âMany people,â announces Gerry Mackie in his book Democracy Defended, ... treating Riker's attack on populism as an attack on democracy as such. This.
Public Choice (2005) 125: 471–475 DOI: 10.1007/s11127-005-4602-1
C
Springer 2005
Book review
Gerry Mackie, Democracy defended. Contemporary Political Theory, ed. by Ian Shapiro. Cambridge and New York: Cambridge University Press, 2003. xvi +483 pages. USD 85.00 (cloth); 30.00 (paper). “Many people,” announces Gerry Mackie in his book Democracy Defended, “suggest that democracy is impossible. The main purpose of this book is to argue against that view” (p. 16). Mackie’s primary target is William Riker’s Liberalism against Populism. In that book, Riker did not attack democratic decision-making procedures per se (voting for representatives, etc.), but instead argued for the superiority of one kind of theoretical grounding for those procedures – liberalism – over another – populism. Mackie, however, argues that “What almost everyone means by democracy is what Riker calls populist democracy” (p. 3). For this reason, Mackie believes he is justified in treating Riker’s attack on populism as an attack on democracy as such. This is a defensible move. Riker clearly believed that the intellectual defense of democracy that one accepted influenced one’s expectations of what democracy could and should do. Since Mackie takes certain of those expectations – expectations that Riker’s liberalism denies can be fulfilled – to be the heart and soul of democracy, it is understandable that he should regard Riker as antidemocratic. Unfortunately, Mackie’s critique of Riker’s argument is hard to follow, and his defense of populist democracy is weak. Indeed, Mackie seems to think that his critique of Riker is sufficient to qualify as a defense of populism, even though it raises more questions than it answers. In what follows, I shall briefly reconstruct what I take to be the heart of Mackie’s critique of Riker, noting points of confusion along the way. I shall then even more briefly describe the sort of constructive argument for which Mackie’s critique suggests the need, an argument Mackie himself does not provide. Mackie accuses Riker of professing “democratic irrationalism,” the belief that “the rational individual opinions or desires of citizens cannot be amalgamated accurately and fairly” (p. 3). Setting aside the question of fairness, however, it not at all clear as to what wanting to see opinions amalgamated accurately involves. Surely Riker was not denying that votes could be counted more or less correctly, with no hanging chads, etc. Instead, Mackie believes
472 that some amalgamation of the preferences is supposed to be an accurate measure of something, and takes Riker to be arguing that democratic decisionmaking procedures cannot do this. Despite, however, Mackie’s repeatedly expressed desire to investigate “the possibility of the accurate and fair amalgamation of opinions and wants” (Mackie’s emphasis; pp. 30–31), it is not until page 34 that he says anything about what this amalgamation is supposed to be accurately measuring, and then only in passing. There he points to the goal of having “democratic decisions variably approximate. . . the general will or public good” (Mackie’s emphasis). The general will and the public good, assuming both can be clearly defined, are conceptually distinct, despite Mackie’s efforts to conflate them here. One can coherently say, for example, that a group willed to send a lot of money to Africa in order to help starving children, and thereby acted against its own good for the sake of the good of others. Most of Mackie’s references from this point on, however, are to the “common good” or “public good” as the goal of the democratic process, and so it is reasonable to assume that he understands Riker to be denying that such a process could yield an accurate measurement of the common good. This interpretation of Riker is problematic. Granted, Riker appears to conflate the common good and the general will very much as Mackie does, defining the general will as “the objectively correct common interest of the incorporated citizens” (Riker, 1982, p. 11). But Mackie attributes to Riker the view that democratic decision-making procedures cannot accurately measure this common good. Riker did not argue this. If he did, he presumably would have indicated what alternative procedures might ascertain the common good better than democratic ones. Instead, he argued that there was no common good for any procedure (democratic or otherwise) to measure. “Populism as a moral imperative,” says Riker, depends on the existence of a popular will discovered by voting. But if voting does not discover or reveal a will, then the moral imperative evaporates because there is nothing to be commanded. If the people speak in meaningless tongues, they cannot utter the law that makes them free. Populism fails, therefore, not because it is morally wrong, but merely because it is empty (My emphasis; Riker, 1982, p. 239). Mackie occasionally recognizes this fact, as when he claims near the end of his book that Riker surreptitiously makes use of the idea of a common good in defining liberalism even while denying its existence when critiquing populism (p. 415). But if this is the correct way to understand Riker, then Mackie is simply confusing matters by claiming throughout the book that Riker denies that democracy can be “accurate.” One does not spend time denying the possibility of accurate measurement of something that one does not believe exists; one does not speak of measuring it at all.1
473 Riker based his contention that no common good exists on the various impossibility results generated by social choice theory, especially Arrow’s theorem. Each of these results specifies a set of properties that a preference aggregation rule must possess before one could claim that it says something about the common good, and then demonstrates that no rule possesses all of these properties. If populism requires that democratic procedures be valued for their ability to identify the common good, then the results of Arrow et al. suggest that populism cannot be the proper ideal against which to evaluate these procedures. In its place, Riker would substitute the ideal of liberalism, which sees the value of democratic voting in its ability to remove leaders who abuse their powers. It is this substitution that Mackie is concerned to combat. To be sure, Mackie does expose some serious problems in Riker’s argument. He does, for example, an excellent job of explaining just how big a loss would be incurred by embracing Riker’s rejection of populism and the notion of a common good. Those who reject populism, Mackie writes, claim that our intuition that there is a public good is an illusion. They do not realize that our moral intuitions are data of a sort that themselves we want to fit into some coherent scheme, that are constrained by considerations of coherence. Thus they do not notice, that without the idea of the public good we become unable to say that one political institution is better than another – our moral world no longer makes any sense. To give up the idea of the public good leaves only nihilistic criteria such as superior force or blind tradition for the evaluation of political institutions (Mackie’s emphasis; pp. 384–385). Any normative evaluation of institutions requires some idea of what is good for society. Indeed, as Mackie points out, Riker must make tacit use of some notion of the common good simply to define the “abuse of power” central to the liberalism he advocates as an alternative to populism – an alternative that is supposedly to be embraced because there is no common good (p. 415). The question democratic theorists must therefore face is not, as Riker believed, what should democratic theory say now that social choice theory has dispensed with the notion of the common good, but instead, how should we understand the common good in the wake of social choice theory. If, as Mackie contends, people cannot function in politics without some notion of the common good, then one of two conditions must hold. Either some property that social choice theorists have deemed essential to any rule that could be said to reveal the common good is not actually essential, and therefore some rule does possess all the truly essential properties; or else the common good is to be identified by some process other than the aggregation of preferences. An argument that successfully defends either condition would vindicate populism. Had Mackie offered such an argument, had he shown
474 exactly how social choice theory failed to foreclose on talk of a “common good,” he would have made an invaluable contribution to democratic theory. Unfortunately, Mackie offers nothing so constructive in his book. While he does provide extensive discussions of the impossibility results of Arrow et al., he never draws a clear answer as to how he believes the common good is to be determined. Instead, he offers a veritable blizzard of critical points and undeveloped suggestions, none of which cohere together very well. Mackie at one point claims that social choice theory is “indispensable even if some of it is sterile and scholastic” (p. 27). He thus appears to endorse the social choice project of defining the common good (assuming it can be defined) in terms of preference aggregation rules possessing certain properties. But it’s unclear what, if any, properties he endorses, and why. He spends time denouncing virtually every condition associated with Arrow’s theorem – both the admittedly controversial independence of irrelevant alternatives condition (chapter 6) as well as the seemingly unassailable weak Pareto (pp. 92–93) and nondictatorship (p. 92) conditions. The same problem occurs when he discusses the spatial model. He recognizes the potential instability of most aggregation rules when preferences are set in a spatial model, but claims that this poses no problem because the median of medians is the obvious “normatively attractive point of aggregate subjective welfare” (p. 191). Unfortunately, he provides no clear argument either as to why the median of medians is normatively desirable or why anyone should expect voting procedures to achieve it.2 Moreover, he gives no indication as to whether the spatial model or the more general axiomatic treatment of preference orderings is the more reasonable way to discuss preference aggregation with an aim to achieving the common good. If Mackie wants people to believe finding the common good is an exercise in preference aggregation, then he has little to say as to how exactly this aggregation should be done. But Mackie does not even consistently endorse the standard social choice framework of preference aggregation as a way of locating the common good. He occasionally speaks of maximizing the sum of social utility as a goal of the democratic process (p. 48). This requires that he admit cardinal rankings of utility to enter into the preference aggregation process, in opposition to most social choice theorists. Mackie recognizes this as a possible problem, but his only solution is to suggest that maximizing utility only occasionally should be the social goal (pp. 67–68). The circumstances under which maximizing utility should be equated with the common good, and why it should be so equated under some circumstances but not others, Mackie never says. Finally, Mackie is ready on some occasions to abandon the preference aggregation framework entirely, and treat the democratic process as an exercise in aggregating epistemic judgments a` la the Condorcet Jury Theorem (p. 55). Again, Mackie offers no defense for this view of democratic decision-making, and again he makes clear that he does not believe that democracy is all about epistemic
475 considerations (p. 64). But he never explains how or why the common good is to be realized through an epistemic process under some circumstances but not others. In short, Mackie’s book might convince a reader that a common good exists, but it will leave that same reader none the wiser about how the democratic process might identify this common good. Mackie’s book totals 483 pages, making it considerably longer than the book it was written to critique – Riker’s Liberalism against Populism. While the book offers some intriguing suggestions as to where Riker went wrong, it does not offer what one would expect from a book of such length – a coherent critique coupled with a constructive alternative argument. Anyone desiring a successful vindication of populist democracy in the face of Riker’s attacks will have to look elsewhere. PETER STONE, Political Science Department, Stanford University, Stanford, CA 94305–6044, U.S.A. Notes 1. In fairness to Mackie, Riker is not always consistent in his own choice of words. At one point, for example, Riker says that “Populism is supposed to reveal a substantive will, a proposition with content. Yet if voting can fail to reveal such propositions accurately and if we do not and cannot know in any particular instance whether failure has occurred, then none of the propositions supposedly revealed can be believed” (Riker, 1982, p. 291 no. 3; quoted by Mackie on p. 411). What Riker says here requires that he believe that there are indeed propositions that voting could more or less accurately reveal. This contradicts the thrust of the rest of his argument. (It’s also a rather silly point for him to make; by the logic he employs here, statistical inference would be impossible. After all, even extremely reliable statistical estimators can fail occasionally, and we “do not and cannot know in any particular instance whether such failure has occurred.”) 2. This is probably the appropriate place to mention that a huge portion of Mackie’s book – seven chapters, totaling 181 pages – is devoted to examining every single reported case of preference cycles in the social science literature, with an eye to demonstrating that none of them really involve cycles at all. Space prohibits me from discussing this part of the book, although it makes an interesting read. All I can say here is that if the theoretical issues discussed here are not properly sorted out, then the question of the prevalence of cycling in multidimensional space is moot. There’s no point in discussing the limitations of pairwise majority rule as a voting device – which is the focus of the entire debate about cycles – without a clear understanding of what voting devices are supposed to do.
Reference Riker, W. (1982) Liberalism against populism: A confrontation between the theory of democracy and the theory of social choice. Prospect Heights, IL: Waveland Press.
