A letter in today's Times from Professor Vicky Randall, Chair of the Political Studies Association, and others explains that AV doesn't violate the principle of one person, one vote. The idea that it does, they say, is 'based on a fallacy' (£):
Under AV, everyone's vote has the same value. Voters can rank the candidates in order of preference as far as they want. Counting begins by focusing only on first preferences. If one candidate garners more than half of these, that candidate wins. If no one gets half the vote, the bottom candidate is knocked out and their second preferences are added to the totals for the candidates who remain.
This might create the appearance that the bottom candidate's supporters are getting a second vote. But that is wrong. In the second round of counting, every voter again has one vote. If their favourite candidate is still in the race, it is their first preference that counts. If their top choice has been eliminated, their second or lower preference counts. No one has more than one vote.
This is pretty much what I said on Friday in my post about AV - speaking of two 'rounds' of voting (in a three-candidate contest) and of the voters getting two 'turns'. Shuggy construes me as claiming that AV is like 'the ballot system as used in French Presidential elections', in which there are two distinct votes by the electorate, taking place at different times, and thus allowing voters to change their preference ordering in the later ballot in the light of the outcome of the earlier one.
That's a simple misunderstanding - also expressed in an email to me (from OL). I spoke of two rounds or turns only in the sense that in ranking three candidates, each voter in effect casts her vote in more than one contest; but it's also obvious from what I said ('Every voter got to express their full preference-ordering...') that the voters do this all at once. Thus if I rank A, B and C in that order, I am voting simultaneously that A gets my vote should any candidate come out with 50 per cent or more, but that if this doesn't happen and A is eliminated in the first round, then B gets my vote over C. I am also, of course, voting - again, at the very same time - that if C should get eliminated then A gets my vote over B in this particular two-candidate contest, and that if B should get eliminated then A gets my vote over C in that particular two-candidate contest. So, if you want, I'm voting in four simultanteous rounds or getting four turns, though in my original post I left it at two since it seemed sufficient for purposes of clarity to do that. In any event, nothing I wrote implied assimilation of AV to two-round voting where the rounds are chronologically successive and separate from one another.