Referential expressions (REs) have been investigated in L2 English but to date there is no single study that systematically and simultaneously analyzes the development and acquisition of the multiple factors that constrain the choice of REs in natural discourse production. We investigate L1 Spanish–L2 English learners across three proficiency levels versus an English control group from the COREFL corpus. An analysis of both the RE and its antecedent(s) reveals that different intra- and extralinguistic factors constrain the choice of REs (information status, activated antecedents, syntactic configurations, characterhood, within-task effect, and proficiency level). L2 learners (L2ers) are sensitive to some factors but are unable to fully attain native-like levels even at advanced stages. They do not transfer null subjects from their L1 contrary to previous L2 research, and do not find all contexts at the syntax-discourse interface equally problematic, thus confirming previous theoretical proposals and empirical findings.

The study of crosslinguistic influences (CLI) has proven that morphosyntactic features exhibit CLI. Technical development and novel resources have enabled detection-based approaches, where potential CLI are revealed based on their observed frequencies and on differences between learners with different language backgrounds. The two research questions are as follows: (i) How construction-specific typological (dis)similarities between L1 and L2 affect the frequencies of linguistic features? (ii) Can such (dis)similarities be detected by comparing feature frequency data of L2? The data come from the International Corpus of Learner Finnish, and the methodology applied is the key structure analysis. The results support the applicability of the method: they show that constructional similarities may trigger CLI construction by construction, irrespective of the general similarities or genealogical categorizations. The results further imply the importance of controlling the genre-related and topical variation to account for skewed nature of the data when dealing with naturally occurring learner language data.