White (2021) has observed that some clause-embedding predicates (esp. doxastic attitude verbs like believe, non-veridical preferential predicates like hope) vary w.r.t. their selection properties: While these predicates commonly combine with declarative complements, they sometimes accept interrogative complements. My paper notes a similar selectional variability for fiction verbs like imagine: while imagine is typically taken to reject polar interrogative complements, some uses of imagine whether are acceptable. Curiously, this acceptability cannot be explained through the techniques (e.g., highlighting, no presupposition, multiple senses) that have recently been used to explain the acceptability of believe and hope whether. To still account for the ability of imagine to take whether-complements, I draw on recent work on attitudinal parasitism (see Blumberg 2019). This work assumes that some cases of imagination depend, for their reference, on the objects of another experience (e.g. visual perception). My semantics holds that imagine whether is felicitous only when the truth of the embedded TP is decided at the possible world of which the experienced scene is a spatio-temporal part. This condition is more easily satisfied when the verb in the TP has future tense (will), or when imagine is embedded under a negated ability modal or under try.