The present study investigates the effects of the introduction of visual text in the shadowing of Japanese. The study involved one group of native speakers of Japanese (experiment 1) and a second group of advanced learners of Japanese whose native language was Chinese (experiment 2). Both groups were asked to perform two types of shadowing tasks: one with a text (parallel reading) and one without (shadowing). Performance was measured for the degree of comprehension, oral response time and oral accuracy. The main results were as follows: (1) In parallel reading, the comprehension level of both the native speakers and advanced learners was as high as for shadowing; (2) The Japanese learners required a longer response time in parallel reading than in shadowing; and (3) For the Japanese learners, parallel reading was performed with greater accuracy than shadowing.