News

Loading...

Wednesday, July 07, 2010

Methodological Termitism

"Our understanding of the world is built up of innumerable layers. Each layer is worth exploring, as long as we do not forget that it is one of many. Knowing all there is to know about one layer -- a most unlikely event -- would not teach us much about the rest." -- Erwin Chagraff

Consider the oft-despised termite. The obvious 'individual' is the little bug you see crawling around in your wall, obviously 'separate and distinct' from all of the other little bugs. (Yes, I know this is not the technical usage of 'bug'.)

But, for many purposes, it turns out, the real unit of analysis should be the colony, which, in many ways, functions as a super-organism with, for instance, one reproductive organ (the queen), one nervous system, and so on. Or, glancing in the other direction, the 'individual' termite itself appears as a colony, made up of the insect and numerous mixotricha paradoxa (and other symbiotes) occupying the insect's gut and enabling it to live on wood, which the insect, by itself, cannot digest. Now, surely those little guys are indisputably 'individuals', right?

Well, they were called 'paradoxa' because Sutherland, the woman who discovered them, was puzzled that these protists seemed to have both cilia and flagella, which had never been seen on the same critter before. ('Critter' is the correct technical term here, I believe.) Well, they turned out to be stranger than she thought -- the thousands of tiny 'cilia' were discovered each to be 'individual' bacteria symbiotes -- and where cilia usually 'plug in' to a protist at a basal structure in the cell wall, the 'oarsmen bacteria' turned out to be anchored to yet another bacterium!

Anyway, the point of all of this is that sometimes, when looking at termites, the 'right' level of analysis will be the species. Sometimes, the 'right' level will be the colony. Sometimes, it will be the 'apparent' individual, the termite. Sometimes, it will be the termite as itself an ecosystem. And sometimes it will be the components of the termite ecosystem as themselves ecosystems.

Why, other than the desire of economists for simplified analysis or libertarians for a dunk-shot justification of their preferred political policies, should we expect social analysis to be any different? Why would social analysis display one uniquely 'right' level of analysis?

6 comments:

  1. Are you worried that methodological individualism leads economists into error, that it blinds them to compatible but equally important truths, or both? (I assume it's at least one of the two.)

    ReplyDelete
  2. There are a variety of reasons why this approach would be problematic. Though it may be possible (and if it is I'd like to know how), I cannot conceive of a way an individual human being may observe mass human interaction in the way a scientist observes a colony of termites. All means of doing so I am aware of today, from statistical analysis to market analysis to sociological polling, must necessarily use a limited sample and usually alters the environment in the process of gathering the data. One may suggest gathering data in an experimental "laboratory" environment in which variables can be controlled, and this is certainly desirable, but this leaves one to wonder how to keep the participants ignorant of the experiment, and if they are not ignorant, how that will affect the data. Furthermore, this supplies information only of humans interacting in that experimental environment with that quantity of humans. It would be akin to attempting to derive conclusions regarding the nature of monkeys through observing monkeys in captivity. If you draw conclusions about humans in general from the observations you make in a particular controlled experiment, you may come to some useful conclusions, but these conclusions would be limited to the conditions of the experiment and cannot appropriately be applied universally. Doing so will result in errors, as anthropologists for the past few hundred years have found.

    I have no problem studying human interaction at the "societal" or species level provided you consider the limited capacity of your research method. Given this caveat I just don't know why you'd want to. The advantage of analyzing humans at the individual level is that you have the entire human before you. Psychologists have a difficult enough time understanding the complex decision-making processes, emotions, and thoughts of individuals, even with the whole human at their disposal. With many humans, you may create models and projections, but that is all they would be, and they would always rely on certain assumptions (hence the limitations of economics). Beyond that it becomes a problem of one's capacity to calculate a complex system in real time.

    In short, humans are not bacteria, and they are not termites. You cannot look at a human civilization under a microscope or confine it to a laboratory. If nothing else, humans are just too large, but more importantly, human interactions and motivations are simply too complex. Observing civilizations would require a much greater observational capacity than any scientists are capable of as far as I am aware. It is impossible to observe and measure the interactions involved in the production of a single good, much less an entire economy, and the same can be said of most other social interactions. When there is a way of analyzing human societies as wholes in a scientific way, with control groups and variable groups and so on, with one reasonably confident of what variables are changing and which ones are not, it may be appropriate to do what you are suggesting. Until then, any conclusions drawn about such matters are probably drawn through faulty scientific procedures, and as such, should be taken with a grain of salt.

    ReplyDelete
  3. Well, an "isolated" individual's conduct is as readily affected by observation.

    That said, I see little to object to in methodological individualism—at least as applied to economics. It amounts to little more than this: We start with some truisms about human nature, deduce from them laws that describe (with tolerable accuracy) the economic activities of any given individual, and derive from those laws the probable effects of some action (e.g., raising prices) on markets.

    ReplyDelete
  4. "Are you worried that methodological individualism leads economists into error, that it blinds them to compatible but equally important truths, or both?"

    More the latter, just like total reliance on homogenization tended to blind biologists to the importance of cell structure.

    ReplyDelete
  5. "There are a variety of reasons why this approach would be problematic."

    Stargazer, you write as if I am proposing some NEW approach social scientists ought to follow! What I'm writing about is what actually goes on, and it's not perfect, but good results have been produced. (Consider, e.g., linguistics, which doesn't focus on individuals at all, but language communities.)

    "I cannot conceive of a way an individual human being may observe mass human interaction in the way a scientist observes a colony of termites."

    Looking?

    "Observing civilizations would require a much greater observational capacity than any scientists are capable of as far as I am aware."

    Ah, this explains it. You are not aware of the science of history. Check it out, you'll like it -- it studies civilizations all the time.

    ReplyDelete
  6. Oh, Stargazer, two more points:

    1) I got four posts from you, but they all seemed pretty much the same. So I rejected three -- if I left out something important you had to say, my apologies.

    2) Do you have any idea how much work it is to study insect colonies, and how hard it is not to alter their behaviour while studying them? I suggest taking a look at Ants at Work, describing one scientists years spent in the desert trying to figure out the basics of colonial (and inter-colonial) behaviour of just a single ant species, and the ingenuity she had to exercise to collect data without messing the ants up so they reacted to her instead of other ants.

    ReplyDelete