Let's say that _I fully know that p_ if I know that _p_, I know that I know that _p_, I know that I know that I know that _p_, and so on. Let's say that _I partially know that p_ if I know that _p_ but I don't fully know that _p_. What, if anything, do I fully know? What, if anything, do I partially know? One response in the literature is that I fully know everything that I know; (...) partial knowledge is impossible. This response is in tension with a plausible margin-for-error principle on knowledge. A different response in the literature is that I don't fully know anything; everything that I know, I partially know. Recently, Goldstein (2024, forthcoming) defended a third view, according to which I fully know some things and I partially know other things. While this seems plausible, Goldstein's account is based on denying the margin-for-error principle. In this paper, I show that the possibility of both full knowledge and partial knowledge is consistent with the margin-for-error principle. I also argue that the resulting picture of knowledge is well-motivated. (shrink)