Voting systems may differ in what they try to accomplish. Some philosophies are:
Majoritarianism[edit | edit source]
A majoritarian voting method passes the majority criterion, which means that if a majority prefers a certain outcome (or set of outcomes) to all others, that outcome (or set) wins.
A deterministic majoritarian system can be seen as a DSV method if based on a method where a majority can always force an outcome if it votes in a strategic manner. The reasoning for always letting a majority win is then the same as for any DSV method: that by doing automatically what a base method can do under strategy, one levels the playing field. However, it has been argued that majorities should not automatically overrule minorities simply because they have the opportunity to do so; that the voters should be able to not maximally escalate their opinion unless they feel the situation calls for such extreme measures.
Alternatively, majoritarianism can be argued as a default position when it's impossible to know the voters' utilities, but that every voter's voice is of equal importance and that the voters can't express utilities on an absolute scale. Then more voters preferring an alternative is preferable to fewer voters doing so, and in a two-candidate election, the candidate preferred by a majority should win.
Total utilitarianism[edit | edit source]
A total utilitarian (often called just "utilitarian") method attempts to choose the outcome or set of outcomes that maximizes the electorate's utility. These methods are in the spirit of utilitarian ethics.
One problem with strict utilitarian methods is incommensurability, which is that different people may have different utility scales and that there is no clear way to calibrate the ratings in a cardinal method so that summing the ratings also sums the voter's utilities. As such, cardinal methods are sometimes said to approximately maximize utility rather than to achieve the goal of maximization. Bayesian regret simulations can be used to argue that even given incommensurability, cardinal methods come closer to maximizing utility than do majoritarian methods.
Minimax utility (Rawlsian justice)[edit | edit source]
A minimax utility method attempts to choose the outcome or set of outcomes that provides the greatest benefit to the least advantaged: in utilitarian terms, that maximizes minimum utility. These methods are associated with consensus or unanimity and are in the spirit of John Rawls' theory of justice.
Since the minimum and maximum statistics are not robust, many of these methods are particularly vulnerable to strategy. A voter might claim that a particular candidate would be absolutely unacceptable, and thus force the method to choose someone else. However, there exist nondeterministic methods that approximate (while not exactly achieving) the minimax ideal, using randomness to deter strategy.
Typical deterministic minimax utility methods are Minimax approval.
Typical nondeterministic minimax approximation methods are the Nash lottery and Maximal Partial Consensus.
References[edit | edit source]
- Heitzig, Jobst; Simmons, Forest W. (2020-06-10). "Efficient democratic decisions via nondeterministic proportional consensus". arXiv:2006.06548 [cs, econ, q-fin].