American Political Science Review Vol. 101, No. 3 August 2007 DOI: 10.1017/S000305540707030X Democratic Theory and Political Science: A Pragmatic Method of Constructive Engagement ARCHON FUNG Harvard University T his article develops two conceptual tools to synthesize democratic theory and the empirical study of institutions. The first is a standard to assess conceptions of democracy called pragmatic equilibrium. A conception of democracy is in pragmatic equilibrium just in case the consequences of its institutional prescriptions realize its values well and better than any other feasible institutional arrangements across a wide range of problems and contexts. Pragmatic equilibrium is a kind of Raw- lsian reflective equilibrium. The second is a method of practical reasoning about the consequences of alternative institutional choices that brings conceptions of democracy closer to pragmatic equilibrium. These two ideas are then applied to four conceptions of democracy—–minimal, aggregative, deliberative, and participatory—–and to two governance problems—–deciding rules of political structure and minority tyranny—–to show how each conception can improve through reflection on the empirical consequences of various institutional arrangements. The tragedy of the world is that those who are imaginative have but slight experience, and those who are experienced have feeble imaginations. Fools act on imagination without knowledge, and pedants act on knowledge without imag- ination. (Alfred North Whitehead, “The Universities and Their Function” [1927]). S harp boundaries currently separate normative political theory from empirical political science. Rogers Smith (2003, 76) notes that the major works in political theory from the 1960s through the 1980s composed by Strauss, Rawls, Nozick, Derrida, Dworkin, Foucault, Ackerman, Riker and Habermas “display only limited direct engagement either with contemporary political issues or with empirical social science.” Ian Shapiro (2003, 2) laments that “Norma- tive and explanatory theories of democracy grow out of literatures that proceed, for the most part, along separate tracks, largely uninformed by one another.” This division of labor has become a segregation of thought that now poses a fundamental obstacle to progress in democratic theory. Debates among the ma- jor contending conceptions of democracy—–direct, rep- resentative, participatory, minimal, deliberative, and aggregative—–proceed largely without the benefit of empirical evidence about whether the arrangements and practices they recommend are feasible, could be stable, or whether they would produce desirable out- comes. Democratic theorists frequently help them- selves to so-called “stylized facts” to construct argu- ments that are valid but not necessarily true. Ascer- Archon Fung is Associate Professor of Public Policy, John F. Kennedy School of Government, Harvard University, 79 JFK Street, Cam- bridge, MA 02138 (archon fung@harvard.edu). I am indebted to Eric Beerbohm, Joshua Cohen, Elisabeth Gar- rett, John Gerring, Gary Herrigel, Jane Mansbridge, Philip Pettit, Charles Sabel, Lee Sigelman, David Austen-Smith, Clarence Stone, Dennis Thompson, Chris Winship, Kenneth Winston, and three anonymous APSR reviewers for comments that guided the devel- opment of this article. Previous versions were presented at the 2005 American Political Science Association meetings, the 2006 Chal- lenges to Participation Workshop of the California Institute of Tech- nology/University of Southern California Center for the Study of Law and Politics, and the 2006 Deliberative Democracy Conference at Princeton University. taining truth requires replacing those stylized facts with real ones. That, in turn, entails bridging normative and empirical research. But how, precisely, should demo- cratic theorists engage with empirical political science? At least two research strategies have combined nor- mative and empirical inquiry. Some scholars, often seeking to identify more desirable forms of politi- cal life, have sought to understand particular insti- tutions that exhibit attractive democratic character- istics. Carole Pateman and Jane Mansbridge—–two of the most important theorists of participatory democracy—–studied actual worker-managed organi- zations (Mansbridge 1980; Pateman 1975) to develop their ideals of democracy. Similarly, scholars have pur- sued inductive research on institutions such as the New England Town meeting (Bryan 2004; Mansbridge), urban bureaucracies (Cohen and Sabel 1997; Fung 2004), and experimental deliberative arenas (Fishkin 1995). This research strategy has revealed democrati- cally promising institutional possibilities. But the dif- ficulty of generalization is its fundamental limitation. The successes of these practices often depend upon the rarified circumstances of their construction. It is therefore difficult to connect these inductive projects to broader conceptions of democracy that are intended to apply across a wide range of problems, circumstances, scales, and times. William Galston (2003), among others, notes that “grand political theories are rich sources of empirical conjectures.” Political scientists have extracted falsifi- able hypotheses from democratic theories or used em- pirical findings to qualify and criticize conceptions of democracy. Studies of juries, for example, reveal that their deliberations often reflect gender and social bi- ases (Sanders 1997). Cass Sunstein (2000) has found that deliberation can cause participants to polarize rather than to moderate their views or reach consensus. In a recent book, Diana Mutz (2006) uses survey evi- dence to argue that there is a trade-off between the so- cial heterogeneity that democratic theorists favor (e.g., Young 2002) and the extent of political engagement. Decades of research in political participation has es- tablished that well off individuals participate more than 443