The Values of Democratic Proceduralism

Abstract

A standard justification of democratic voting is that it is a fair procedure, providing for the equal treatment of voters. Democratic theorist David Estlund challenges the adequacy of that justification: flipping a coin between alternatives is also a fair procedure, but no one would propose substituting random draw for voting. Estlund provides several arguments that fair proceduralism is an untenable view, and this article counters those arguments. He concludes that what distinguishes democratic voting from random choice is its better epistemic value in approximating a standard of justice independent of the procedure. The author replies that a less controversial distinction between voting and coin flip is that voting tends to select what is thought best by the most people.

Keywords:

• democratic theory
• voting
• procedural value
• David Estlund

Acknowledgements

The author thanks Eddie Hyland for a long conversation about democracy in Dublin, and more importantly for his accomplishments in democratic theory. The author thanks Corey Brettschneider and Thomas Christiano for thoughtful comments and David Estlund for patient and thorough comments on two drafts, correcting many confusions and errors.

Notes

Riker's views are rebutted in Mackie (2003).

Democracy is much more than majority vote over two alternatives. This sketch concentrates on that for simplicity of exposition. A fuller argument would have to incorporate a more general account of voting, public deliberation, delegation including representation, constitutionalism and other features of modern political democracy.

The consistency of various conditions with a given voting procedure does not exhaust the meaning of that procedure, because the procedure behaves differently with different patterns of inputs, actual humans in actual votes may act in ways not contemplated by the formal model, and so on.

May (1952: 681) calls it the equality condition, says it has to do with equal influence of individuals on the outcome, and says that his ‘simple majority decision’ is meant to be the same as Arrow's ‘method of majority decision’. Arrow's (1963: 46) definition of the method of majority decision does not refer to symmetric inputs but instead refers directly to counting the number of individual humans voting for each alternative.

Estlund's (2009) replies to Christiano are implicitly addressed at various points herein; to rehearse explicitly those arguments and my responses to them is not possible within the available space.

Each of the five has preferences not only over their most favored alternatives, but also over their less-favored and least-favored alternatives. The idea of preference is not limited to top-ranked preference.

It is a mistake to consider plurality voting, which considers only each voter's top-ranked alternative, the exemplary democratic voting rule. Voting rules that aggregate all of each voter's rankings allow us to talk about an alternative that, although not top-ranked by any, is most wanted by all, as will be shown in the upcoming example.