This paper focuses on automatically assessing language proficiency levels according to linguistic complexity in learner English. We implement a supervised learning approach as part of an automatic essay scoring system. The objective is to uncover Common European Framework of Reference for Languages (CEFR) criterial features in writings by learners of English as a foreign language. Our method relies on the concept of microsystems with features related to learner-specific linguistic systems in which several forms operate paradigmatically. Results on internal data show that different microsystems help classify writings from A1 to C2 levels (82% balanced accuracy). Overall results on external data show that a combination of lexical, syntactic, cohesive and accuracy features yields the most efficient classification across several corpora (59.2% balanced accuracy).

Antipsychotic treatment resistance affects up to a third of individuals with schizophrenia. Of those affected, 70–84% are reported to be treatment resistant from the outset. This raises the possibility that the neurobiological mechanisms of treatment resistance emerge before the onset of psychosis and have a neurodevelopmental origin. Neuropsychological investigations can offer important insights into the nature, origin and pathophysiology of treatment-resistant schizophrenia (TRS), but methodological limitations in a still emergent field of research have obscured the neuropsychological discriminability of TRS. We report on the first systematic review and meta-analysis to investigate neuropsychological differences between TRS patients and treatment-responsive controls across 17 published studies (1864 participants). Five meta-analyses were performed in relation to (1) executive function, (2) general cognitive function, (3) attention, working memory and processing speed, (4) verbal memory and learning, and (5) visual−spatial memory and learning. Small-to-moderate effect sizes emerged for all domains. Similarly to previous comparisons between unselected, drug-naïve and first-episode schizophrenia samples v. healthy controls in the literature, the largest effect size was observed in verbal memory and learning [dl = −0.53; 95% confidence interval (CI) −0.29 to −0.76; z = 4.42; p < 0.001]. A sub-analysis of language-related functions, extracted from across the primary domains, yielded a comparable effect size (dl = −0.53, 95% CI −0.82 to −0.23; z = 3.45; p < 0.001). Manipulating our sampling strategy to include or exclude samples selected for clozapine response did not affect the pattern of findings. Our findings are discussed in relation to possible aetiological contributions to TRS.

Until recently, use has been made almost exclusively of text-based concordancers in the analysis of spoken corpora. This article discusses research being carried out on Padua University's Multimedia English Corpus (Padova MEC) using the multimodal concordancer MCA (Multimodal Corpus Authoring System, Baldry, 2005). This highly innovative concordancer enables the retrieval of parts of video and audio from a tagged corpus and access to examples of language in context, thereby providing non-verbal information about the environment, the participants and their moods, details that can be gleaned from a combination of word, sound, image and movement. This is of use to language learners of all levels because if “communication is to be successful, a relevant context has to be constructed by the discourse participants” (Braun, 2005: 52). In other words, transcripts alone are not sufficient if learners are to have anything like participant knowledge and comprehend spoken language. In the article it will be demonstrated how language functions expressed in the multimedia corpus of spoken English are retrieved using MCA. Online learning materials based on the multimodal concordances take into consideration not only language, but also the way in which it co-patterns with other semiotic resources, thereby raising the issue of the importance of learner awareness of the multimodal nature of communication.