Noun-adjective collocations in L2 Portuguese: Exploring the effects of proficiency and native language in a large-scale learner corpus

In this study, we investigated the development of phraseological knowledge in Portuguese second language learners, focusing on noun–adjective combinations through the lenses of phrasal frequency, association strength, and accuracy. We analysed essays from 693 learners across five proficiency levels, representing 18 typologically diverse L1s. Learners’ use of noun–adjective combinations was examined using four association measures—phrase frequency, mutual information, log dice, and ΔP—as well as the accuracy of use. Findings reveal that phraseological development is not linear: advanced learners produced less frequent and less strongly-associated combinations, suggesting a shift toward more varied or creative expressions. L1 word order typology did not influence phrase frequency or association strength, but learners from noun–adjective dominant L1s produced collocations more accurately, indicating a dissociation between accuracy and idiomaticity. Accuracy followed a U-shaped developmental trajectory. Overall, the results highlight the complex interplay of learner- and input-related variables in L2 phraseological development. Keywords: collocations, phraseological development, association measures, L2 proficiency,learner corpus research, word-order typology