Publications

This paper explores the relationship between two classic social evaluation procedures: the Borda count, and (an extension of) the Condorcet criterion. We provide a straightforward way of identifying and comparing those evaluation protocols, dispensing with the transitivity of individual preferences. Our approach uses individual pairwise comparisons of alternatives as informational inputs, with complete social orderings as informational outputs. We show that, keeping Arrow’s framework but weakening the property of independence of irrelevant alternatives to independence of separate pairs (the evaluation of each alternative only depends on how people compare this alternative with each other), opens the door to Borda and Condorcet evaluation functions. The key difference between these two protocols is the type of monotonicity assumed.