Session 1B

Several studies have been conducted on the relationships between linguistic complexity and L2 development. One of the most widespread assumptions is that complexity grows over time, so that increasing proficiency in a second language would imply that one's productions become more and more complex. However, this assumption needs to be qualified and further investigated. Firstly, complexity does not grow at the same rate in different linguistic sub-domains. While the lexicon may increase constantly over the life span, with no exact limit to the number of words being learned, morphology and syntax have clear 'ceilings' with respect to the maximal levels that can be achieved. Secondly, linguistic complexity varies across different tasks and modalities, so that it is not always the case that 'the more, the better' - there are optimal levels of complexity, and sometimes more can actually mean worse, at least for certain linguistic structures and sub-domains (Biber et al 2011, Ortega 2003, Pallotti 2009). In other words, linguistic complexity must be interpreted, not just measured. This is particularly true today, with the availability of automatic tools calculating scores of complexity measures, and the associated risk of accumulating results with little reflection about how they should be understood in terms of language development and communicative appropriateness.
The presentation will first review previous literature discussing issues such as non-linear trajectories of complexity growth and optimal levels of complexity in different text types and genres. It will then present the results of two studies on the acquisition of Italian as an additional language. The first is a longitudinal four-year project involving adolescent learners, together with native speaker controls, performing a number of oral communicative tasks. The analysis will look at telephone calls, demonstrating that the degree of syntactic complexity (length and structural intricacy of clauses and AS-units) depends on the taks's interactional requirements. More specifically, higher levels of syntactic complexity compete with optimal levels of interactional complexity, that is, the need to rapidly exchange turns and to direct the interlocutor's attention. As learners progress in the L2, their syntactic complexity in this particular task tends to decrease, approaching native speakers' levels, while they develop more sophisticated interactional skills. The second study concerns primary school pupils' written productions, that were assessed in terms of both their functional adequacy (Kuiken & Vedder 2017) and their linguistic complexity. While lexical complexity, measured as the Moving-Average Type/Token Ration (MATTR, Covington & McFall 2010), was positively correlated to functional adequacy ratings, the relationship with morphological complexity (calculated with the Morphological Complexity Index, MCI, Brezina & Pallotti 2019) was more nuanced, and had to be interpreted also in light of the appropriateness of verb tense choices.
The conclusions are that the complexity of linguistic productions should be evaluated taking a number of factors into account. Some proposals will be made for interpreting complexity in the wider context of linguistic proficiency, communicative development and functional adequacy.

Second language development (SLD) has long been kept equal to measurable linear changes in L2 proficiency (Verspoor et al., 2021), mostly through the use of standardized tests (e.g., TOEFL) or the analysis of task-based oral or written performance data through a complexity, accuracy, and fluency (CAF) assessment framework. Although the latter has provided a broader understanding of L2 development, no form of systematicity has been found to confirm that each CAF construct conjointly develops over time (Bulté & Housen, 2020). The significance of the problem here is that if we learn more about the nature of the relationship between, for instance, syntactic and lexical complexity, we might provide useful insights to language classroom pedagogy. 
The current study operationalizes the nature of L2 development drawing on the tenets of complex dynamic systems theory (CDST). It aims to investigate the developmental trajectories of a group of sojourners with regard to written syntactic (SC) and lexical complexity (LC) over the course of an academic semester. As part of larger project (Pérez-Vidal, 2014), 26 Catalan/Spanish tertiary level sojourners learning English provided weekly diary entries over 14-17 weeks spent abroad regarding their experiences related to target language use, interaction, and host culture. Using the written performance data elicited through diaries, the researcher compiled a learner corpus of 383 diary entries including approximately 250K words. Following Kyle et al. (2021), the dataset was coded for various measures of syntactic and lexical complexity via CLAN (MacWhinney, 2000) and TAASSC (Kyle, 2016). To determine the nature of relationship (e.g., competitive, supportive, growing, or precursor) among the indices selected, a series of detrended correlations were run. Additionally, a series of linear mixed-effects models (LME) were developed to examine the relationship between time and these SC and LC measures. The preliminary results indicated a competitive relationship between syntactic and lexical complexity, along with high levels of individual variation confirming the CDST argument of individual learning trajectories. These results are anticipated to bring useful insights to language classroom pedagogy.


Bulté, B.& Housen, A. (2020). A DUB-inspired case study of multidimensional L2 complexity development: Competing or connecting growers? In W. Lowie, M. Michel, A. Rousse-Malpat, M. Keijzer, and R. Steinkrauss, Usage-Based Dynamics in Second Language Development (pp. 50-87). Bristol: Multilingual Matters.
Kyle, K. (2016). Measuring syntactic development in L2 writing: Fine grained indices of syntactic complexity and usage-based indices of syntactic sophistication (Doctoral Dissertation)
Kyle, K., Crossley, S., & Verspoor, M. (2021). Measuring Longitudinal Writing Development Using Indices Of Syntactic Complexity And Sophistication. Studies in Second Language Acquisition, 43(4), 781-812.
MacWhinney, B. (2000). The CHILDES project: Tools for analyzing talk. Mahwah, NJ: Lawrence Erlbaum Associates.
Pérez-Vidal, C. (2014). Language acquisition in study abroad and formal instruction contexts. Amsterdam: John Benjamins.
Verspoor, M., Lowie, W., & Wieling, M. (2021). L2 Developmental Measures from a Dynamic Perspective. In B. Le Bruyn & M. Paquot (Eds.), Learner Corpus Research Meets Second Language Acquisition (Cambridge Applied Linguistics, pp. 172-190). Cambridge: Cambridge University Press. doi:10.1017/9781108674577.009

Recent L2 complexity research has captured linguistic complexity across a wider range of language domains, specifically targeting complexity of phraseological units through the operationalization and measurement of phraseological diversity and sophistication (Paquot, 2019). Preliminary research has indicated that the diversity and sophistication of three types of grammatical dependency relations (verb + direct object, adjectival modifiers, and adverbial modifiers) contribute significantly, though not exclusively, to the prediction of L2 Dutch proficiency at the B1 and B2 CEFR levels, when modeled alongside traditional measures of syntactic and lexical complexity (Rubin et al., 2021).
While existing measures of phraseological complexity have focused on a rather limited set of phraseological units consisting of co-occurring lexical items, developments in the automatic annotation of linguistic information have expanded the potential for the extraction of phraseological units that consist of co-occurrences between lexical items and non-lexical grammatical patterns. The present study aims to incorporate such phenomena into the battery of measures tapping into the phraseological dimension of linguistic complexity, specifically targeting verb-argument structures in Dutch. To that end, diversity and sophistication of verb-argument structure units, respectively operationalized as the root type-token ratio (RTTR) and mean mutual information (MI) score of these units, are computed alongside previously investigated diversity and sophistication measures for dependency-based phraseological units (i.e. verb + direct object dependencies) in order to determine whether these phraseological units situated more closely to the lexis-grammar interface can also contribute to the prediction of learner proficiency and performance.
The analysis was carried out using 1,172 texts from an L2 Dutch corpus consisting of written extracts of the CNaVT (Certificate of Dutch as a Foreign Language) exams, administered by the Centrum voor Taal en Onderwijs at the KU Leuven. For each individual exam level, the overall numeric score for the exam is modeled as a function of the diversity and sophistication measures computed for both the dependency-based and verb-argument structure phraseological units in a linear regression model, controlling for the topic of the prompts at the B2 and C1 levels.
Measures of verb-argument structure complexity emerge as important and significant predictors of L2 Dutch proficiency across all three exam levels included in the present analysis, alongside measures of verb + direct object complexity. Some variation is observed across the three models in terms of the effects of the measures in the models, the direction of the effects, and the role of topic in moderating these effects. For the construct of phraseological complexity, measures of verb-argument structure sophistication, in particular, demonstrate strong potential as a quantitative measure to include alongside more established measures of dependency-based phraseological units.


References
Paquot, M. (2019). The phraseological dimension in interlanguage complexity research. Second Language Research, 35(1), 121-145.
Rubin, R., Housen, A., & Paquot, M. (2021). Phraseological complexity as an index of L2 Dutch writing proficiency: a partial replication study. In S. Granger (Ed.), Perspectives on the L2 phrasicon: the view from learner corpora. Bristol: Multilingual Matters.