We investigate how a textbook as input is composed and how L2 beginners' written production as output is affected by the distributional properties of the textbook. We found supportive evidence for input-output relations in L2 development, as well as cross-linguistic influences on the course of acquiring target language knowledge.
Usage-based accounts assume frequency as a nucleus for language development. Korean particles (postpositional function words indicating grammatical information) provide an attractive testing ground for this issue due to their polysemous nature. Previous research generally focused on either textbooks or production by proficiency, lacking the consideration of possible input-output relations pertaining to learning outcomes. With that, this study investigates the degree to which textbook input affects L2 beginners' written production of three Korean particles (-i/ka, -(n)un, -(l)ul) and their prototypical functions (NOMinative, TOPic, ACCusative), respectively. Whilst -(l)ul is mapped strongly onto ACC, -i/ka and -(n)un are often interchangeable to indicate NOM and TOP colloquially.
We examined distributional properties involving the three particle types in a beginner-level Korean textbook, and then compared learner writing (cartoon-based narrative essays from 26 English-speaking undergraduates who learnt that textbook in the previous semester) to input characteristics with respect to the polysemous nature of these particles.
The textbook provided perfect form-function associations of individual particles, with different proportions of occurrence: -(n)un~TOP (26%) > -i/ka~NOM (22%) > -(l)ul~ACC (11%). L2 beginners' writing, in general, also demonstrated the same kind of form-function pairings involving these particles. However, the tendency of particle use in learner writing diverged such that -i/ka~NOM (31.5%) and -(l)ul~ACC (21.3%) were over-produced whereas -(n)un~TOP was under-produced (17.8%) in comparison to the textbook.
Learners' use of -i/ka~NOM and -(n)un~TOP is ascribable to the subject-prominence in English and the interchangeability of the two particles, possibly interfering -(n)un and instead choosing i/ka even for TOP. Their production of -(l)ul~ACC implies the early entrenchment of this particular mapping, possibly leading to its over-production compared to input distribution in the textbook. Together, our findings lend empirical support to input-output relations, favouring the usage-based accounts in L2 development, and possible impacts of cross-linguistic differences on the beginning stage of L2 acquisition.
