At 02:41 AM 3/6/2007, Jobst Heitzig wrote: >Dear Abd ul-Rahman, > >you wrote: > > If we > > are going to link the randomization that allegedly eliminates > > injustice to "minority" voters, we must have open voting, we can't > > have secret ballot. If we have secret ballot, and there is some > > hidden process that randomizes the results, well, tell me, would > > *you* trust that such a process was not being manipulated? After all, > > there would be no way to check. > >It seems to me you are repeating old arguments. First of all, voting >systems that remember past votings need not be open at all, they can >use cryptographic methods to keep each voter's voting history secret. >Second, randomization in practice is always done by pseudorandom >devices whose seed values can be determined deterministically from the >ballots, so that the result can be repeated and checked.
I feel like I'm seeing an insistence on an idea that results in a failure to understand what is being said. This is the second time in a row on this topic. So, sure, I may be repeating old arguments, *because they have not been answered.* Rather, the response, certainly, fails to indicate to me that the issue has been seen and, perhaps, actually countered. Sure, you can use cryptographic methods to keep each voter's voting history secret. In theory. Know of any actual examples? Unstated here is how the voter's record can be both secret and available for analysis. The vote must be *open* to some process. Now, it may be indeed possible to keep the identity of the voter secret, but also maintain a history for individual anonymous voters. (And, I'd suggest, it might not even be possible. A voter moves from one jurisdiction to another, so the vote takes place in a different precinct -- and it is impossible to separate voting locations from voting, since many *elections* are locally based. So, as soon as it can be determined that voter X moved from location 1 to location 2, in a given period of time, voter registration records can be used to determine, in very many cases, the identity of voter X.) So perhaps all this secrecy can be maintained. But how is it then checkable? That is, you may have a system which is theoretically sound, but with the level of secrecy required, it becomes extremely difficult to verify that it is actually working as desired. With open voting, it would be trivial. I understand about pseudorandom numbers. That's not the problem at all. You may design a system that you consider ideal, and it is even possible that a collection of "experts" can consider it secure, but the public knows that experts have been wrong in the past and it will want something far simpler. If you can design a simple system, great. But then you still face the problem that you are essentially reducing the intelligence of the result. No social value has been shown for giving "the minority" the right of decision by some statistical means. Range Voting *does* provide a kind of right like this where it counts. I.e., where it is important to a minority and not important to the majority. And the very core of my objection is that "the minority" is not a fixed group, such that it is deprived by not getting its way. The thinking behind this proposal seems to be that every citizen deserves to "get their way," yet "getting their way" is not the goal of electoral choice systems, the goal is maximization of benefit, and benefit is maximized by making choices which actually are the best for society, and individual opinions are only a clue to what these choices are. The theory is that majority opinion, particularly if informed, is more likely to be right than wrong. And the converse of this is that in the presence of controversy, minority opinion is more likely to be wrong, so following the opinion of a minority merely because of the outcome of a random process is more likely to increase error. That's noise. It is not a *reason* to follow the minority, i.e., some reason to expect that the minority opinion is more likely to be correct. Generally it is not. Range Voting does allow a reason: strength of opinion. In particularly, more knowledgeable people -- in some cases -- will have strong opinions, stronger than the opinions of the ignorant. In other cases, it's true, the ignorant will have strong opinions and the knowledgeable weak ones, in fact. But in such situations the knowledgeable are generally aware of the problem and will, given their own assessment of their relative knowledge, amplify their preferences to compensate. But if I'm knowledgeable and I say, well, I'm not sure which of these is best, A or B, but the ignorant are quite sure that A is best, *there is no harm in allowing the ignorant "their way."* It is only when I see -- or believe I see -- that ignorance is causing the majority to hold the *wrong* opinion, that I may find it necessary to amplify my opinion, which also, of course, assumes that my knowledge has led me to have an opinion of some strength. Yes, I quite easily acknowledge that Range Voting can, under some circumstances, devolve to Approval style voting. I don't consider this a problem at all, the only reasonable assertion that I've seen of a problem involved with this is that this makes the alleged high cost of the ballot and counting process useless. That cost, though, is generally trivial compared to the importance of the decisions being made. It's a phony argument. Range Voting need not be complicated, even Range with resolution higher than is probably optimal -- in terms of cost-benefit -- isn't particularly difficult. My point is that there is no value in giving any individual "their way" unless they perceive that, for some reason, their needs are being neglected in the process. Range methods allow such people some means of indicating this, and it is interesting that many here consider this a problem! I have no problem with allowing minorities the right of decision *under some circumstances.* Essentially, these are circumstances where there is reason to expect that the minority preference might be the better one for society. And where minority opinion is strong and majority opinion is weak, this is precisely such a situation. But none of this addresses the *real* problem, which is access and communication. When large numbers of people related to the system as this big *thing* that does not care about them, that does not listen to them, that does not personally communicate with them, they become alienated and even, under some circumstances, hostile. The kind of system Jobst has proposed would actually make this worse. It would not increase the sense of individual voters that the system personally responds to them, because the *response* isn't personal at all. The voter is just a statistic. Missing in all this is that I don't want leaders to act according to my preferences. Yes, I don't want leaders to act according to my personal preferences. Where I have delegated power to others, I want them to act according to *their* best judgement of my *benefit*. Yes, where I have a clear and strong opinion, I expect them to *consider* that, but if I don't trust them to make the best decision, given not only my opinion but also their own knowledge and judgement, I would not delegate power to them. I want to be *heard*, not *obeyed.* The problem arises because our systems of representative democracy don't allow me to *choose* my representative. Instead some anonymous and often unverifiable process chooses them. It may choose for me a representative that I wouldn't trust with the office of dogcatcher. By far. There are alternatives. They are known, they have been known for hundreds of years. They are a common-law right where property is involved, generally. So why do we have something less with politics? History and inertia and cynicism and despair. We could have this tomorrow if we woke up. (Almost literally tomorrow....) ---- election-methods mailing list - see http://electorama.com/em for list info
