[EM] Re: Voting Systems Study of the League of Women Voters of Minnesota

[Abd ul-Rahman Lomax]

At 12:50 AM 6/9/2005, Russ Paielli wrote:
>Abd ulRahman Lomax abd-at-lomaxdesign.com |EMlist| wrote:
>
>>So promoting Approval voting might be as simple as pointing out the 
>>injustice of [discarding overvoted ballots]. I can't see any reason for 
>>*preventing* a person from voting for more than one candidate. Allowing 
>>it merely adds to the freedom of the voter without complicating the 
>>process. For me, the question is "Why not" rather than "Why?"
>
>You make an excellent point. Rather than defending Approval, Approval 
>advocates should go on the offensive and let the opponents explain why the 
>voter *shouldn't* be allowed to approve more than one candidate.
>
>Having said that, let me play devil's advocate and give you a potential 
>reason.

One that I had not thought of, probably because I don't think of democratic 
processes, in general, as properly "forcing" voters to do anything.

>The basic principle of voting is that you, the voter, are supposed to 
>specify who *you* think should be elected. If you were the only voter, 
>your vote should choose the winner. But if you approve more than one 
>candidate, you have not specified who you think should be elected. You 
>have not made a final decision. You have only narrowed the field.
>
>Plurality forces you to make up your mind!

This is correct. Note that this is basically an argument for IRV. IRV with 
overvoting allowed (I'm trying to get in the habit of not calling it 
Approval, even though the effect of counting overvotes is to implement 
Approval voting) allows the kind of preference to be expressed that the 
devil would be claiming it is proper to force the voter to express.

Typically of the devil's arguments, a partial truth is used in an attempt 
to generate a false or misleading conclusion. The most clever liar will lie 
with the truth, it is done all the time.

The partial truth is that discarding overvotes essentially punishes the 
voter for failing to make a choice (within the approved set). However, the 
goal of elections is not merely to produce a winner. A coin toss will do 
that quite handily. In some jurisdictions, a tie vote is indeed resolved 
with a coin toss, and the hypothetical situation of the single voter 
marking two candidates could likewise be expeditiously resolved in the same 
way.

In that election, also, if there were only two candidates, the voter could 
express equality of preference simply by not voting for either. It is only 
when more than two candidates are involved that overvoting becomes a method 
of duplicating an option which is already available in single-candidate (or 
likewise with yes/no questions).

Indeed, this is exactly what I do when I have no preference or, what 
amounts to the same thing, I do not have sufficient knowledge to make an 
intelligent choice. In that case, my decision is to leave the decision to 
others, I don't mark the ballot for that race or question.

Since it is possible for a voter to express equal preference in a 
two-candidate election, why should the voter not be allowed to express the 
same if an additional candidate is added to the ballot.

>Well, that's one way to look at it.

There are as many ways to look at things as there are people. I really 
appreciate Mr. Paielli's examination of this question. We should have a 
conversation with the devil more often. By playing devil's advocate, Mr. 
Paielli has made that possible for us. The *real* devil will try to avoid 
conversation, for he knows that open conversation will expose his schemes. 
So, usually, someone must play his role in order for these deceptions to be 
exposed.

(I'm using the word "devil" for "the force of deception that operates, as 
it says in the Qur'an, from a place where we do not recognize him. My 
opinion is that this place is within ourselves. What I write about "the 
devil" in this forum and in other fora is intended to refer to a 
personification of that force, which *does* exist in the way that other 
abstractions exist, quite clearly, and not to some specifically religious 
concept, about which much more disagreement legitimately exists.)

Now, I wonder if there is anyone on this list who thinks that overvotes 
should not be counted?

If not, I'd suggest that this indicates a consensus of the community on the 
topic. Because this is not an FA/DP organization of the community, this 
appearance could be false, because this list is likely not a representative 
sample. The theoretical ability of FA/DP organizations to quickly and 
efficiently discover consensus is one reason why I think they should be 
tried. The other reason is that it costs very little to try.

So, please, if anyone here either thinks that overvotes should be 
discarded, or can think of other reasons why it would be so argued, please 
let us know and, if possible, tell us why!

To summarize what has come before, two arguments were presented:

(1) An overvote could be the result of criminal alteration of the ballot. 
[True.] Therefore the vote should be discarded. [non sequitur, discarding 
the ballot accomplishes the goal of the criminal, at least in part.]

(More could be written on this. Allowing the overvote does allow the 
additional fraudulent vote to remain, but under the present system it would 
be unlikely that the overvote would be for the candidate that the criminal 
desired to win; more likely the overvote would be for someone else. To give 
an example, some Florida overvotes could have been generated by Bush 
supporters even though an unusual number of such votes were cast for 
Buchanan, not for Bush, thus throwing the election to Bush. [no charges are 
being made here that this actually happened, it is just a hypothetical 
example. In the actual case, there was a reasonable explanation, the 
butterfly ballot. Of course, reasonable explanations often conceal the 
truth....]

(2) Allowing overvotes allows the voter to fail to choose between two 
candidates [True], thus potentially creating a "failed" election. [False.]

[And again, more could be written about this. For example, not only may the 
voter abstain from making that choice anyway, the election has not failed 
if it produces a tie; rather it has demonstrated that the electorate does 
not prefer one candidate over another, so the election may fairly be 
resolved by a coin toss, as is actually done. Indeed, were there a way to 
set a clearly fair criterion, close elections in general might wisely be 
settled by coin tosses, especially when they are for representation in 
large bodies. This is done in Alcoholics Anonymous where elections for 
Conference delegate fail to produce a supermajority, at least according the 
publications. The argument for it is that the minority should at least 
occasionally be represented. Since I find this fascinating, I'll add that 
it would be more equitable to arrange the random choice method so that the 
odds of winning approximated the actual vote result(s), thus resulting, 
overall, in proportional representation. But one of the most famous mottos 
originating in AA is KISS, Keep it Simple, Stupid. AA attempts to reach 
consensus at the Conference level, so precision of vote strength is not 
terribly important and diversity of representation is more important. AA is 
a model Free Association (FA). And, typical of FA conditions, the 
Conference does not actually make binding decisions. Rather its votes are 
reported for guidance to the standard nonprofit corporation which holds 
actual AA property, which corporation is almost entirely supported through 
small donations from the members and which has no legal obligation to 
follow Conference decisions (but a strong tradition to respect a Conference 
supermajority); were it to stray seriously from Conference consensus, the 
members could either shift the board composition, or bankrupt it in less 
than a year and form a more congenial corporation.... This is how FAs could 
exert tremendous power where necessary, but they are much more likely to 
find paths to cooperation, coercion would be quite unusual.]