I’ll cheerfully admit that until last weekend I had no idea that Mississippi decides tied elections by drawing straws — much less that other states flip a coin. The only comparable version of planned randomness I’d heard of was the ancient Greek practice of choosing annual leaders by lottery.

My first instinct on hearing about the Mississippi State House election that was resolved Nov. 20 in favor of the Democratic candidate was that this arbitrary practice should obviously be changed.

On reflection, I’m not so sure.

A strong case can be made that elections should reflect the intentions of the voters, and that deciding them by reference to luck makes a mockery of the idea that the people are choosing.

Yet simultaneously, elections in the real world turn on a range of random factors, such as weather, traffic and the uncertainty of individual voters’ whims and willingness to turn out.

And the truth is that elections don’t literally reflect the will of “the people,” because not all of them vote, especially in relatively minor races like those for State House. (Quick, name your state representative. See what I mean?)

The idea that the people have spoken is a useful myth, captured in the medieval adage, “vox populi, vox Dei,” the voice of the people is the voice of God. It isn’t, not unless you believe that the hand of God is directly present in elections. It’s a fiction that serves our interests. And drawing straws or flipping coins once in a great while may actually be a great way to remind us of this fact.

The modern, rationalist impulse to reject the use of randomness to resolve ties rests on the idea that the purpose of democratic elections is to enable each individual to contribute to the community’s collective decision-making.

This same ideal of valuing everyone’s input equally explains the modern notion of “one person, one vote.” That idea is pretty recent, having been adopted by the U.S. Supreme Court in the 1964 case of Reynolds v. Sims.

Before that, most theories of representation held that a subset of the citizens was adequate to represent the rest of the public. Expanding suffrage to everyone (except children) is consistent with the objective of determining the will of the people via maximal participation of actual individual humans.

This ideal sits badly with resolving elections by a straw or a coin, for the simple reason that a straw or a coin isn’t a human who bears the agency of voting and choosing. That’s why, before the Mississippi straw-draw, one of the candidates said the practice was “archaic” and “medieval” and should be changed. The archaism lies in the denial of individual agency.

Yet the modern idea that we all participate in determining the results of elections is no less a myth than the old idea that God could speak through the voters. The reason is that we actually don’t all participate in elections, at least not in free countries. If voting were mandatory, as it sometimes is in dictatorships, then we could say with more confidence that the results reflected the public’s views.

Because we don’t force people to vote, we get a statistical sampling of the voting-eligible public — not the voters taken as a collective whole. We like to tell ourselves that the sample is representative, but we know perfectly well that it isn’t. Poor people in particular have low voter turnout, giving reasons that include both a metaphorical sense of disenfranchisement as well as the actual practical difficulties of finding time to go to the polls on an ordinary, busy workday.

And beyond observable, unpredictable turnout differences, further elements of randomness are introduced by daily life.

The other day I missed a local city council election in Cambridge, Massachusetts, where I live, through a combination of a crazy workday and dropping off children at school. I say I missed it, but what I really mean is that I gave priorities to other responsibilities over the civic responsibility to vote, even when I could’ve done otherwise, as I usually try to do.

Maybe a statistician or a psychologist could figure out a pattern in when I vote and when I fail to, but that pattern, too, would include some component of randomness.

Multiply that randomness across the whole voting-eligible population, and you’ll immediately see that electoral results contain some randomness.

That randomness is hard to reconcile with the rationalist theory of elections. That’s why we’ve largely forgotten about the confusing butterfly ballot that cost Al Gore the 2000 election in Florida and the presidency.

We remember the dimpled chad and the struggle to figure out the voters’ intent, a rationalist enterprise. But we mostly repress the embarrassing fact that some significant number of sweet little old Jewish ladies in Palm Beach County voted for Pat Buchanan when they thought they were voting for Gore and that nice boy Joe Lieberman. The randomness is almost too much for us to bear.

Yet it’s worth being reminded of how random elections — and life itself — can sometimes be. Tied elections are pretty unusual, maybe rare enough that we can tolerate resolving them by luck because they’ll teach us to be humble about our institutions and their capacities. The will of the people is a useful myth, but it’s just a myth — and we’d do well to remember it every so often.

Noah Feldman, a Bloomberg View columnist, is a professor of constitutional and international law at Harvard.