Response from David Brink on November 1, 2007

Insofar as a system is chaotic, it is impossible to predict all the consequences of actions we perform. Many systems, while not literally chaotic, are still complex, with the result that predictions about the total consequences of one's actions are difficult and not fully reliable. But it's not clear that these facts about complexity and chaos compromise the claims of classical utilitarians, such as Mill. The classical utilitarians (e.g. Bentham, Mill, and Sidgwick) all insisted that utilitarianism -- the demand that we act so as to promote human happiness -- be understood as a standard of duty or right action, rather than as a decision procedure. That is, what determines the moral assessment of one action in relation to others is the comparative value of its actual consequences and not the comparative value of its expected consequences. Whether we should try to apply the utilitarian principle in various contexts is itself a practical question, the answer to which depends upon the comparative consequences of doing so. Calculating utility is time consuming (and so has opportunity costs) and is often subject to ignorance (as in complex or chaotic systems) or bias and distortion (e.g. due to self-interest). For this reason, we may often better promote utility, not by trying to, but rather by adhering to various discrete moral precepts that make no direct reference to utility, such as familiar precepts about non-aggression, honesty, fidelity, and fairness. Mill called these precepts seconadry principles and insisted that the utilitarian first principle, which supplied the standard of right conduct, be supplemented by various secondary principles. He thought we should review the value of our secondary principles from time to time, and that we should resolve conflicts among secondary principles by direct appeal to the principle of utility. But otherwise our conduct should be more or less strictly regulated by such secondary principles. Of course, in cases of conflicts among the secondary principles, direct appeal to utilitarian calculation may prove unreliable, due to complexity or chaos. All we can do, it seems, is do our best. This won't determine what is best to do -- that's a matter of what in fact maximizes utility -- but it may mean that what we do is blameless, even if it results in action that is suboptimal (i.e. wrong, ny utilitarian standards).