Chris linked to an essay by Razib Khan arguing that empiricism implies conservatism. Khan defines empiricism as the conjunction of two attitudes: (1) we should “cautious of overturning practices and customs that have served society and individuals in good stead” (preference for the status quo), and (2) social change should occur through “an interative process that sifts optimal solutions by trial and error and maintains previous patches along the way” (preference for incrementalism). Khan grounds his preference for the status quo and for incrementalism in epistemic humility. “Human affairs are complex, contingent, and difficult to tease apart in their interrelationships,” so we should be wary of tinkering with them least we cause them to come completely unraveled.
Khan believes that these two attitudes, undergirded by epistemic humility, support empirical conservatism, which he defines as the view that “there doesn’t have to be a reason for a practice; its very persistence across generations is a mark in its favor, because the prejudice is that persistent practices are not harmful at worst and may be positively essential at best.” To clarify what this prejudice means in practice, Khan explains how the empirical conservative approaches whether to allow gay marriage:
Let us observe homosexual relationships and see if they flourish. On an empirical conservative worldview, the aim of marriage is not to fulfill a simple principle such as equality but to develop human life in full. Monogamy, whether heterosexual or homosexual, seems to fit that bill. In contrast, polygamy and polyamory seem suboptimal for a variety of reasons, although individuals make the arrangements “work” in some cases. Instead of reducing the argument for gay marriage to one of legal equality, one should enumerate the social benefits to a wide swath of individuals that legally sanctioned monogamy allows. Additionally, the reality is that gay marriage as a practice has been introduced in various locales without any negative consequences. The experiment has been attempted and seems to be a success.
Khan’s essay has at three flaws. Its first flaw, clear from the portion purporting to apply empirical conservatism to the problem of gay marriage, is that it does not grasp the sweeping scope of the empirical conservatism that it has defined. Khan believes that the empirical conservative should eschew arguments for gay marriage based on legal equality in favor of arguments based on the continued success of communities where gay marriage has been introduced. He therefore thinks that empirical conservatism implies a (arguably pseudo)-scientific methodology for deciding whether a particular change should be supported. This is not the case. On Khan’s definition, empirical conservatism is a general prejudice in favor of persistent practices. An entire category of persistent practices are decision procedures for deciding whether a particular change should be supported. These procedures, including rights-based legal argumentation, are longstanding and venerable (I dare say, far more longstanding and venerable than Khan’s pseudo-scientific methodology for evaluating the desirability of gay marriage). The empirical conservative thus should not recommend eschewing arguments from legal equality; he should embrace those arguments as part of a deep-rooted practice of, uh, politico-legal social dynamism. Indeed, the empirical conservative should embrace virtually all of the common forms of politico-legal argumentation and reform that operate in America today for precisely the reason that they are common and they operate in America. He therefore should have no cause to intervene in any political debate that is recognizably a political debate; his concerns should be confined to preventing radically different forms of politics from emerging.
The second flaw in Khan’s essay is that it does not provide for any principled way to define a practice presumptively worthy (a fortiori) of preservation. Practices can be defined at varying levels of generality. Take, for example, the practice of restraining people who violate criminal laws. One statement of this practice is just a list of all the variegated ways in which we confine people. Another statement is more general but still quite specific; the practice of confining people for certain periods of time in certain conditions for certain conduct and under certain public policy justifications. Another is even more general still: the practice of confining people for certain periods of time (never mind the conditions). Which one is the practice? Define it too specifically and Khan’s conservatism becomes ridiculous (even on its own terms — how many times have we made minute adjustments to legal regimes without experiencing catastrophe?). But how broad can one go? This is a matter of crucial importance for concrete policy questions; until it is answered empirical conservatism can provide no guidance.
The previous two criticisms reduce empirical conservatism to a vague warning against radicalism. This broad anti-radicalism, too, cannot stand because it rests on a faulty inference (one that is at odds with what conservatives believe in other contexts), viz. from the premise that human relationships are complex and difficult to tease apart to the conclusion that therefore we should walk on eggshells when trying to make policy, lest we disrupt the contingent web and cause all to come tumbling down. But surely the complexity of human relationships, and the difficulty in teasing them apart, actually supports the view that they are resilient, and so it takes quite a lot to disrupt them. (This is, indeed, a lesson conservatives are happy to absorb in other contexts, e.g. when focusing on the abject failure of Stalinist efforts to create an idealized Soviet Man.) And if human relationships are, indeed, resilient then risks from significant social interventions that all would recognize as catastrophic (e.g. every single person starves) are vastly more remote than the empirical conservative would have you believe.
Assuming the falsity of empirical conservatism’s thesis that society is fragile does not require rejecting empirical conservatism’s anti-radical attitude. It does, however, require recognizing its partiality. In light of the fact that even major social interventions are seriously unlikely to promote results that are catastrophic for all, empirical conservatism’s anti-radicalism is the proper posture only for some. Risk aversion is for people who have something to lose. For people whose lives are intolerable — slaves, criminals, the miserably poor, (until recently) gay people — a roll of the dice is rational. It wouldn’t be if the assumption of empirical conservatism that human society is always perched on the brink of unthinkable catastrophe was correct, but it isn’t.
In short, empirical conservatism is not a universalist panacea and politics remains a mired in dissensus. Who’d’ve thunk.