History is littered with unfulfilled promises that emerging technologies – from radios to televisions, and from computers to mobile phones – would completely transform teaching and learning. Now the same promises are being made of generative artificial intelligence (AI). This presentation argues that we should not be focusing on educational revolution, but instead on educational evolution. Education is a complex social, cultural, and political endeavour, serving multiple purposes and multiple stakeholders, and technology is just one of many elements in this large ecosystem.

Focusing on the context of language teaching and learning, this presentation discusses what has changed technologically, and suggests what could and should change educationally. It shows that ChatGPT and a range of other generative AI tools can contribute to language and literacy development in a number of ways, but that we need to be wary of their pedagogical, social, and environmental risks. Educators must develop the AI literacy necessary to take a more nuanced view of generative AI, and we must help our students to do the same.

This paper is based on a keynote presentation delivered at the English Australia Conference in Perth, Australia, on 12 September 2024, with some elaborations for the written version alongside minor updates to reflect more recent developments and publications.

We investigated which objective language proficiency tests best predict the language dominance, balance, English and Spanish proficiency scores relative to Oral Proficiency Interview (OPI) scores (averaged across 5–6 raters). Eighty Spanish–English bilinguals completed OPIs, picture naming, semantic and letter fluency, lexical decision tests and a language history questionnaire. Except for letter fluency, objective measures explained more variance than self-report variables, which seldom and negligibly improved proficiency prediction beyond objective measures in forward regression models. Picture naming (the Multilingual Naming Test (MINT) Sprint 2.0) was the strongest predictor for most purposes. Lexical decision and category fluency were next best predictors, but the latter was time-consuming to score, while the former was easiest to administer (and does not require bilingual examiners). Surprisingly, self-rated proficiency better predicted the OPI scores when averaged across modalities (i.e., including reading/writing instead of just spoken proficiency), and lexical-decision (a written test) was as powerful as picture naming for predicting spoken Spanish (but not language dominance).

Code-blending is the simultaneous expression of utterances using both a sign language and a spoken language. We expect that like code-switching, code-blending is linguistically constrained and thus we investigate two hypothesized constraints using an acceptability judgment task. Participants rated the acceptability of code-blended utterances designed to be consistent or inconsistent with these hypothesized constraints. We find strong support for the proposed constraint that each modality of code-blended utterances contributes content to a single proposition. We also find support for the proposed constraint that – at least for American Sign Language (ASL) and English – code-blended utterances make use of a single derivation which is realized using surface forms in the two languages, rather than two simultaneous derivations, one for each language. While this study was limited to ASL/English code-blending and further investigation is needed, we hope that this novel study will encourage future research comparing linguistic constraints on code-blending and code-switching.

Despite the central role of language teacher educators (LTEs) in contributing to the development of language teachers in higher education and non-higher education contexts, there is a lack of theoretical and empirical work on their professional lives. One such area that remains largely unexplored concerns the psychology of LTEs. This paper argues for the need to embrace a research program that systematically investigates aspects of LTE psychology in the face of the unique demands, challenges, and pressures this professional group needs to navigate in their complex situated reality. We first position LTEs as key agents in the educational enterprise and go on to problematize the current state of scholarship on this under-researched population in universities, schools, and other practical settings. We then present an empirically grounded discussion to justify why a more explicit focus on LTE psychology is essential, followed by a brief review of what is already known in this respect. In what follows, we outline several key directions future empirical work might take to generate a more in-depth and holistic account of LTE psychology. Overall, this paper portrays LTE psychology as a promising but under-explored area which merits particular attention in its own right.

We investigate the processing of scalar inferences in first language (L1) and second language (L2). Expanding beyond the common focus on the scalar inference from ‘some’ to ‘not all’, we examine six scalar expressions: ‘low’, ‘scarce’, ‘might’, ‘some’, ‘most’ and ‘try’. An online sentence-picture verification task was used to measure the frequency and time course of scalar inferences for these expressions. Participants included native English speakers, native Slovenian speakers and Slovenian speakers who spoke English as their L2. The first two groups were tested in their L1, while the third group was tested in their L2. Results showed that the English-L2 group resembled the Slovenian-L1 group more than the English-L1 group in terms of inference frequency. The time course for scalar inference computation was similar across all groups. These findings suggest subtle pragmatic transfer effects from L1 to L2, varying across different scalar expressions.