[ExI] Fwd: elections again; was [Time Magazine: Person of the Year: Putin(!),my vote instead: Anna Politkovskaja]

J. Andrew Rogers andrew at ceruleansystems.com
Fri Dec 28 19:45:08 UTC 2007

On Dec 28, 2007, at 10:00 AM, daniel radetsky wrote:
>> On Dec 27, 2007 4:48 PM, J. Andrew Rogers < andrew at ceruleansystems.com 
>> > wrote:
>> As a meta-comment, "fair elections" as most people would describe  
>> them
>> are generally a hard problem, even if every vote is always counted
>> correctly.  See: Gibbard-Satterthwaite Theorem, Duggan-Schwartz
>> Theorem, and Arrow's Theorem (among others).
> These only apply to ranked choice voting schemes. See Range Voting.

Range voting is not immune to the basic problems of ordered schemes,  
and appears to collapse to something equivalent to conventional  
systems in most cases.

> If you assume that certain properties are non-negotiable, and those  
> properties imply manipulability, then manipulability is guaranteed.  
> But why should we assume that properties implying manipulability are  
> non-negotiable? If a property leads to manipulability, that's a sure  
> sign that it ought to be negotiated.

While it has not always been true in the past, concepts like "one man,  
one vote" and "every vote counts equally" are going to be really hard  
to negotiate away today.  If you look at the various alternative  
conditions for maximizing preferences that cannot (easily) be gamed,  
they are culturally unacceptable for the most part, at least in  
western European culture.

The philosophical nature of the process is quite apparently *at least*  
as important as the quality of the result.  There is an assumption  
that people choose voting systems primarily to ensure a particular  
result with respect to population preferences but it seems obvious to  
me that people choose voting systems to primarily ensure particular  
voting rules.

J. Andrew Rogers

More information about the extropy-chat mailing list