Three experiments examined whether the representations underlying recognition memory familiarity can be episodic in nature. Recognition without identification [Cleary, A. M., & Greene, R. L. (2000). Recognition without identification. "Journal of Experimental Psychology: Learning, Memory, and Cognition," 26, 1063-1069; Peynircioglu, Z. F. (1990). A feeling-of-recognition without identification. "Journal of Memory and Language," 29, 493-500] was used to isolate familiarity processes. In order to test whether the representations underlying familiarity can be episodic, participants were exposed to stimuli that should not have pre-experimental representations, pseudowords and nonwords. The first two studies found evidence of recognition without identification for pseudowords and nonwords, a result that is inconsistent with views proposing familiarity only arises from existing representations. The third study found evidence that recognition without identification for words, pseudowords, and nonwords is stronger when study and test modality match. These results are interpreted within the framework of global-matching views of recognition memory, which claim that familiarity arises from the matching of test items to episodic representations in memory. (Contains 2 figures and 7 tables.)