The paper compares the effectiveness of balanced and skewed input at facilitating the acquisition of a novel construction by learners of English as a foreign language. The study replicates and extends Casenhiser and Godberg’s (2005) seminal paper on L1 construction acquisition. The study argues that the previous research did not pay sufficient attention to the frequency-prototypicality connection, as observed in natural language usage, and makes an attempt to overcome this limitation by adopting a more principled approach to this issue, which takes into consideration the frequency effects. Polish university students (N = 80) were randomly assigned either to skewed frequency groups, where the input was skewed towards the prototypical exemplars of the novel appearance construction, or a balanced frequency group, where the input was evenly distributed among the exemplars of this construction. Both groups were tested on comprehension of the target structure. The results suggest a facilitative effect of balanced input, as compared to the skewed input. This result was found to be correlated with the degree of cognitive/experiential entrenchment of the situations coded by the test exemplars, that is, the balanced frequency group significantly outperformed the control group only on tasks depicting highly familiar and frequently encountered scenes.
English