The rationalists mostly refer to "structural unemployment" and "cyclical unemployment" where cyclical unemployment is what comes and goes with recessions and seasons while structural unemployment means "that job is gone, bro." From a "hysteresis" standpoint, structural unemployment is those jobs that take a long damn time to come back in any form because they involve retraining at a bare minimum.
I have to think there is a political element of this hysteresis as well. These monetary policies are being practiced in representative democracies, and the bubbles and busts are a big part of political sales in the election cycles that follow. It might be that Keynesian exercises reward politicians that practice disharmonious fiscal policy. When your bank account charges negative interest rates and cash becomes suspect, are people going to elect the guy calling for more of that? If the Obama administration was working with a Democratic legislature these last seven years it's likely that our monetary policies and fiscal policies might have been more harmonious, and Summers would never had felt the need to write this. Just like Communism, 'new Keynesian macroeconomics' might look best on paper.