Subjects listened to digitized statements that were played backward or
in the usual forward manner. After listening to the statements, subje
cts were tested for recognition memory. All the statements were played
backward on the test. Subjects who had heard backward statements show
ed reliable old-new discrimination but subjects who had heard forward
statements did not. Then subjects rated the probable truth of statemen
ts that were typed in normal orientation. This time it was the subject
s who had heard the statements played forward who discriminated the ol
d from the new statements by rating more of the old ones true. The sub
jects who heard the statements backward did not. These results indicat
e that the forward meaning of backward statements does not ''leak'' th
rough, even when the backward statements themselves are memorable. Tho
se who are concerned about the possible influence of backward speech i
n rock music can rest more easily.