But wait, you say — hasn't Trump already locked up the majority of delegates he needs to win?
Indeed he has. All along, though, there's been a "nuclear option" that could deny him the nomination nevertheless.
To block Trump, a majority of delegates at the convention would simply have to vote to change their rules, to free themselves up to vote for whichever candidate they want. And then they'd just … nominate someone who's not Donald Trump.
This may seem shockingly undemocratic, but it's completely possible given how the convention works. The delegates get to set their own rules, and no higher body can overrule them. So they're perfectly free to change those rules at the last minute — even if those changes are obviously meant to alter the outcome of the race.
Now, few people currently believe anywhere even close to enough delegates would actually be willing to take such an extreme measure, because so blatantly overturning the results of the primaries would obviously provoke a massive uproar.
Furthermore, state law relating to delegate binding is complicated — some delegates may have obligations to vote for certain candidates under their state's laws even if that national-level binding rule is removed.
Still, if a die-hard anti-Trump movement were to gain sufficient strength among the delegates — and felt confident enough to defy the apparent will of the voters — nobody could step in and stop them from using the nuclear option. Even if pressing that button could destroy the Republican Party.
Many "Trump delegates" might not actually support Trump
One of the reasons a nuclear option scenario isn't completely implausible is that many supposed "Donald Trump delegates" headed to the convention might not truly support Trump at all.
By the current rules of the party, about 95 percent of delegates will go to the convention "bound" to cast their ballots for a particular presidential candidate. The primary and caucus results in their states determine which candidate they're bound to.
But these delegates are also actual people. And while in a few states the candidates get to pick which people will be their bound delegates, in many others they have little or no say in the process.
Indeed, it's common for states to select their delegates through obscure or arcane processes that have absolutely no connection to the primaries. One common method is through a series of caucuses or conventions, often at the county or state level.
Depending on the political situation and particular delegate selection rules in each state, these processes could end up being dominated by the states' political establishment, or by highly motivated and organized conservative activists. They could also be influenced by the candidates themselves — but by most accounts, Ted Cruz's campaign was much more organized and tuned in to this stage of the process than Trump's has been.
So many of Trump's delegates could actually be what GOP operative (and would-be uncommitted Virgin Islands delegate) John Yob has dubbed "SINOs," or "supporters in name only." Though bound to vote for Trump on the first ballot, these people could in fact hope to deny him the nomination.
And here's the crucial point: Delegates are only "bound" for the nomination vote. They have no obligation to vote in their candidate's interests on earlier rules votes.
So, theoretically, if Trump has a majority of delegates but many of them are SINOs, the overtly anti-Trump delegates and the SINOs could join together in an effort to change the rules in a manner that would hurt Trump.
Meaning they could vote to use the nuclear option — to change the rules so that all delegates would be completely unbound.
Ted Kennedy tried (and failed) to get the Democratic nomination with this strategy in 1980
This may sound like a far-fetched scenario — but Ted Kennedy made a serious attempt to get Democratic delegates to do just this in 1980, as Elaine Kamarck describes in her book Primary Politics.
Heading into the convention, incumbent President Jimmy Carter had clearly won the Democratic delegate race — he had a lead of several hundred delegates. But he trailed Ronald Reagan badly in polls, and Kennedy won all the late primaries. Furthermore, the month before the convention, news of the "Billygate" scandal, in which it was revealed that Carter's brother Billy had been paid hefty fees to lobby for the Libyan government, broke.
Kennedy sensed that Carter was vulnerable, and that Democrats had serious doubts that he could win in the fall. So Kennedy and his allies launched an effort to convince the delegates to change the rules and unbind themselves, so they could vote with their consciences instead. As Kamarck writes:
For the next month, the Carter campaign called and recalled every one of its pledged delegates to explain the rule, the controversy, and the motives of the Kennedy campaign. The Kennedy campaign organized an equally impressive delegate tracking operation, hoping to convince delegates that they could abandon Carter and vote for him...
...As the preconvention campaign mounted in intensity, Kennedy forces dubbed Rule F3(C) "the robot rule" — splashing the slogan "free the delegates" across posters and buttons and printing pictures of a robot with a red slash across it.
Kennedy's effort failed. Sure, Democratic delegates were worried about Carter's prospects. But few were willing to defy the apparent will of the voters in their states to try to depose him. When the "robot rule" came up for a vote, it survived easily.
Trump, of course, is not his party's sitting president, so he doesn't have that advantage going for him. But it's also worth remembering that in 1980, the idea that the public and not party insiders should pick the nominee was relatively new — the nomination process had been overhauled less than a decade ago.
The nuclear option would be a really, really tough sell today
Nowadays, the norm that the people and not the party choose presidential nominees is much better cemented. To the public, this "nuclear option" would look a whole lot like the nomination was being stolen from Trump, and with good reason. This modern democratic norm is the most formidable obstacle to this plan.
Accordingly, when I've broached the nuclear option topic in conversations with Republican insiders, most have expressed extreme skepticism that it would ever be used. And Republican National Committee spokesperson Sean Spicer has said that the party needs to "unite behind whomever the voters select as the nominee."
But the choice will truly be up to the delegates, not national party insiders. So to understand how likely the nuclear option actually is to be used against Trump, then, we have to understand two things about the delegates: first, their true candidate preferences, and second, how willing they'll be to flout democratic norms and cause an immense public uproar.
As in any gathering of 2,472 people, there will likely be a spectrum of political opinion — from out-and-out Trump supporters to die-hard #NeverTrumpers and everything in between. So how many of Trump's delegates will truly support him, and how many will be SINOs? And how many delegates overall will be truly hell-bent on stopping Trump, versus willing to tolerate his nomination?
Just as crucial is the delegates' tolerance for risk and willingness to court controversy. The delegates know full well that these decisions aren't being made in a vacuum, and those with political ambitions could face political consequences back home. They also know that changing the rules to depose Trump would likely make for a chaotic and potentially disastrous mess of a convention, and a new schism in their party.
Delegates who deeply believe Trump is a dangerous threat to their party or to America could well be willing to risk this backlash to take him down — indeed, there's an argument that since they have the power to stop Trump, they have the responsibility to.
But the more blatantly the delegates use their power to tip the scales against Trump, the more intense and volatile the reaction will be. And the nuclear option doesn't just tip the scales — it, well, nukes them. It would be a very, very tough sell.
This article was initially published in March 2016. It has been updated to reflect recent events.