This dissertation examines syntactic complexity in learner Swedish. Syntactic complexity is explored through both absolute and relative complexity. Absolute complexity refers to linguistic, structural complexity, and relative complexity can be seen as complexity in relation to the learner. Thus, syntactic complexity is defined as a structural property of learner language that can be observed in the varied use of long and embedded units of language and in the use of demanding structures. Syntactic complexity is inherently multifaceted and influenced by many factors. The present study focuses on the effects of language proficiency, mode of production, and individual variation. The theoretical underpinnings of the study are drawn from cognitive and usage-based approaches to language learning, and the dissertation incorporates ideas from both dynamic and systemic views of learner language development. This study also provides suggestions for teaching and assessing language.

The dissertation consists of three original publications and this summary. The data comprise written and spoken productions by 31 Finnish-speaking university students. A variety of syntactic complexity measures targeting different linguistic levels (e.g. the clausal level) are used. The first two substudies explore differences in absolute and relative complexity at two proficiency levels (lower vs higher) using quantitative analyses. The first substudy examines the effect of learners’ language proficiency on the syntactic complexity and accuracy demonstrated in their language use, as well as on the relationship between these aspects of learner language in written production, whereas the second substudy investigates the differences in syntactic complexity between speech and writing. The third substudy is an individual-level enquiry that focuses on absolute complexity. In this study, complexity profiles in learners’ productions are explored using a mixed-methods approach.

The results of the dissertation corroborate earlier findings demonstrating that syntactic complexity is a complex feature of learner language. Somewhat unexpectedly, the results of the first substudy do not indicate any statistically significant correlations between measures of syntactic complexity and accuracy. In line with several earlier studies, the results show that the learners at the higher proficiency level produced more accurate structures. However, in contrast with some earlier research, the learners at the lower proficiency level were able to produce relatively complex structures in their writing. As the second substudy shows, the difference in syntactic complexity between the two proficiency levels was more noticeable in speech, indicating that the mode of production employed has a strong effect on the resulting syntactic complexity. At the lower proficiency level, the written productions seemed more syntactically complex than the spoken productions, whereas the syntactic complexity in the two modes of production did not differ to a great extent at the higher proficiency level. In line with many previous studies, production in the spoken mode seemed more cognitively demanding for the learners at the lower proficiency level. Based on the results of the first two substudies, individual style of producing language seems to be one of the key determinants of syntactic complexity. The results from the third substudy, conducted at the individual level, validate this assumption: while some learners have distinctively individual styles of producing structures, others show more variation in their preferred complexification strategies. The results of this dissertation highlight the importance of considering the methodological choices when interpreting the results of both group- and individual-level studies, as not every aspect of syntactic complexity can be identified using only traditional quantitative measures of complexity.

This dissertation critically discusses the operationalisation of syntactic complexity and the problem of choosing suitable complexity measures. Future studies would benefit from analyses that focus on several different linguistic levels and, ideally, combine quantitative and qualitative approaches. The new method of categorising productions into complexity profiles explored in the third substudy could be further developed in future research to shift the focus from statistically significant differences between groups of learners to the individual-level examination of syntactic complexity.

Individual-level variation is especially important in language teaching and assessment. By examining the different complexity profiles together with learners, teachers can guide the learners and their productions accordingly. Language learners should also be aware of the fundamental differences between speech and writing and recognise how these differences affect their use of syntactically complex structures in these two modes of production. Equally important for teachers and learners is to acknowledge the relationship between complexity and other features of learner language, such as accuracy. Due to the non-linearity of development and the constant interaction between complexity and accuracy, for instance, these features can show simultaneous progress and regress in learner language. All in all, it is crucial to regard syntactic complexity as a multifaceted construct in both teaching and research.