological pluralism in the social science, constrained always by the requirement of microfoundations.

Peter Demeulenaere on analytical sociology

Here is another take on the core features of analytical sociology, this time from Peter Demeulenaere in the introduction to his very interesting recent collection on the subject, Analytical Sociology and Social Mechanisms. The volume includes thirteen essays by leading experts grouped around “Actions and Mechanisms,” “Mechanisms and Causality,” and “Approaches to Mechanisms,” and it is an important further contribution to the framework of analytical sociology. (The volume is also available in an affordable Kindle edition; Analytical Sociology and Social Mechanisms.)

Demeulenaere makes several basic points. First, he holds that AS is not just another new paradigm for sociology. Instead, it is a reconstruction of what valid explanations on sociology must look like, once we properly understand the logic of the social world. He believes that much existing sociology conforms to this set of standards — but not all. And the non-conformers are evidently condemned to being judged non-explanatory. So this sets a claim of a very high level of authority over the whole field — what Ronald Jepperson and John Meyer refer to as micro-chauvinism (link, 57). Here are several of Demeulenaere’s statements of the intended scope of AS:

Analytical sociology should not therefore be seen as a manifesto for one particular way of doing sociology as compared with others, but as an effort to clarify (“analytically”) theoretical and epistemological principles which underlie any satisfactory way of doing sociology (and, in fact, any social science). The social sciences already command a considerable stock of substantive descriptions and explanations; and some of the alternatives to these are either redundant, or resistant to proof, even false or imprecise, quite regardless of their status with respect to one or other established paradigm. Analytical sociology should seek to define a set of sound epistemological and methodological principles underlying all previously established and reliable sociological findings. The aim of analytical sociology is to clarify the basic epistemological, theoretical and methodological principles fundamental to the development of sound description and explanation. (Kindle loc 121)

The most important aspect of the analytical approach should be to clarify the strategy by which we endeavour to separate and conceptualize different elements entering into descriptions and explanations of the social world, so that we might understand their mutual relationships, and in particular the causal links existing among them. (Kindle loc 139)

Second, Demeulenaere provides a long, detailed, and helpful analysis of the doctrine of methodological individualism and its current status. He believes that criticisms of MI have usually rested on a small number of misunderstandings, which he attempts to resolve. For example, MI is not “atomistic”, “egoistic”, “non-social”, or exclusively tied to rational choice theory. He prefers a refinement that he describes as structural individualism, but essentially he argues that MI is a universal requirement on social science.

The two core ideas behind MI, first expressed by John Stuart Mill and Carl Menger, and subsequently by Weber, can be expressed very simply:

— Social life exists only by virtue of actors who live it;

— Consequently a social fact of any kind must be explained by direct reference to the actions of its constituents.

These two simple propositions remain central to the analytical approach; we therefore have to address the problem of the relationship between MI and AS. This section is directed to a brief exposition of the problem. (Kindle loc 183)

This is important, because it involves two separate assertions: social things are composed of individual actions and nothing else; and “therefore” social outcomes must be explained on the basis of facts about individual actions. But the inference doesn’t hold — more in an upcoming post on explanatory autonomy.

Demeulenaere specifically disputes the idea that MI implies a separation between society and non-social individuals. He believes that even the originators — Watkins and Mill, for example — recognized that individuals are social organisms. But he believes this recognition can be folded into a consistent elaboration of MI that he describes as structural individualism:

The combination of these two approaches can be called “structural individualism” (Wippler 1978; Udehn 2001). Any serious attempt to reflect on a social situation should deploy both in turn. Their combination is in some respect illustrated by Coleman‘s famous “boat” (1986, 1990). It remains a central aspect of analytical sociology. (kl 246)

That said, Demeulenaere fully endorses the idea that AS depends upon and presupposes MI:

Does analytical sociology differ significantly from the initial project of MI? I do not really think so. But by introducing the notion of analytical sociology we are able to make a fresh start and avoid the various misunderstandings now commonly attached to MI. (kl 318)

Third, Demeulenaere holds that AS depends closely on the methodology of social causal mechanisms. The “analytical” part of the phrase involves identifying separate things, and the social mechanisms idea says how these things are related. Demeulenaere gives a compressed history of the development of this framework over the past decade or so.

Analytical sociology incorporates an affirmation that “social facts” are generated, triggered, produced, brought about, or “caused” by individual actions which themselves are in some sense “caused”, or at least partly determined by, the constraints presented by the social environments and situations in which such actions take place.

Therefore the focus has to be on the causal “process” occurring at the action level. The idea that there are laws directly implemented at a macro level can be easily rebutted, since the effectiveness of the outcome necessarily leads to the “active” level, the level of action. One general implication of the notion of “mechanism” is to move analysis away from an “inactive” level to an “active” level, where effective actions occur. A strong correlation between variables should not therefore be interpreted in causal terms unless a mechanism linking the two dimensions is identified, mechanisms involving effective actions. (kl 378)

So causal mechanisms are expected to be the components of the linkages between events or processes hypothesized to bear a causal relation to each other. And, more specifically to the AS approach, the mechanisms are supposed to occur solely at the level of the actors–not at the meso or macro levels. So this means that AS would not countenance a meso-level mechanism like this: “the organizational form of the supervision structure at the Bopal chemical plant caused a high rate of maintenance lapses that caused the accidental release of chemicals.” Both features are meso-level factors or conditions, and it would appear that AS would require that their causal properties be unpacked onto individual actors’ behavior.

I, on the other hand, would maintain that this is a perfectly legitimate social mechanism because we can readily supply its microfoundations at the behavioral level. So this suggests that we can legitimately refer to meso-level mechanisms as long as we are mindful of the microfoundations requirement. And this corresponds as well to the tangible fact that institutions have causal force with respect to individuals — a point Demeulenaere acknowledges here:

Hence we commonly encounter the fact that, one set of actors having at a certain point adopted a norm, or built up an institution, it then become more difficult for actors subsequently coming onto the scene to adopt another norm, or replace the institution. But this is never impossible. (kl 372)

There is quite a bit of the framework of AS that I find appealing and constructive. Moreover, several of the core premises line up well with assumptions I’ve argued for in my own philosophy of social science — microfoundations, causal mechanisms, and agent-centered ontology, for example. Where there appear to be a few inches of separation between the views is on the subject of reductionism and explanatory autonomy. I will return to these differences in an upcoming post.