College students were aurally presented a set of 48 mutually unrelated sentences. After 60 sec, they received 1 of 6 recognition lists, each consisting of 48 sentences of which every one out of 6 was kept identical or transformed in 5 different ways; and they were required to judge whether each sentence was the same, formally as well as semantically, as any sentence that had been aurally given. The average proportions of "the same" responses were 73, 49, 38, 40, 26 and 19% for identical, converse, obverse, contrapositive, contradictory and irrelevant sentences, respectively. These results show that Ss do not retain the exact meaning of these sentences.