I expect different departments have different attitudes towards published work. But in the departments I've worked at (Harvard, Princeton, NYU), there's been no requirement or expectation that the leading candidates will already be published. Having a paper that's good enough to be in J Phil will certainly be to your benefit; you should use it as your writing sample. The extra time it would take you to really get it into J Phil would (I think) often be more productively spent writing another paper that good, or making your dissertation project better developed, more polished, or more marketable. So I never push my grad advisees to publish.
There's a lot here I want to unpack, and I'm not going to get to it all today. Let me start by saying, these are exactly the reasons why, until recently, my own senior profs told me not to worry about publishing.
Notice the key idea here is that I'm supposed to write a paper good enough to get into a good journal, but then not bother with the painfully slow process of actually getting it past the journal's referees. Okay, fine. But good enough by whose standards? I'm pretty sure most search committees look for publications as evidence that a candidate's work is good. The whole fucking point of that part of your CV is to show the level you're working at.
It’s no accident Pryor's talking about his experience with search committees at Harvard,