Rosenkranz (2021) offered a logic and a detailed account of justification, according to which justification that p can be analyzed as a form of second-level ignorance: ¬K¬Kp. An intuition behind the analysis is that the justified subject has the potential, at least in a nearby world, to either come to know p or come to know ¬Kp. However, given Rosenkranz’s hyperintensional semantics for modeling knowledge states, we can always construct, out of an ¬K¬K-agent’s knowledge state, epistemic possibilities that prohibit the (...) agent from having that potential. This shows that Rosenkranz’s analysis is still insufficient for carrying all the epistemic weight underlying the notion of justification. A better analysis would require a deeper look into an agent’s knowledge structure and its involved modalities, which could be a lot more complex than what the ¬K¬K rule tells us. (shrink)