Background: Background sounds disrupt short-term memory, even if the target information is presented visually (Ellermeier and Zimmer, 2014, doi:10.1250/ast.35.10). Since speech has been found to be the most disruptive sound, this phenomenon was termed the Irrelevant Speech Effect (ISE). Previous studies have focused on various factors that contribute to the ISE, e.g., the type of sounds, speech intelligibility, or familiarity with the presented language. The present study focuses on the tonality of tonal languages. Compared with non-tonal languages, tonal languages contain more pitch changes and thus more changing state that causes the ISE. We hypothesize that, at least for native Mandarin speakers, tonal languages will cause a bigger ISE.

Method: 38 native Mandarin and 38 native English speakers were recruited and asked to memorize the order of a nine-digit sequence that was presented visually while the participants listened to background sounds. These sounds were either English, Mandarin, or Cantonese speech, or a continuous pink noise.

Results: Both Mandarin and English speakers confirmed that the least disruptive sound was continuous pink noise, with an average of 8.36 and 6.65 digits correctly recalled, respectively. This difference that Mandarin speakers could recall about one more digit than English speakers was found in each sound condition.English speakers experienced a larger disruptive effect in English (5.01) than in Cantonese (5.69) and in Mandarin (5.47). This result differs from previous studies that did not use tonal language sounds and found no effect of native language. Mandarin speakers experienced the highest ISE in Mandarin (6.51), followed by English (6.80) and Cantonese (6.94).

Conclusions: The results suggest that native language increases the ISE in tonal languages. The small ISE for Cantonese speech presented to Mandarin speakers suggests that the underlying cause is rather the native language than additional changing state due to tonality. A possible reason for the Mandarin speakers memorizing more digits than English speakers could be that Mandarin digits can be pronounced and thus rehearsed more quickly than English digits. We are presently investigating this further in a follow-up study where we present digits and Latin letters as to-be-memorized material.

We investigated whether the rehearsal language would affect the short-term memory and ISE. We found that both background sounds and to-be-memorized material significantly affect the correct rate, but only the background sounds significantly affect the size of ISE. This indicated that rehearsal language will only affect participants’ short-term memory but will not affect ISE.