No abstract

Articles pose a particular challenge to second-language learners whose first language does not have them. Variability in article production in these learners is often explained in terms of first-language influence, but there are also suggestions that frequencybiased regularities in the target language itself might play a role. While most secondlanguage research on articles has focused on English, a language with a relatively simple article system, the present study explores first-language influence and inputfrequency effects by focusing on Swedish. Swedish expresses definiteness using a complex noun-phrase structure including several free-standing and bound morphemes, some relatively frequent in input, others less frequent. An oral-production task elicited adjectivally modified and non-modified noun phrases in indefinite and definite contexts from 23 foreign-language learners of Swedish who were native speakers of Russian, an inflectional language without articles. The analysis revealed that the learners were more likely to supply high-frequency morphemes than low-frequency ones. Furthermore, while the learners were equally likely to supply bound and free-standing morphemes, only their suppliance of free-standing morphemes was negatively affected by adjectival modification; their suppliance of bound morphemes was not. While the role of cross-linguistic influence should not be neglected, these findings suggest that probabilistic regularities in the linguistic input are a key factor in second-language acquisition of functional morphology.

The Common European Framework of Reference has stimulated research on the relation between linguistic competence and communicative competence. Such studies have often used linguistic measures from the Complexity-Accuracy-Fluency (CAF) paradigm to objectively tap into the linguistic competence of language learners; others have used Processability Theory. This study contributes to this research by measuring grammatical accuracy and complexity, in terms of both CAF and level of processability, in learners of second language Swedish taking a speaking proficiency test. The results show that a higher level of processability primarily enables learners to use a more complex, rather than a more accurate, language. Nevertheless, assessors focus primarily on accuracy. This raises questions about the reliability of communicative language testing, and the article argues for a more psycholinguistic and experimental approach to language testing.