Owen Densmore wrote at 11/08/2012 08:36 AM: > The 1 & 2 party systems are the only ones avoiding the pitfalls of Arrow's > Impossibility Theorem. > > http://www.udel.edu/johnmack/frec444/444voting.html
> 1. If and individual or group prefers A to B and B to C, then A is > preferred to C (transitivity). > 2. The preferences must be restricted to the complete set of options. > 3. If each individual prefers A to B, then the group must also. > 4. No individual's preferences can necessarily dictate group preferences. > 5. The group's pairwise preference ordering is independent of irrelevant > alternatives, i.e. determined solely by individual's pairwise preference > orderings. I'm sure I'm being dense. But I don't see any need for rules 2, 3, or 5. And 1 is suspect, as well. So, I wouldn't accept this as an argument against >3 viable parties. Can each of these rules be defended? ... with any kind of evidence (as opposed to ideology)? > So I wonder what's it like in a true multi-party system like most of Europe > has? Is it effective? interesting? confusing? fun? Are the populations > aware of Arrow? Does it avoid grid-lock? I've been told (sans evidence) that multi-party systems risk a situation where each party represents a geographical region. I can also _imagine_ that parties would form around single (or clusters of) issues. That sort of thing makes me think that there should be an upper limit on the number of parties. But what's the limit? And what's the limit a function of? Perhaps the limit could be a function of (clusters of) land area, population diversity, and issue diversity? For example, I'd love to have two axes, in the US: fiscal (conservative vs. liberal) and social (conservative vs. liberal). I can imagine this would nicely lead to a party limit of 9: 1. Fiscal Conservative (FC), Social Conservative (SC) 2. Fiscal Moderate (FM), SC 3. Fiscal Liberal (FL), SC 4. FC, Social Moderate (SM) 5. FM, SM 6. FL, SM 7. FC, Social Liberal (SL) 8. FM, SL 9. FL, SL If there's an upper limit, then there should probably be a lower limit. If the limits are based on clusters of region, demographic, and issue, then there can never really be a single party. Perhaps a utopian ideology would allow it, but no reality would. I can, however, imagine a large distance between the most important issue (say emergency preparedness or WAR!) and the rest of the issues. That scenario would allow a single axis with a party on each side and perhaps in the middle. That implies that 2 or 3 is the lower limit. Frankly, if someone started a "moderate" party, I might actually register as a member, something I've never done and will never do as long as there are only 2 nationally viable parties. One thing that would be interesting is if I were allowed to affiliate locally with 1 party but state-wide with another, and nationally with yet another. -- glen ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org