This collection of articles on satisficing is essential reading for anyone interested in this topic in practical reason, but at the same time it should put the topic to rest for philosophers, at least for the foreseeable future. On the one hand, a weak thesis regarding satisficing is universally acknowledged to be true, but no longer philosophical news. On the other hand, a stronger thesis seems clearly false despite heroic efforts by its few advocates to revive it with different justifications (initially only unconvincing purported examples) each time.
The weak thesis, first put forward by Herbert Simon, views satisficing as a rational stopping rule for deliberative decision-making under uncertainty. Lacking adequate knowledge, an agent does not attempt to assign probabilities and utilities to the possible outcomes of alternative actions (necessary for maximization), but instead chooses an option with all satisfactory outcomes. This is a time- and resource-saving strategy for choice under uncertainty, and as such, it is used locally for the sake of global optimizing. It is rational because the pursuit of particular goods competes for resources against the pursuit of other goods. We should not maximize any quantity that is not a measure of overall good; optimizing involves trade-offs to achieve the best mix. As David Schmidtz and Jan Narveson point out in their contributions to this collection, such local satisficing makes our lives go better overall, given our limited budgets of time, energy and other resources. Living the best lives we can generally does not involve trying to maximize wealth, brilliance and beauty in our companions, fame in our careers, or other particular goods.
In fact, optimizing overall, if that involves aiming at the best mix of a true plurality of diverse goods, does not involve aiming to maximize anything. Globally, there is no single good or value to maximize and hence nothing in regard to which to satisfice either. As several of these authors, including James Dreier, point out, personal utility, as a theoretical construct from the choices of agents that meet certain constraints such as transitivity, is assumed to be rationally maximized. There is no question for the self-interested agent of satisficing in regard to it, and nothing else at the global level to fit this bill. But while we neither maximize nor satisfice globally, we do try to optimize, to lead the best lives we can.
In regard to particular goods, we have seen that no one would deny satisficing as a cost-saving strategy in the service of global optimization. The stronger thesis is that it can be rational to choose less than the best overall among known alternatives. It seems clear that this thesis cannot be right. To be rational one must act on one's strongest reasons. Reasons count in favour of actions: they are what make actions better than alternatives. If satisficing is rational, there must be reasons in favour of it as strong as the reasons favouring maximization. But these reasons then make satisficing better or optimal from the global perspective. It will not then involve choosing less than the best known option or acting in less than the best way overall. For global satisficing to be rational there must be reasons in favour of an option that do not make that option better or better to choose, but that does not seem possible. Variations on this argument are offered by contributors Schmidtz, Narveson, Dreier, van Roojen, and Byron.
Michael Slote, the original defender of the stronger thesis, as well as Michael Weber and Thomas Hurka, argue for it in different ways. Slote, following Christine Swanton, here argues that strong satisficing can be seen as a manifestation of the virtue of moderation, itself constitutive of certain goods. There are three problems with this position. First, if it is better to be moderate in order to have the goods that moderation generates, then moderation is optimal, not less than best overall. Second, if moderate persons are perfectly content with moderate amounts of some good, then seeking more of it is not better for them. Hence once more they are not choosing less than is best for them. Third, as Schmidtz points out, what is satisfactory for an individual may not be moderate: moderation and satisficing are distinct.
In his argument, Weber posits two temporal perspectives on life: the momentary and the whole-life perspectives. What is significant from the former perspective, particular pleasures and pains for example, is negligible from the latter, while pursuing only what is significant from the point of view of one's life as a whole involves sacrificing momentary goods. Weber then argues that satisficing in each of these domains for the sake of the other is rational. What he fails to show is that this is not just another case of trade-offs to achieve the best mix of goods so that our lives go as well as possible. Once more there is no single good available from both perspectives to maximize, but this is compatible with an overall goal of optimizing. Competition among goods or perspectives breeds satisficing, but only in the weak sense.
Several of the authors, including Hurka and Dreier, endorse moral satisficing. Hurka points out that we are not required to ensure others' maximal happiness even if we could; it is permissible if within our power to provide them with only a satisfactory level of well-being. We are required only to prevent misery. Dreier notes that a moral theory with defeasible side constraints calls upon us to satisfice with respect to goodness before applying the side constraints. Checking that the lives of millions are not at stake before we agree not to torture does not seem to be satisficing, however, but avoiding disaster. And even if we grant that satisficing is the proper term for minimal moral requirements, such requirements to produce only less than the best we might provide derive yet again from the fact that moral considerations compete against others, usually self-interested reasons but also possibly aesthetic ones, and so on. It is such competition that makes some acts supererogatory and allows one not to perform them. But this does not imply a category of the supererogatory in the domain of the globally rational: it cannot be a virtue to be moderately rational.
