Procedure

Participants completed three tasks (distributional learning task, familiar song task, and language and music background questionnaire), the order of which was randomized across participants. The entire experiment took approximately 50 min to complete.

There were three phases to the distributional learning task: pretest, training, and posttest. At pretest and at posttest, participants completed an ABX discrimination task (e.g., in a trial in which participants hear /na33/–/na241/–/na33/, the correct answer is A, or the first sound) for test stimuli of both domains. There were 32 trials per test phase, with eight repetitions of the four target minimal pairs (2 Thai lexical tones + 2 musical chord minimal pairs). The test trials were not blocked by test domain but presented in a randomized order for each participant. Participants were required to respond within 1 s in order to maintain vigilance and there were no replacement trials for slow responses. Prior to pretest, participants were given four practice trials with feedback.

During training, participants were randomly assigned to be trained on either Thai lexical tones or novel musical chords and then further assigned to either a Unimodal or a Bimodal condition within each domain. Listeners in both distribution conditions heard the same number of training tokens (i.e., 256) but the conditions featured different modal peaks of the distribution; listeners in the Unimodal condition heard Tokens 4 and 5 most frequently whereas those in the Bimodal condition heard Tokens 2 and 7 most frequently (Figure ​Figure11). Listeners in both conditions were presented with Tokens 1 and 8 (i.e., the stimuli used in pretest and posttest for the minimal pairs for both domains) the same number of times. In order to ensure that participants listened to the entire set of training tokens, they were required to complete a cover task—to mark on a response sheet with the numbers 1–288 whenever they heard a ‘beep’ by circling the sound number to which the ‘beep’ corresponded (Ong et al., 2015). A total of 32 beeps were interspersed randomly within the training tokens. The training phase took approximately 6 min to complete.

Frequency of occurrence for each training token heard by the Unimodal condition and the Bimodal condition.

In each trial of the familiar song task, participants were shown a song title and the artist who performed the song and were asked to indicate whether they were familiar with the song. If they were unfamiliar with the song, they moved on to the next trial with no replacement trials. If they were familiar with the song, they were then presented with two excerpts: an original and a transposed version, the order of which was randomized, and they were required to choose which excerpt was the original. There were two blocks of 20 trials, with one block corresponding to a pitch transposition of one semitone and the other of two semitones. The presentation order between the blocks and the trials within each block was randomized. Across all participants, the average number of songs chosen as familiar, and hence, presented to the participants, was 27.96 (SD = 6.61, Range = 10–40).

In the language and musical background questionnaire, participants were asked to provide their demographic details as well as to list all the languages known to them and rate on a 5-point scale how well they (i) read, (ii) speak, (iii) write, and (iv) understand each of those languages. They were also asked to indicate whether they have had musical training and if so, the age of commencement, the duration of training in years, and the instruments played.