Monday, January 9, 2012

Assumptions and Reality

While I've not been not blogging, I've been catching up on my reading.  Part of that has been reviewing some basic texts in political theory as a bit of a refresher before starting my PhD next year as well as reading a bunch of history while I have time for it.  Reading these two subjects together made me reflect a bit on a basic truth, that far too  many people take their assumptions far too seriously.

Let me explain.  Assumptions, as well as theories and basically all the methods of scientific inquiry, are simply tools.  Assumptions and theories are never accurate reflections of reality.  Rather, they are crutches we use because human beings are incapable of dealing with the world as a whole, such levels of comprehension are left to God.

So, we use observation, theory, and assumptions in order to explore the world around us one piece at a time.  For most of history, the purpose of inquiry was to make our observations match with our theories about how the world worked, since we knew how the world worked the task of inquiry was to discover why our observations matched so poorly with our theory and assumptions.  This led to such achievements of human ingenuity and brilliance as the Ptolemaic geocentric model.  Brilliant, creative, and a stunning example of the powers of human inquiry.  But also dead wrong.

Modern scientific inquiry works a bit differently.  We take our observations as being relatively reliable (though only a fool trusts observation implicitly, it's better than forcing the world to fit preexisting conceptions, but it ain't perfect) and test our theories and assumptions against these.

One problem I notice rather frequently however, is that the success of a theory at explaining one piece of a puzzle inevitably leads some goofball to thinking that they've struck upon the whole thing.  This is nonsense.  The specific area I want to look at is the relation of the individual and society.

Now, this isn't a terribly deep diagram.  All I want to get across is that there is a holistic relation of the individual to society.  Various theories tend to take different cuts of this overall relationship, for my purposes today I've divided this into three spheres, political, by which I mean power relations (not the best definition of politics but it's the one that draws out the contrasts I am concerned with today), economics, by which I mean exchange relations between individuals that are independent of power relations (like buying a orange from the grocery store), and social relations, by which I mean things like friendships and other relations where there are no clear power disparities or any observable exchanges as part of the activity.  There is a substantial part of the diagram where I've added no circle, there are aspects of the relation of individuals to each other and to groups that cannot be encompassed within these three spheres, they're irrelevant for today.

Now, early thought generally took a more holistic look at these relations then moderns do.  For the Greeks, man was a political animal and economics was the study of household management.  Pretty much all three spheres I outlined were collapsed into a continuous view of the nature of man.

This is probably pretty accurate, it's also pretty much useless for making theories that can be tested against observations.  So the Greeks ended up having a pretty stagnant economy and descended into tearing each other apart while a bunch of barbarians conquered them, so it goes (oversimplification, of course).

Modern thought doesn't try to grapple with everything at once.  Instead, we posit that the individual is a rational being seeking to making utility according to a set of utility functions, or that individual action and preference is the result of the relation of the individual to the means of production and the contradictions inherent in membership in a social class, or that an individual is defined by their membership in a culture and the system of symbols associated with that culture (badly paraphrasing some arguments I'm less familiar with here), or that and individual is the product of the immediate impressions of the senses on the brain (trying to get at Hobbes here, but it has been a long while), and so on ad infinitum. 

From these assumptions and associated theories we can reduce the problem of the individual and society into manageable chunks that can be tested and analyzed.  We've learned a lot doing this, leading to a much better relation of individuals to each other in all spheres, whether economic, political, or even just social.

The problem is when some boneheaded researcher gets it in their head that the success of explaining individual action in, say, the economic sphere, means that all relations are in fact economic and decides to go and promote this idea in the public sphere and try to base policy on it.  Never mind that their assumptions don't include basic political concepts like legitimacy, loyalty, or authority.  Since their economic assumptions worked so well obviously these things can be explained in terms of rational choice (or rational choice could be explained as epiphenomena resulting from the relations to the means of production and the class structure at a given point of development, taken past the limits of good sense, any theoretical perspective can be explained as meaningless chimera in terms of another).

This is a problem since, as with all assumptions and theory, this view is deeply wrong.  Theory and assumptions are tools used to explore a narrow domain, if what you want is the Truth, with a capital T of course, the more holistic Greek idea is better.  As I said though, it's useless analytically.  But when the point isn't to analyze data and discover relations but to make policy the more inclusive concepts are necessary.

From this perspective, the purpose of scientific inquiry into social processes is to inform and guide, not dictate policy.  Inquiry has done an excellent job in describing to policymakers the tradeoffs they face and given them a better sense of the potential costs and benefits of policy (in a probabilistic sense).  However, when theory and assumptions go beyond informing policymakers as to the tradeoffs and instead try to dictate policy (usually with some variant of the notion that if policy prescriptions are followed exactly some sort of amazing change in human welfare will result) disaster usually results.  This occurs because the assumptions and theories used by an researcher necessarily only explores a small segment of the total relation between the individual and society (as well as between groups within a society).  The stuff that gets left out of any theory will necessarily bring any utopian program based solely on one perspective to a disastrous end.

Theories that assume strict methodological individualism and voluntary contract, for instance, will be brought crashing down because of the objective existence of groups, bonds of loyalty between family members, and power relations (I have Ryan's various pipe dreams in mind here).  Too much is simply left out of this perspective to provide a firm foundation for proposals as sweeping as those proposed.  It ignores, for instance, how economic relations are rarely as simple as those of walking into a market and buying an orange (simple economic relations can also be neutral between modern floating prices and just price theory, this remains economic under both modern and ancient conceptions of the economy).  Many economic relations become politicized, even absent the state.  In the case of health care, we know that power relations, usually expressed in economic terms simply as buyer or provider bargaining power, play a big role.  Bigger, more powerful actors are able to secure better deals than smaller actors.  We also know that different groups have different tendencies to combine to exercise power, supplier interests have a powerful tendency to combine while consumer interests rarely combine, absent state support.

There is no good purely economic reason for this.  It would be in our interests to combine into large consumer groups to bargain with providers.  But this doesn't happen.  This particular failure can be explained largely using economic reasoning, it's a simple collective action problem.  However, dealing with this isn't really possible purely within the economic system, we need some form of political intervention.  Absent any real theory of politics or of how groups forms it appears plausible that spontaneous coordination of free agents should be able to compete against existing groups if this would be economically efficient.  In practice, this does not seem to occur, or when it does to be efficient (spontaneous coordination really needs to deal better with things like peasants binding themselves to lords in the middle ages or mutual protection societies that eventually evolved into guilds, modern economic efficiency owes a lot to these spontaneous bonds decaying because they came into conflict with the state, spontaneous coordination does occur but most of the time it is inefficient and needs an outside factor to force it to rationalize, that one theoretical perspective says they should work together because of superior local information doesn't amount to evidence that the wealth of information possessed locally is in fact the most salient information necessary to efficient and rationalized enterprise, I don't think the research on these questions really supports spontaneity position).

Anyway, I think I've made my point and descended into rambling.  The assumptions and theories we use to analyze the world are necessarily incomplete and can explore only a portion of reality.  That one theory works does not have any impact on another theory, just because one perspective is reasonable and logical does not mean that an opposing perspective is not, just that they are looking at a different portion of reality.  The test for any theory is how much evidence it explains, how few assumptions it is forced to rely on (and how easily it handles what assumptions it has being problematized), and most importantly, how effective it is at making predictions.  That a theory accords better with common sense or is argued in a more appealing fashion is irrelevant, performance against data is all that matters.  It is important to note that people often mistake correspondence to common sense as evidence, since common sense changes across cultures and time periods this is obviously nonsense as a test.  Also, just as with most historical perspectives, people tend to be more interested in justifying the current way things are and their prejudices and stereotypes than they are in doing honest analysis of observations.  It is important to always be on guard that a theory's plausibility has more to do with confirming our biases and describing how things are the way they are, rather than in actually dealing with how things work or how change happened to make things the way they are.

Policy, of course, must rely on using all available theory and evidence, it cannot simply dismiss inconvenient evidence because people will suffer for poor decisions.  Too often, it seems, people in positions of power forget that they are responsible for the lives and well being of others and instead focus on adhering to their own set of beliefs, rather than the indicators that are agreed upon by the bulk of the people they govern.

No comments:

Post a Comment