The confusion between sentential figurativeness and conventionality found in many of the experiments on figurative language comprehension is here disentangled by factorially crossing the figurativeness of a proverb (determined by discourse context) with conventionality (determined by familiarity of use). Familiar proverbs are conventionally used in their figurative (and not literal) sense whereas for unfamiliar proverbs the literal meaning (and not the figurative sense) is more available. Multiple dependent measures were employed: the time taken to read the target (experiments 1, 2 and 3), incidental recognition tests of target (experiments 1 and 2), recognition errors (experiments 1 and 2), interpretation errors (experiment 2), and recall aided by context-appropriate or inappropriate cues (experiment 3). Reading time data indicated that unfamiliar proverbs used figuratively took longer to read than the same proverb used literally or literal paraphrase controls. Familiar proverbs were read equally fast, whether understood as a literal or figurative statement. The pattern of memory errors and cued-recall data indicate that conventional meaning and literal meaning are both available in context-appropriate and context-inappropriate conditions, whereas unconventional meaning is available only in context-appropriate conditions.