ABSTRACT
This chapter highlights the need for further conceptual clarification and operational refinement of Second Language (L2) complexity constructs, reliable and valid corpus annotation, and automated retrieval of non-redundant measures. Gradually, with the development of Natural Language Processing tools, researchers started applying a wide range of lexical measures to the study of L2 complexity. The overwhelming majority of syntactic complexity studies have based their findings on the material of academic L2 English essays written by university students. M. Paquot found that these lexico-grammatical measures reliably discriminated between upper-intermediate to advanced proficiency levels in a written L2 English corpus, unlike separate syntactic and lexical complexity measures. Register differences have been extensively studied by D. Biber and colleagues in their First Language corpus research spanning several decades since the 1980s. The effects of cognitive task complexity on the complexity of language production have received considerable attention in Task-Based Language Teaching Second Language Acquisition studies.
