[EM] Approval and Condorcet

[Juho Laatu]

On 2.6.2012, at 15.32, Jameson Quinn wrote:

> There's been a lot of back-and-forth over which is better. As an Approval supporter myself, but one who doesn't agree with a lot of the pro-approval arguments that have been made, I'd like to state my own position – once. I won't respond in this thread because I hope the whole debate dies out soon.
> 
> Is it possible to support Approval over Condorcet? Condorcet over Approval?
> 
> Manifestly, yes. Further argument will not change these facts.
> 
> Is it logical to support Condorcet over Approval? Approval over Condorcet?
> 
> Again, yes to both.
> 
> Condorcet over Approval: 
> Approval lacks a unique definition of "honest votes". But for some such definitions, which (however misguidedly) actual real-world voters spontaneously believe, honest Condorcet clearly does better (by any reasonable definition of "better")
> Some voters truly do not like ambiguity of how to vote. This is an aesthetic preference and not subject to logical browb- I mean, argumentation.
> Some voters do not like being strategically forced to elide the preference for their favorite over their second. Again, aesthetic preference, though it would furthermore have the consequence of more spoiled elections under approval insofar as voters eschewed rational strategy for this reason.
> Approval over Condorcet:
> There are voter models for Approval which lead to better satisfaction (BR) than ANY voter models for Condorcet.
> Dishonest strategies can be rational under Condorcet, 
> This can rationally lead to strategic results far worse than Approval.
> If dishonest strategies are irrationally overused, it gets even worse.
> Some voters have aesthetic objections to the dilemma of whether to vote rationally or honestly.
> Given that reasonable people can take either side of this debate, what's the point of arguing?
> 
> It's only productive insofar as it helps us unite our activism. Since we're never going to agree 100%, that is very limited. Still, I think it's worth pointing out that approval is much, much simpler, and thus seems more likely to be implemented in the short term.
> 
> Is it possible to see either Approval or Condorcet as worse than plurality?
> 
> Possible? Again, manifestly yes.

Two possible approaches to answering yes:

1) It is possible to think that a two-party system is the ideal political system. It certainly has some well known benefits like e.g. very efficient and unified governments when compared to the other common approach, proportional representation. Also people that feel that the system does not work satisfactorily at the moment may seek reform within the framework of Plurality and two parties (e.g. changes only in the financing of the political parties).

2) Also people who want to have a reform that removes the limitations of Plurality and two parties may oppose this proposal because the whole system has been designed or has evolved around the idea of "alternating two-parties". Changing just one component to a different one with different nature, without replanning the whole system and all the components that this one change will impact, could make the dynamics of the political system worse, not better.

Juho



> 
> Are those positions logically defensible?
> 
> With apologies to Bruce, I'd have to say no regarding approval; there is no defensible reason to prefer plurality over approval. Bruce argues that since only two candidates are viable under plurality, it never forces him to choose between supporting his favorite viable candidate over his second-favorite, or his second-favorite over one he dislikes. It is true that approval can be harder to vote in this way. However, by the same argument, dictatorship never forces you to make any difficult choices at all, and thus should be preferred to plurality. The fact is, plurality has "only 2 viable candidates" precisely because it's reduced your choices, and often (I'd argue usually) that's because it has implicitly eliminated the best choices before you even see them. 
> 
> Regarding Condorcet... well, I can't be as categorical. There's no logical reason to prefer plurality's results. (And no, Condorcet's strategy incentive is not strong enough to lead to real DH3 scenarios worse than plurality; because plurality can and frequently does lead to worst-two-are-viable scenarios anyway.) But I can't completely disparage the argument that Condorcet is too complex, even though I don't buy that argument.
> 
> Bonus factious argument: Is there an amoeba's worth of distance between Democrats and Republicans?
> 
> Yes, there is. The median prominent Democrat politician today (for any prominent set size, including measure 1) is to the right of Richard Nixon on many issues (I'd say more than half of them where it's possible to check), and there are several issues (I'd say important ones) on which the parties are indeed functionally indistinguishable. But it's just crazy to say that they're the same; in fact, objective measures of voting records show that the gap is wider now than any time in the last 100 years. You just can't debate that, there's many ways to prove it. US democracy is indeed very sick, but hyperbole discredits only the person who can't abandon it.
> 
> Again, I hope this argument dies soon, so I won't contribute any further to continuing it; but I wouldn't mind it if others who haven't already made their points ad nauseum would state their position once, as I have here.
> 
> Jameson
> ----
> Election-Methods mailing list - see http://electorama.com/em for list info

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120602/7a3255b2/attachment-0004.htm>