At first glance, it'd seem that the argument against the NPV would also be an argument in favor of the bill under consideration in the Pennsylvania legislature that would end the state's practice of awarding all its electoral votes to the winner of the statewide popular vote for President. As it turns out, Will also opposes this plan. His opposition is not partisan: he notes with disapproval that the bill has been pushed by Republicans who hope to diminish Democrats' electoral-vote total next year. In one way, on the other hand, it is partisan -- not in the sense that he seeks advantages for one party over another, but in the way that it exposes Will's bias in favor of Bipolarchy over alternatives.
While Will notes with implicit approval that "the existing system handicaps third parties," he worries that "Pennsylvania's plan would encourage third parties to cherry-pick particular districts, periodically producing 'winners' with only national pluralities of electoral votes, leaving the House [of Representatives] to pick Presidents." For one so reverent toward the intentions of the Framers, it seems odd that he should deplore this possibility. It is only what the Framers intended, and for all we know they may have expected presidential elections to go to the House more often than they have. It was also the Framers' intention to give each state discretion in how its electoral votes are distributed; as Will notes, two states already follow the practice Pennsylvania may adopt. How, then, does doing so violate the Framers' intentions?
The answer has something to do with Will's notion (if not the Framers') of the right sort of majority rule. The "certain kinds of majorities" he favors "do not materialize spontaneously.
They are built by a two-party system's candidates who are compelled to cater to entire states and to create coalitions of states. Today's electoral system provides incentives for parties to alter the attributes that make them uncompetitive in important states. It shapes the nation's regime and hence the national character. The electoral college today functions differently than the Founders envisioned -- they did not anticipate political parties -- but it does buttress the values encouraged by the federalism the Founders favored, which Pennsylvanians, and others, should respect.
Will's peroration begs a big question: if the Framers intended politicians to seek the not too small, not too big, but just right approval of statewide majorities, why on earth did they divide the states into electoral districts? Why is it not one state, one electoral vote instead? The answer probably has less to do with the intentions of the Framers of the Constitution than with the ambitions of the framers of the two-party system. It's more than implicit in Will's commentary that winner-take-all rules for states serve the interests of Bipolarchy best of all. The existence of Bipolarchy, and the desirability of its persistence, justifies the different function of the Electoral College that Will favors. Why Bipolarchy itself is desirable -- its main virtue, from Will's account, is that it compels candidates to compromise their principles -- is apparently a question for another time.