Towards a distribution-based corpus analysis of transfer-susceptible NP modifiers. A case of advanced users of L2 English.

The present study investigates transfer phenomena in Polish users of English at varying competence levels. We inspect the frequency of morphemes which are postulated to be transfer-susceptible (Hulk and Muller, 2000; c.f. Bu 2012). For our analysis, we chose four functional morphemes: the possessive ['s], the genitive [of], and two English articles, the indefinite [a] and the definite [the]. Our data consisted of a corpus of student writing (N=150 B2 Polish learners of English, N=224 C1 non-Polish learners of English, and N=192 Polish C1 students, mean age=20, SD=4), supplemented by an NS reference group (N=32). Our analysis afforded significant variation among conditions, F(6,2292) = 109.01, p < 0.05. Post-hoc analysis showed that the mean score of the B2, L1-Polish group (M = 5.58, SD = 4.11) was significantly lower than that of the C1-Polish group (M = 12.55, SD = 8.79). However, no significant difference was observed between the C1 groups, with the C1-L1 non-Polish group (M = 12.24, SD = 8.95) having a similar score to its Polish counterpart. The frequency of all studied morphemes can be correlated with learner level, while any effect of L1-background remains negligible. This finding marginalizes the influence of transfer on the distribution of grammatical morphemes in monitored L2 outputs and points to linguistic sources of morpho-syntactic and lexical errors, perhaps along the lines of Collins et al. (2009).