Abstract
One reliabilist option against the problem of bootstrapping is to argue that circular reasoning is bad, but reliabilism can avoid circular reasoning. Vogel dismisses this option on the grounds that reliabilists need circular reasoning in order to circumvent skepticism. Briesen argues, however, that although reliabilists need circular reasoning to block second-order skepticism, they do not need it to block first-order skepticism. But I argue in this paper that reliabilists cannot legitimately reject first-order skepticism unless they can block second-order skepticism. In particular, I argue that reliabilists cannot meet the non-undermining provision for justification unless they can somehow block second-order skepticism.