What is the unit of analysis appropriate for the social sciences?

In economics, methodological individualism is one of the pillars of the Neoclassical school. Its 20th century rivals focused on group-level analysis of various sorts; the German Historical school treated the nation as an organism, while Marxists focused on class interest.

I was educated in the Neoclassical tradition, heavily supplemented with insights from the Austrian school. There are big differences between these schools, but those differences are not located in their respective choices for unit of analysis: both are methodological individualists through and through.

For that reason, I tend to be most critical of methodological individualism. It is the framework I have explored the most, its insights are most readily available to me and so, therefore, are its shortcomings.

But I don’t want to give the impression that its historic rivals are therefore superior, in my eyes. It’s more that I’ve become acutely aware of some basic questions that I have never seen addressed to my satisfaction. It is a serious possibility that they have been addressed, and I have simply missed it. If so, I would be immensely grateful to anyone who could point me to the works that do so.

Nevertheless, I find myself wishing for a basic social ontology. To that end, I will sketch one out below. Emphasis on “sketch”.

Levels of Abstraction

Taking for granted modern physics, chemistry, and indeed the very phenomena we can observe ourselves, my first question is what makes anything properly a unit at all.

Take an individual organism. Even if we can say it would be a useless or misleading unit for evolutionary science or economics to take as its starting point, just what is it that makes it a candidate to begin with?

The methodological individualist says, “groups are just aggregations of individuals.” But individuals are just aggregations of cells, which are aggregations of molecules, which are aggregations of atoms, which are aggregations of subatomic particles.

So should every science begin with the most fundamental particles we know of as their unit of analysis?

But how are even these considered units? Just by virtue of being indivisible?

The disciplines which appear, to my amateur eye, to be the most promising in terms of producing answers to this question, are philosophy of mind and complexity theory. In philosophy of mind they speak of “levels of abstraction.” See this paper by Christian List and Kai Spiekermann applying the concept of the micro-macro divide in economics. In complexity theory, there is the concept of emergent systems that cannot be reduced to their parts for the purposes of explaining the behavior of that system.

The view that I have in mind, for our purposes, is what List and Spiekermann refer to as “causal-explanatory holism”:

Some causal relations (of the kind that a social scientific explanation would describe) are distinct from (and not redescribable as) any individual-level causal relations.

Thus a group is a relevant unit of analysis if it involves specific causal relations that are distinct from individual-level causes. The group in question becomes a level of abstraction worthy of scientific study.

In philosophy of mind, this is used to argue, for example, that beliefs or ideas are distinct causal relations from neurons firing. You can get more useful information about how a person will behave in a given situation by analyzing in terms of their beliefs than by looking at neuron firing patterns.

I think the German Historical economists were onto something with their biological metaphor. In biology, there isn’t really one unit of analysis, though the organism is the whole that biology is oriented around. But you can spend your whole career studying liver cells, or viruses, or the immune system, or the relations among organisms such as reproduction or the spread of disease. There are many levels of abstraction you can choose from, all of which have important causal relations of their own.

In human society, there are individuals—irreducible, I believe, for reasons explored in philosophy of mind—but there are also social entities with their own existence. We need not even go as far as macroeconomics to see this; the price system, the primary focus of microeconomics, itself is such a level of abstraction. You only need to read Hayek’s famous paper on knowledge, or Leonard Read’s more poetic take on it, to see that there are system-level causal relations occurring beyond the individual level. Individuals provide feedback into the system, to be sure, but the system provides feedback to them. In short, causation flows both ways across the entire hierarchy of levels of abstraction.

Let us set to the side the status of this concept, the level of abstraction. Exploring that is the purview of metaphysics, and I am only interested in ontology here. Suffice it to say that there’s a strong case to be made that this concept is at minimum a useful tool.

Systems, Groups, and Individuals

I said the German Historical school was onto something, but that’s as much credit as I’m prepared to give them. The relationship between individual human beings and larger social entities is nothing like the relationship of cells or organs to an organism.

This should be obvious, almost trivially so. You can immigrate to a different country. You can change employers. You can leave your church. You can move out of your neighborhood. The social organism is a poor metaphor.

Here is my stab at a taxonomy:

• There are individuals; thinking and talking animals who by virtue of what it means to think and talk cannot be further reduced.
• There are social entities—groups, if you like—that are able to exist through the unity of ethics, politics, and rhetoric. The relationship of individuals to these entities is holistic.
• There are processes which govern the relations between social entities, which I will refer to as systems. Thus, the price system is not a group, but a particular set of processes that connect disparate groups together—these days, on a global scale. The relationship of individuals to these systems is conjective.

I could go on, but I think I’ve reached the limit of what I can usefully say on the subject at this point. I am sure I will revisit it; social ontology is an itch I cannot stop myself from scratching.

Further questions not addressed here:

• Does a social entity retain a meaningful identity once all original individuals have turned over? If so, why?
• Is there really a meaningful distinction between the processes that generate social entities, and the processes I have referred to as systems?
• What are the implications, morally and politically, for acknowledging the existence of social entities?

I welcome all feedback on any of the topics touched on in this post.

 

 

 

Featured image is Gentile Bellini’s Procession in St. Mark’s Square