The Proper Functionist account of warrant – like many otherexternalist accounts – is vulnerable to certain Gettier-style counterexamples involving accidentally true beliefs. In this paper, I briefly survey the development of the account, noting the way it was altered in response to such counterexamples. I then argue that Alvin Plantinga's latest amendment to the account is flawed insofar as it rules out cases of true beliefs which do intuitively strike us as knowledge, and that a conjecture recently put forward by Thomas Crisp is also defective. I conclude by presenting my own suggestion as to how the account can be made less vulnerable to counterexamples involving accidentally true beliefs. Although I stay within the confines of Proper Functionism here, I think that my proposal (modulo a few details) could be attached to other externalist accounts of warrant as well.