Because nobody knows why people vote the way they do. At least, not in any useful sense. The four main theories of voter behavior — micro-sociological, macro-sociological, socio-psychological, and economically-rational — are as narratively compelling as sociology, psychology, and economics at explaining why something happened, but (like those disciplines) are basically useless as prediction tools.

You can read some of this between the lines of a revealing, month-old, two–part interview with Mike Murphy on his plan to cinch the nomination for Jeb Bush through the mega-bucks of the Right to Rise SuperPAC. Amongst the various details that may or may not be misdirection — that the 45 days leading up to March 15 are key; that targeting the southern states for 10 days coming out of New Hampshire could cost $35 million in media buys; that Right to Rise’s war chest is funded by a few thousand donors; how they’re looking to link your mobile phone location data to your voting history; etc. — Murphy refers to his “theories about the Iowa caucus electorate, the New Hampshire electorate, and the South Carolina electorate.”

“Theories.” Because even in a piece that is part advertisement, part disinformation, and part application for his next job, Murphy knows he doesn’t know why people vote the way they do because none of us do. So any argument about “electability” is an exercise in rhetoric, not science. And should therefore be treated with an appropriate level of seriousness: none.