Although I’ve developed something of a reputation as a Trump booster (I can’t imagine why), I’ve always been aware of the potential difficulties actually winning the nomination that Ross Douthat identifies. Because of his extraordinarily low level of support from party regulars, if Trump doesn’t actually get an outright majority of the delegates, enabling him to win on the first ballot, it’s difficult to see him winning additional support on a second, third or later ballot.

But just as Douthat would like to see regular observers grapple more seriously with that difficulty when they assess Trump’s chances, I’d like to see Douthat and the rest of the brokered convention crowd grapple with the difficulties attendant on the effort to deny Trump the nomination if he wins a substantial plurality of delegates.

The issue isn’t that winning a plurality means you “deserve” the nomination in some abstract sense. Legitimacy is always and everywhere a matter of perception. If those subject to an authority perceive it as illegitimate, then that authority’s ability to function suffers, and it really doesn’t matter whether that authority is duly constituted according to one or another theory. Standing on the ground of principle in such circumstances will be more likely to discredit the authority’s theory than to bolster the legitimacy of the authority.

Trump is running explicitly against the Republican party as it has historically understood itself. It’s a rebellion. If 40+% of the Republican party primary electorate supports Trump’s rebellion, and nobody else comes close to that level of support, and the delegates at the convention simply ignore that expressed preference, and pick somebody more to their liking – the sort of person Trump is running explicitly against – their actions will not be perceived as legitimate. Nor should they be, though that doesn’t really matter. What matters is that if, after all that has transpired, the party leadership blithely stands on the ground that they have the right to choose the nominee, and so nobody can legitimately complain, they will face the same rude awakening as Richard II – if not Louis XVI.

Were Trump a normal factional candidate who had plurality support but majority opposition, what you’d expect the party to do at the convention is make a deal with his faction in exchange for unity behind a compromise candidate. The problem for the party is that it’s not clear what kind of deal there is to make with Trump, not because he’s too extreme or too out of sync with the party’s historic positions, but because he’s Trump. Advocates of a brokered convention should acknowledge the problem this poses, and suggest possible deals and/or explain how the party should navigate in the rough seas that their advice would sail them into. Because as Douthat notes, the party leadership is ultimately going to act in their self-interest, and they are going to be really reluctant to see splitting the party as in their interest. Which is a major reason they’ve been as passive in the face of Trump’s revolt as Douthat laments they have been.

Then there’s the problem that Ted Cruz, the only candidate who might come close to Trump in the delegate count, isn’t a good party man either. Indeed, Cruz’s interests are distinctly at variance with those of party regulars, and this needs to be factored into any analysis of how a convention might be brokered. Cruz and Trump are allied in opposing the established party leadership, allied in trying to keep any other candidate from consideration on either the first or subsequent ballots, and allied in having little future if they don’t get nominated this time. Trump won’t have the kind of loyalty from his delegates that a typical nominee has – but Cruz very well may. If party leaders don’t agree to make Cruz the nominee, he might well throw his support to Trump in exchange for the Vice Presidency.

Of course, the uncommitted delegates at Cleveland will know about that before the first ballot. How might they react to the prospect of blackmail? Is it not possible that Trump might use that possibility to win over some among these delegates who have a deep antipathy to Cruz? And what about Kasich himself? Bear in mind that while Trump is still only winning pluralities in a three-person race, it’s entirely possible that, if Kasich dropped out, Trump would be winning majorities.

In other words, if the convention tries to give the nomination to someone other than Cruz or Trump, they may discover that Cruz would prefer to ally with Trump against the party regulars, and Trump wins the nomination. And if the convention tries to give the nomination to Cruz, they may discover that enough party regulars prefer to appease Trump than to lose with Cruz to put Trump over the top – provided Trump comes close enough to a majority that he doesn’t need to convince that many of them.

All of which means: the primaries really do matter.

Let’s say Cruz wins an overwhelming victory in the Wisconsin primary today, and this propels him to unexpected victories in Pennsylvania and Maryland, and a better-than-expected second-place finish in New York. Then he racks up lopsided wins in Indiana, Nebraska, Oregon, Washington, Montana, New Mexico, South Dakota, and – crucially – California. Even in this scenario, Trump ends up ahead, limping into the convention with a plurality of around 1,050 delegates, with Cruz nipping at his heels at around 900. This would basically make Cruz the equivalent of Reagan in 1976 – but Trump lacks the regular party support that Ford had. So Rubio and Bush instruct their delegates to vote for Cruz on the first ballot, and the lion’s share of uncommitted delegates follow suit. Kasich, who wins only his home state, comes under ferocious party pressure to do likewise, and he does so, giving Cruz a solid first-ballot majority.

Could that scenario transpire? Yes. Could the party avoid open civil war in that scenario? Probably – particularly if Cruz did well enough in Pennsylvania and California that his popular vote totals began to rival or even surpass Trump’s.

But let’s say Cruz’s victory in Wisconsin has no material impact on voting in the Mid-Atlantic later in the month – that, once again, it’s just demographics, not momentum. Trump wins an outright popular majority in New York, and either strong pluralities or outright majorities in Connecticut, Delaware, Maryland, Pennsylvania and Rhode Island. John Kasich finally drops out, which enables Trump to win Indiana in a head-to-head contest with Cruz. Trump loses Nebraska, Oregon and Washington, as well as Montana, New Mexico and South Dakota, but wins a clear majority in California, and 2/3 of the state’s delegates, as well as New Jersey, which is winner-take-all. He comes into Cleveland just 50 pledged delegates shy of a majority, with Cruz far behind, below 800. In that scenario, I think it’s fanciful to imagine that the party could unite behind anybody but Trump. In which case, the question facing the convention would be: should we try to unite behind Trump, or accept that the party is going to split for 2016, and start planning for how to heal before 2020.

If that’s the question, is Douthat sure the party will choose schism over apostasy?