Dealing with incommensurability for dessert and desert: comments on Chapman and Katz.

• Author: Bix, Brian
• Position: Response to articles by Bruce Chapman and Leo Katz in this issue, p. 1487, 1465 - Symposium Comments: Law and Incommensurability

Incommensurability is the claim that different choices, or different values underlying choices, cannot be measured on a single metric.(1) The related, but different claim of incomparability is that there are items or choices of which it cannot sensibly be said that one is better than the other (or that the two are equally good).(2)

The phenomenon of incommensurability can be recharacterized as the irreducible plurality of values. The implications of incommensurability for practical reasoning, moral reasoning, and law are highly contested. Isaiah Berlin famously argued, in a series of books, that the plurality of values showed the basic flaws of monistic and Utopian strains in political thought, strains that go back to the Enlightenment and that include, in modern times, Socialism and Marxism.(3)

In moral reasoning, the fact of the plurality of goods is often used to argue against the possibility of utilitarianism, and its sister forms of analysis, law and economics and rational choice theory. Indeed, the phenomenon of incommensurability would appear to pose problems (though not necessarily insoluble problems) for all forms of analysis that depend on the ability to reduce all options to a single good that is being optimized, or to a single continuum of measurement.(4)

It is sharply contested whether incommensurability entails incomparability: that is, whether incommensurability means that one cannot speak meaningfully of one option (which instantiates one value) being better than a second (which instantiates a different value).(5) The claims that choices can be incomparable, and that this sort of incomparability occurs with some frequency, would seem to threaten basic aspects of the way we think about morality, aesthetics, law, and dozens of other aspects of our lives.(6)

Even if incommensurability does not entail incomparability, there remain interesting connections to draw between incommensurability and moral or practical reasoning. It is tempting to ask, if options cannot be measured on a single metric, how can rational choice (understood in its colloquial, not its technical, sense) be possible? The paradoxical, but (I believe) persuasive response is that it, is incommensurability that makes rational choice possible. If all options were reducible to units of some good an individual sought to maximize, there would be no need for "choice," understood as an act of judgment. Any automaton can choose $500, when the alternative is $100; and similarly if one way of life "equals" 5000 "units of happiness/contentment" while an alternative way of life "equals" only 1000 units.(7) The difficult choices in life are not like these; they involve a sense that whichever way one chooses, one will be giving up something of significant value. Incommensurable choices are the choices that are later viewed with some regret, or even labeled as "tragic," even when one continues to believe the choice made is the correct one.(8) For example, one reasonably might choose to spend more hours working in order to create greater economic security and opportunity for one's family, but might regret (both now and later) the lost chance to spend more time with one's family.

The two articles on which I am commenting for the most part stay clear of such general reflections on the nature and consequences of incommensurability, considering instead what implications might follow once we accept the irreducible plurality of values. Bruce Chapman, in Law, Incommensurability, and Conceptually Sequenced Argument,(9) tries to show how our decisionmaking processes often find ways to give plural values distinct and ordered consideration, and how legal decisionmaking may offer a number of particularly effective instances of this possibility. Leo Katz, in Incommensurable Choices and the Problem of Moral Ignorance,(10) demonstrates some difficulties in one usual understanding of incommensurability, and also uses incommensurability as a starting point for an inquiry regarding moral intuitions, deontological logic, and the criminal law. I will discuss the articles in turn, offering in passing some reflections on the connections between the two.

1. SEEKING, ORDERING, AND AVOIDING ISSUES

   Professor Chapman takes as his starting point that values are plural and that some sets of values are incommensurable. He notes that this already puts him at odds with traditional economic analysis and rational choice theory. Those approaches assume that all options can be ordered based on people's preferences: Even if we claim to be unable to choose between, say, friendship and money, or to reduce one to the terms of the other, our choices belie our assertions. By being willing to break a lunch date with a friend to pursue a lucrative opportunity, but not being willing to break the appointment for a far less lucrative opportunity, we have, the argument goes, begun to specify the dollar value of that friendship. Similarly, by being willing to accept an increased risk of death in exchange for an increased pay, with sufficiently subtle calculation, we can figure out how much our lives mean to us.(11) Many writers have argued that this inference and equation is not justified.(12) While Chapman does not focus particularly on that debate, his article opposes the foundational assumptions of economics and rational choice theory in another, perhaps more indirect way: by showing that the consideration of plural values in a noncommensurating way is both common (in both individual and institutional decisionmaking) and sensible.

   Chapman's portrayal of "conceptually sequenced argument"(13) offers an alternative to two different sets of traditional (economic) positions. First, regarding the commensurability of values: (a) many argue that there is only a single value or that all values can be effectively combined in a single evaluation; and (b) those who recognize problems with commensurating some values often hold out as the only alternative the lexical ordering of values, in which one value has absolute priority over the other(s). Against those positions, Chapman argues that we can accommodate plural values in a decision procedure, without either assuming that they can be commensurated on a single metric or requiring that one value always maintain absolute priority over others (that is, allowing for at least a limited tradeoff of values). Second, when values or criteria are applied in a sequence, the conventional view appears to be that there can be no reason or order in the way that a series of evaluations is done.(14) Here, Chapman's contrary position is that there can be reason for choosing one partition rather than another, and one path rather than another, in creating a sequenced decision.

   Chapman's position is that although there is no way sensibly to combine plural values in a single evaluation, one can have a multipart evaluation that gives a number of different values their individual due. Additionally, in creating the multiple-step process of evaluation, some ways of partitioning the items to be evaluated, and some ways of ordering the types of evaluation, will be better than others.(15)

   For some people who work within rational choice theory, one of the most contentious aspects of Chapman's article is his conclusion, partly argued for and partly assumed, that there are (or can be) failures of transitivity in people's preferences, and that these failures, when they occur, are or can be sensible (that is, not "irrational").(16) While I note the importance of the issue and the possibly controversial nature of Chapman's position, I must leave that issue to others. To the extent that taking a position is essential for the purposes of my argument, this Comment will side with Chapman by assuming that there is intransitivity, and that such intransitivity is not necessarily irrational.

   Chapman illustrates how the sequence in applying criteria of selection sometimes affects the evaluations, in the sense that different orders will yield different outcomes.(17) He gives the example of the diner whose dessert choice will vary depending on the order in which pairs of fruit are offered to her. The differing outcomes derive from whether the value of etiquette is presented, which will occur if, and only if, the diner must choose between a small apple and a larger one. In such a circumstance, the value of etiquette overrides hedonistic values, and the diner would choose the smaller apple. If the fruit is presented in a different order, the choice between different-sized apples would never arise, and a different overall outcome would result.(18)

   This sort of situation presents the issue squarely, as Chapman notes:(19) We sometimes can structure choices in ways that avoid reaching certain values or questions, or in ways that expressly seek to consider and confront such values or questions.

   Is there thus a general principle that we should try to structure our (individual, institutional, or social) decisions in such a way that all issues (values) that might be raised will be raised? Is it the case that since one way of partitioning the dessert-choice problem would raise issues of etiquette, selection procedures or partitions that invoke etiquette are to be preferred to ones that do not? This seems to be Chapman's position,(20) and there is a sense in which that approach, at least as an ideal (limited by the constraints of time, resources, and imagination), seems right. However, we must consider at least two responses to this intuition.

   First, we encounter a practical problem: As with choosing a dessert in a selection among three choices, the number of possible issues may be small and manageable; however, when the choice is among a larger number of more complex options (which career path to choose, or which candidate among a dozen should be given a scholarship), numerous issues which could be considered (and which might become salient if just the right pair of alternatives were combined)...