# A paradox for relational accounts of hallucination **Abstract:** Even though the sense-data account of perception is almost universally rejected, the ‘phenomenal principle’ from which it derives much of its motivation is still, I will claim, very much in play. This is evidenced by the number of relational accounts of hallucination that can be found in the contemporary literature. My aim in this paper is not to argue against relational accounts, but merely to point out a couple of interesting and somewhat paradoxical consequences which arise for these theories. First, proponents of relational accounts have difficulty distinguishing between illusions and hallucinations; certainly they cannot make use of our intuitive way of distinguishing between these experiences. Second, it is difficult for adherents of the relationalist picture to consistently claim that those states we typically classify as hallucinations involve misperception at all.