Article use poses a challenge for second-language learners (Chrabaszcz & Jiang, 2014). Pre- vious research suggests that having acquired articles in an L2 facilitates the acquisition of articles in a third language (e.g., Ariba ̧s & Cele, 2021; Jaensch, 2009), and that article acquisition is to some extent determined by frequency-based patterns in the input (e.g., Ogawa, 2015). However, both suggestions suffer from methodological limitations. The present study addresses those limitations by investigating Finnish-speaking learners of English and Swedish, considering both cross-linguistic influence and frequency-based regularities in the input. Finnish students learn English and Swedish at school and Finnish learners (whose L1 lacks articles) might to some extent rely on their knowledge of English in Swedish article use (Jarvis, 2002; Nyqvist, 2016). For the present study, acceptability data (Likert 1–5) was collected from Finnish adults (n = 57) as well as Swedish (N=133) and English (N=30) control groups. The stimuli consisted of 32 base sentences, each appearing in four versions – with and without an article in English and in Swedish – resulting in 128 unique variations of each sentence. No participant rated the same base sentences in multiple variations. We also collected self-ratings of proficiency. Blog- mix (Spr ̊akbanken, 2017 (1.67M)) and enTenTen (Sketch Engine, 2022 (52G)), both building on colloquial language, were used to establish corpus frequency and article probability (i.e., the probability that the phrase occurs with an article).

For each item, the learners received a target-likeness score based on their deviance from the L1 groups’ mean score. The target-likeness score was used as an outcome variable in two mixed- effects models, one with Swedish and one with English as the target language (TL). Both models included the participants’ relative proficiency as well as item frequency, article probability score, and TL proficiency as fixed effects. Item and participant were included as random effects.

Both self-reported proficiency and participants’ ratings indicate that the Finnish participants were generally more proficient in English than in Swedish. In our preliminary results, TL profi- ciency and article probability show as predictors of target-likeness in both mixed-effects models. However, there are clear differences between the models. Absolute frequency is only a signifi- cant predictor in the English model, meaning that the more frequent the item was, the more native-like the participants’ ratings were. In line with the hypothesis about English affecting Swedish article use, we do see effects of comparative English-Swedish proficiency in the Swedish model.

Selected references

Ariba ̧s, D. & Cele, F. (2021). Acquisition of articles in L2 and L3 English: the influence of L2 proficiency on positive transfer from L2 to L3. Journal of multilingual and multicultural development 42(1).

Chrabaszcz, A. & Jiang, N. (2014). The role of the native language in the use of the En- glish nongeneric definite article by L2 learners: A cross-linguistic comparison. Second language Research 30.

Jarvis, S. (2002). Topic continuity in L2 English article use. Studies in SLA 24(3).Ogava, M. (2015). Input frequency of article-noun combinations as another factor for L2 article use. Ars linguistica: linguistic studies of Shizuoka 21.