Incidental acquisition of formulaic expressions by L1 and EFL learners: Input mode and frequency effects

Formulaic expressions (FEs) subsume all types of lexicalised word combinations including idioms, conversational routines, collocations and binomials. Recent studies (Lin, 2012; 2018) speculated that aural input is more conducive to FE acquisition than visual input. A controlled experiment was conducted to address whether: 1) the outcome of out-of-context incidental FE acquisition indeed differed by input mode; 2) EFL learners and L1 English speakers differed in how well they acquired FEs from aural and visual input; and 3) the number of repetitions affected FE acquisition.

In this experiment, 182 EFL and 30 L1 English speakers learned 36 novel binomials (e.g., silent and sleepy, sample and signal) presented aurally or visually out-of-context. A third of the binomials were repeated 4, 6 and 8 times. Incidental learning was measured using an unannounced form recall task, which required subjects to complete the phrases within 5 seconds either after seeing the first two words plus the first two letters of the third word (e.g., silent and sl___) or after hearing the first two words plus the first two phonemes (e.g. /ˈrɪs.ki ən juː/). Successful recalls were counted for items learned aurally and visually and compared within-subjects.

The results shows that aural input is more conducive to FE learning than visual input for both L1 (t=8.35, df=29, p=3.34e-09; Cohen’s d=1.52) and EFL learner (t=4.70, df=181, p=5.20e-06, Cohen’s d=0.35) groups, although the magnitude and impact of the aural input advantage were significantly larger for the L1 speaker group (t=4.93, df=50.09, p=9.58e-06, Cohen’s d=0.95). The number of repetitions had a statistically significant impact on the magnitude of the aural input advantage (χ 2 (2)=32.2, p=9.95e-8, Kendall W=0.08), but the relationship between the two variables was not directly proportional. The findings will be explained from prosodic, evolutionary and working-memory perspectives.