Factions in Rousseau's Du Contrat Social and federal representation

Abstract

Consider the following two seemingly unrelated questions. First, why does Rousseau (1993 [1762]) believe that the formation of factions or partial associations is not conducive to the general will in Du Contrat Social, II, 3? Second, why do federal assemblies typically strive for some form of degressive proportionality, i.e. a balance between equal and proportional representation, for the countries in the federation? We will show that there is a surprising connection between these questions. We turn to our first question. It is often thought that Rousseau's oppo sition to factions can be interpreted in reference to the Condorcet Jury Theorem (e.g. Grofman and Feld 1988). The Condorcet Jury Theorem states that if voters are more likely to be right on some issue than not and cast their votes independently, then the chance that the majority is correct converges to one as the number of voters goes to infinity. This interpre tation is supported by Rousseau's claim that 'If, when a sufficiently informed people deliberates, the citizens were to have no communication between themselves, the general will would always result from the large number of small differences' (II, 3, emphasis added). If voters form fac tions, then they no longer vote independently. They become subject to opinion leaders and the epistemic advantage of consulting a large number of voters is lost. Or, as Grofman and Feld (1988: 571) put it, the effective number of votes is reduced, which in turn reduces the probability that the majority is correct. This interpretation is tempting, but it is inconsistent with some crucial textual evidence. Rousseau writes, 'If there are partial societies, it is necessary to multiply their number and to prevent inequality from existing among them ... These precautions are the only good ones for ensuring that the general will is always enlightened' (II, 3). Now if Rousseau is interpreted in terms of the Condorcet Jury Theorem in the way described above, then multiplying factions only makes things worse. The effective number of votes is further reduced. Thus, even if factions form, one should retain the independence between the remaining parties as much as possible to safeguard whatever epistemic advantage the majority vote still has. An alternative interpretation of Rousseau's opposition to factions is called for. We turn to our second question. Consider a federation of countries with a decision-making assembly in which each country casts a block vote. On