Replies

There are several salient points to be summarized.

First, despite attempts to stop the recounts, Gore got his selective recounts in his four counties. He was given the 7 days to accomplish the recounts, and when the 7 days expired, he went to court to get the updated returns from those 4 counties accepted after the deadline -- and he won that case as well.

But a funny thing happened -- he didn't get enough votes from the selective recounts in the counties that finished first, and in the 4th county, after some good initial results, they got into cuban sections of the county and started getting more Bush votes than Gore votes. At that point the counters in that county went behind closed doors, and in the ensuing outcry claimed "they had run out of time" and "had to stop", thus ensuring they wouldn't whittle away at the extra Gore votes.

Gore got that partial recount accepted as well, but still did not have enough votes. Meanwhile, the Supreme Court ruled that selective recounts were unacceptable, and kicked the case back to the Florida court.

There was one county left, the ballots were all in a court under a judge's control, and the Florida Supreme court took the opportunity of the Supreme Court decision to issue a NEW ruling, telling the judge in that county to recount the votes, and ordering a recount of all undercounts in the entire state, well after the results of the election had been certified.

That was the infamouse "state-wide recount", being done by retired judges across the state, that we watched for a day on Saturday. The Florida court refused to set a standard for what would be considered a vote, leaving it to each county's election board to make that decision.

It was this NEW recount that went back to the Supreme Court. Interestingly, while the Florida Supreme Court ordered that the recounts be kept secret until completed, the news reports suggested that Bush might actually be gaining votes in that recount.

Anyway, the Supreme Court, seeing this travesty of selective recounting of undervotes using different rules county by county and sometimes "group by group" (counties set up groups of 3 judges to get the job done, and each group could make up their own rules), decided by a 7-2 vote that if you were going to recount, you had to recount the entire state under the same rules, or else you were violating the "equal protection" clause.

That 7-2 vote STOPPED THE RECOUNT that was going on.

The court then voted 5-4 that there was no legal way for the Florida Supreme Court to correct their ruling in time to count the votes in a manner that would pass constitutional muster before the statutory deadlines, and since missing the deadline would disenfranchise ALL the voters of florida, no further counting would be allowed.

BTW, the recount was technically flawed, as we found out from evidence in the one county that was actually in court. Having counted the state twice, they knew the current vote count, and they knew how many ballots did NOT have a vote for President, and how many had more than one vote.

When they decided to recount, they reprogrammed the punch card machines to separate undervotes from the others, and to sort them by precinct. There was no count collected during this separation, as the software was not certified for counting. It was assumed we didn't care about certification for undervotes, since at worst we'd simply miss some of the undervotes, while the actual counting was done by humans.

But they discovered an odd thing (although the lawyers didn't seem to understand the significance). In one precinct, the total number of counted votes, PLUS the total number of "undervote ballots" sitting in the vault, added up to MORE THAN the total number of voters who had showed up to vote in that precinct.

In other words, at least ONE of the "undevote" cards awaiting manual counting had already been included in the actual count, and would therefore be counted twice.

But in fact, the flaw was more serious. If a chad is hanging, it can flip open and closed, in addition to falling off. So not only was it likely that a good number of previous votes might have had a chad flip back closed and now be seen as "undervotes", but a good number of "undervotes" had now turned into votes and were therefore NOT separated out to be manually counted. So it was quite possible that dozens of counted votes for Gore were now sitting in the undervote pile waiting to be counted again, while undervotes which would clearly have been seen as Bush votes were now in fact "countable" and not separated.

This flaw could have been corrected if they had certified the software and re-run the counts. However, they had already learned that each recount through the punch-card machine gave them different results, and in fact showed MORE votes -- and doing it again would have proven to the world that the punch card machines were in fact GENERATING votes.

Which means that you couldn't really use a machine recount and get a good answer either. What votes would they generate? Well, the tendency in punch cards if that, if you press them tightly, an unpunched chad can be pushed into an opening of the next card. So, if you took a pile of unvoted cards, and interspersed them with a pile of Gore votes, and ran the machines a dozen times, you would each time get a few more Gore votes as the chads were knocked off because they were caught in the holes of the next card.

So if you used the machines to sort cards in a county that was majority-Gore, you would get more Gore votes. Of course, if the chads were knocked off, those Gore votes would have simply been sorted to the "not undervoted" pile. So by NOT having the machines recount, we didn't have this problem, but had the other problems.

Also, this would tend to make Bush votes into OVERVOTES, as Gore votes were knocked into the Bush vote cards. The sorter would have kicked those out, but they werent counting overvotes, so if you had run vote totals off the machines a 3rd or 4th time in a highly democrat district, you'd tend to see increasing numbers for Gore, and more overvotes, and possibly fewer Bush votes.

If you are going to recount, which is stupid, it should have to be a full manual recount of very ballot.

But the simple fact is, elections are somewhat random, and more "representative" than absolute. On election day voters get sick, or their cars break down, or the lines are too long, or they get held up at work. There are storms in parts of the state, some people get called out of town, absentee ballots get lost, etc.

Even if you think your process itself is perfect, the actual vote counts only approximate what the ELIGIBLE voters would want, and with all the common mistakes and machine errors, the actual counts only approximate what the actual VOTES were for people who showed up.

So if an election of 6 million people is within 500 votes, the simple fact is that we have no idea WHICH candidate was actually supported by a majority of those 6 million people. But it also doesn't REALLY matter, because about half the people supported each one, so they are equally "representative". It's no real travesty if a candidate wins by one vote and it turns out that we find one voter who poked the wrong hole and otherwise the OTHER guy would have won -- because certainly we don't think that ONE voter was the one that deserved HIS choice of representative.

We count all the votes because it gives us a sense of absoluteness that isn't real but feels good. Al Gore shattered that sense of "rightness" by demanding recounts and by complaining about the voting process, revealing what we already knew but didn't care about -- that in close elections, we can only guess at who really "won", and up until 2000 that was generally OK with us.

In 2006, George Allen lost by 9000 votes. If 4500 more Allen supporters had been able to make it to the polls, (there were certainly 4500 more people who didn't vote who, if they did, would have voted Allen), and 4500 Anti-Allen voters had been held up at work, Allen would be senator.

But nobody can tell me that the votes of those 9000 people would change what it is that Virginia wants from our Senator. It was just the luck of where it was raining, and who was held up, that put Webb in instead of Allen. That's true for any close election (I believe in terms of percentages Allen's was the closest of the races).