Previous research showed that identifiable lures prevented false recognition under the Deese‒Roediger‒McDermott (DRM) paradigm. This suggests that learning ad hoc lists as opposed to associated lists produces less instances of false recognition if the list theme is noticed while learning ad hoc lists. The present study investigated whether participants notice the list theme when learning ad hoc exemplar lists. This experiment implemented the DRM paradigm to compare the frequency of false recognition evoked by learning ad hoc lists with that of the false recognition evoked by learning associated lists. A word designated as the seed of association was used as the critical lure for each associated list in keeping with the standardized procedure of the DRM paradigm, whereas a noun included in the list theme was used as the critical lure for each ad hoc list. A sample of university students (N=24) were instructed to study both kinds of lists and engage in a recognition test. The result showed that learning the associated list produced a higher frequency of instances of false recognition, whereas learning ad hoc lists produced a lower frequency with fewer instances of false recognition. This result suggests that the list theme is identifiable when learning ad hoc lists.