Sunday, August 8, 2010

Liberty, the knowledge problem, and social evolution

We've seen that Jim Manzi's distinction between liberty-as-goal and liberty-as-means libertarianisms runs into some immediate problems and false paradoxes. Better to contrast dogmatic libertarianism with a pragmatic version, that aims simply to increase liberty in an incremental, experimental manner. But there are deeper issues involved in the notion of liberty, both as goal and means, that Manzi also raises, though again in a somewhat tangled fashion.

The first of these is the famous "knowledge problem" (see, e.g., Hayek's "The Use of Knowledge in Society"), which can take various forms, but can be expressed as the idea that the dispersion of decision-making power amongst the individuals affected by decisions takes maximum advantage of their knowledge of their own circumstances. In this sense, we might say that liberty -- that is, the dispersion of decision-making power -- is a means to the end of maximizing individual well-being, in whatever way individuals define it. But this, by its emphasis on the individual's definition of her own well-being, and by its very generality, comes quite close at least to being equivalent to simply making liberty a goal -- we might say that it helps explain why liberty is a goal.

Manzi may be alluding to something like this in his assertion that "We need liberty ... because we are so ignorant of what works in practical, material terms". But, in bypassing the individual emphasis, it misses a critical aspect of the issue, and leads him into the false paradox noted earlier, where liberty, and the ignorance it addresses, is apparently satisfied simply by dividing society into smaller units, each of which is "free" to be relatively repressive or anarchic.

But perhaps Manzi has in mind another kind of ignorance than the Hayekian notion of the dispersion of knowledge -- perhaps he's simply referring to the limits of human knowledge in general. Here, for example, are the sentences that preceded the quote above:
Liberty-as-means libertarianism sees the world in an evolutionary framework: societies evolve rules, norms, laws and so forth in order to adapt and survive in a complex and changing external environment. At a high level of abstraction, internal freedoms are necessary so that the society can learn (which requires trial-and-error learning because the external reality is believed to be too complex to be fully comprehended by any existing theory) and adapt (which is important because the external reality is changing).
 And here (finally) is a clear and important realization. Let's try to make its components even more explicit:

  • First, social/cultural evolution is not something brought about by design -- by social engineers, policy-makers, or bureaucrats -- but rather by the same sort of random, trial-and-error processes that we see in the natural world; it's Darwinian, in other words, as opposed to Lamarckian. This isn't because human knowledge is futile but simply because it has limits, outside of which nature always waits and exerts her influence -- it's just that influence that appears as evolution, as distinct from, say, "improvement", "reform", "institutional design", "revolution", etc.
  • But, second, such evolution isn't something distinct or apart from human agency, but rather consists of human decisions and actions, and of course of the consequences thereof. Manzi is right to point out, therefore, that the dispersal of decision-making power in the face of this sort of unavoidable ignorance confers a distinct evolutionary advantage on a society, in terms of the greater flexibility and adaptability it allows. There are, as he also points out, limits to this dispersal, or issues on which coercion is still an evolutionary advantage (e.g., obviously, laws against murder or theft, or, arguably, laws against prostitution), but everywhere the evolutionary logic is to disperse this power as far as possible, thereby extending the freedom of the individual agent. 

No comments:

Post a Comment

You can use some HTML tags, such as <b>, <i>, <a>