This is the third post in a series. The first one is here. The second, here.

It's easy to remember the 2016 presidential primaries, particularly the Republican National Convention--dramatic conventions with delegates pledged to support certain candidates, a lawsuit enjoining a state law from forcing delegates to vote for their pledged candidate in Virginia, procedural wrangling and voting on the floor, and the like. The Democratic National Convention was less climactic, but many battles over the roles of delegates and superdelegates lingered over the months ahead of the convention.

Many of these battles turned on a fundamental question about what delegates are supposed to do. That is, when voters in a state selected delegates, did they expect them to serve as something like brainless robots, simply reflecting their public pledges to support a candidate? Or did they expect them to think independently, despite their pledge, and perhaps modify their positions based on new information as they deliberated prior to the convention?

It might be the case that settled expectations suggest that delegates ought to be bound to their pledge, at least for the first round of voting. Or it might be the case that we elected human beings as delegates for a reason--to exercise their judgment, even if that means breaking their pledge.

This debate is the same kind of debate surrounding the Electoral College. We select 538 presidential electors, and they meet on December 19 this year in state capitals around the country to vote for the next president and vice president. They rarely exercise independent judgment anymore--indeed, they have rarely done so since 1796. But some still suggest they ought to do so, particularly this year. Why compel electors to vote for a candidate when new information might change their mind? Much can happen between November 8 and December 19. Allow electors the opportunity to decide for themselves.

This, of course, is exceedingly unlikely for a host of reasons. The electors are often selected as party, and even candidate, loyalists. The public suggestion that an elector might vote for someone other than the pledge candidate is enough to force resignation. Most states require electors to take a pledge; a few purport to make that pledge enforceable.

Perhaps that ought to be the case--we have settled expectations, I suppose, that electors should not exercise independent judgment. But, at the same time, we would expect those electors to exercise independent judgment, I think, if the candidate died, or was indicted on a felony, or perhaps exercised some other egregious lack of judgment--no? And if so, is it only limited to egregious circumstances?

It's worth noting that for all the flaws of the Electoral College, the 12th Amendment addressed a few things--reacting to presidential-vice presidential tickets, political parties, and the like. But it didn't remove the judgment reserved to electors. It could have done so--even the 20th Amendment could have done so. This residual authority has long remained with us.

And in most cases, such independent judgment is academic--a D.C. elector in 2000, for instance, refused to vote for Al Gore to protest the lack of congressional representation for the District of Columbia. But it didn't affect the outcome, and perhaps we simply ignore the issue.

But in close elections, there are greater "risks" associated with "faithless" electors--"risks" meaning an increased likelihood that the outcome of the election changes from our expectations on November 9, and "faithless" meaning electors voting for a candidate other than the candidate they are pledged to support. Consider a survey of electors after 2000, which found nearly 10% of electors open to voting for someone other than their pledged candidate. Consider that just two Bush electors voting for someone else would have thrown the election to the House, the "risks" can be quite high.

One can imagine similar pressures, perhaps even greater, in this year's election. Given the history unpopularity of both Donald Trump and Hillary Clinton--for altogether different reasons--a close election might yield unusual pressure on electors to either switch their votes for the other candidate, or to throw their votes to a third candidate.

In the event no one candidate secured 270 electoral votes, the top three vote-getters will be sent to the House of Representatives in what is known as a "contingent election." (The Senate votes among the top two vice presidential candidates.) The House votes not as 435 members, but as 50 states, each state receiving one vote, similar to the Congress under the Article of Confederation. Wyoming, then, has the same voting power as California. Republicans control most of the state delegations and would likely control the choice if they voted as a unit.

Is that a better system? One of electors exercising their own judgment and sending the presidential election to the House? Perhaps discussing the role of the House is a better left for another post. But it's worth noting that in a close election, a couple of stray electors could flip the election to the "losing" candidate ("losing," at least, based on our expectation on November 9).  Or, the judgment of electors might send elections to the House.

It might be that we are so far beyond the notion of electors exercising independent judgment that we simply could not (as a society) accept the result as "legitimate." But it's worth emphasizing that electors have this power, that they have exercised this power, and that there are some good theories of government that might view it as admirable that they have and exercise this power--look no further than the debates that played out in the 2016 presidential primaries. Until we replace electors with robots, we should expect nothing less.