SIGLTA meeting on Thursday 23rd May: Artificial Intelligence (AI) – Changing the Face of Formative and Summative Assessment

SIGLTAYou are cordially invited to attend the Special Interest Group in Language Testing and Assessment (SIGLTA) meeting. SIGLTA is a postgraduate student-led reading/research group within the Faculty of Arts and Humanities.

The meeting is at 17:00-18:30 on Thursday 23/05/2019 in room 1173, Avenue Campus (building 65), and will be led by Dr. Rose Clesham, the Director of Academic Standards and Measurement, Global Assessment, Pearson.

Abstract: Millions of English as a Second Language students are taught and assessed each year on both receptive (listening and reading) and productive (speaking and writing) skills for entry into English speaking universities or professions. These tests are high stakes and prospective candidates apply from across the world. So how can these skills be tested with high validity, reliability and lack of bias and obtain almost immediate feedback, accurate scoring and diagnostic information? This talk will describe and demonstrate how research and advances in Artificial Intelligence (AI) technologies have changed the way one of these testing agencies assesses and measures oral, aural, reading and written skills, on a global scale, using large worldwide data sets. Artificial Intelligence as a concept is not new, dating back over seventy years. However, enormous computing power and algorithmic advances now available enables AI and automated machine decision-making to effortlessly process big data and it is applied to many areas of society, from banking to entertainment.

In an educational assessment context, these AI technologies can be used for formative or summative purposes, and may in time replace both national and international tests and assessments. Public perception in this area has often focused on the lack of human interaction and judgement when automated marking technologies are used. This talk will demonstrate that in many ways, the opposite is true. The use of artificial technologies allows the judgement of hundreds of human assessors to work in unison, increasing validity in terms of broader content representation, and removing bias and low reliability issues. These technologies also significantly reduce teacher workload in terms of marking student work, yet still allow teachers to benefit from diagnostic feedback on their students, and releases valuable time to facilitate personalised learning.

The speaker: Rose is the Director of Academic Standards and Measurement, working in Global Assessment at Pearson. Her career started in teaching, and teacher education, before moving on to Governmental positions, responsible for running national assessment programmes in the UK. Her roles in Pearson have included leading Assessment Design and Research teams, carrying out national and international alignment and benchmarking studies, and presenting at major international conferences. Rose has also worked extensively on OECD PISA assessments, co-writing the 2015 Scientific Literacy Framework.

If you require any further information please send an email to or or see the SIGLTA Facebook page.

Next CLLEAR talk: Findings from GECO: The Ghent Eye-tracking Corpus of Monolinguals and Bilinguals Reading an Entire Novel

CLLEARThe next Centre for Linguistics, Language Education and Acquisition Research (CLLEAR) talk will take place on Wednesday 22nd May 2019 at 4:30pm in Lecture Theatre C, Avenue Campus (Building 65). The talk is entitled “Findings from GECO: The Ghent Eye-tracking Corpus of Monolinguals and Bilinguals Reading an Entire Novel” and will be delivered by Denis Drieghe from University of Southampton. All welcome!

Here is the abstract for this talk:
Up until recently, models of eye movements during reading almost exclusively focused on monolingual reading, even though most people are bilinguals. Relatively few studies examined eye movements in bilinguals and even less had a focus on sentence processing. Those few studies that did look at reading in a second language (L2) examined eye movements on a single embedded target word without taking into account changes in global eye movement behaviour that L2 reading might entail. In my talk I will present GECO (Ghent Eye-tracking COrpus), the freely available monolingual (English) and bilingual (Dutch- English) eye-tracking corpus of participants reading a complete novel (56,000 words). The aim of this project is to establish a more comprehensive description of eye movement behaviour in bilingual reading. We collected this large-scale data corpus from 14 English monolinguals and 19 Dutch (L1) / English (L2) unbalanced bilinguals of intermediate to high L2 proficiency who read the entire novel while their eye movements were being tracked (Bilinguals read half in Dutch, half in English). In this talk, I will present descriptive statistics of reading time measures for first-language (L1) and second-language (L2) reading. I will also present analyses of frequency, neighbourhood size and cognate effects for L1 and L2 sentence reading.

InnoConf19 on 28th June: Treasuring languages: innovative and creative approaches in HE

InnoConf 2019

It is with great pleasure that Modern Languages and Linguistics at the University of Southampton are hosting the 9th annual conference in the Innovative Language Teaching and Learning at University series, InnoConf19, this year. The conference will take place on Friday 28th June 2019. We are delighted to have professionals involved in the teaching and learning of languages share their expertise, innovative approaches and creative solutions to face the challenges ahead surrounding language teaching.

The theme of this year’s conference is ‘Treasuring languages: innovative and creative approaches in HE’. This theme aims to ignite discussions and seek innovative approaches dealing with creative ways to raise awareness about the value of learning languages.

You can follow the conference Twitter account at @innoconf. For further details about the conference, including the provisional programme, visit the InnoConf19 homepage. All welcome – please register for InnoConf19 here! We look forward to seeing you there!

LCAWB talk tomorrow: DeCentring the intercultural and small culture formation on the go – implications for internationalisation, research methods and who we all are.

Language & Culture in the Academic World & Beyond logo
As part of the Language & Culture in the Academic World & Beyond being run by PGR students in Modern Languages and Linguistics, there will be a talk by Professor Adrian Holliday from Canterbury Christ Church University on Friday 10th May 2019 at 4pm in Lecture Theatre B, Avenue Campus (Building 65). The talk is entitled “DeCentring the intercultural and small culture formation on the go – implications for internationalisation, research methods and who we all are.” All welcome – please register for the talk here!

Here is the abstract for the talk:
So much of what has become normalised regarding the intercultural now seems invalidated by Centre structures (e.g. associated with ‘the West’, native-speakerism, the neoliberal university, patriarchy, and methodological nationalism). This includes common large-culture concepts such as equating second language with second culture, quantifying intercultural competence, locating third spaces as between, interpreting integration as mixing with locals. The struggle to deCentre requires finding threads that dissolve assumed boundaries and are often not where we expect, beginning with transient and everyday small culture formation on the go. However, things are often not what they seem. The intercultural is everywhere; interculturality is messy and political; researchers and the researched merge; and data needs to be disturbed. I will attempt to look at these questions through reference to my forthcoming book with Sara Amadasi, Making sense of the intercultural: finding deCentred threads.

Next CLLEAR seminar: A dynamic typology of syntactic change in contact Englishes

CLLEARThe next Centre for Linguistics, Language Education and Acquisition Research (CLLEAR) seminar will take place on Wednesday 15th May 2019 at 4pm in Lecture Theatre C, Avenue Campus (Building 65). The talk is entitled “A dynamic typology of syntactic change in contact Englishes” and will be delivered by Devyani Sharma from Queen Mary, University of London. All welcome!

Here is the abstract for this seminar:
Postcolonial Englishes are an ideal testing ground for the influence of universals and language transfer in contact languages. However, the tendency to conduct static comparisons of contributing grammars to an outcome grammar often overlooks a dynamic aspect of the actuation problem, namely ‘why certain instances of variation become changes and others don’t’ (McMahon 1994). I assess this question in relation to contact settings: Why does only a subset of variable usage become entrenched over time in a given contact variety? An initial comparison of several syntactic features in Indian English and Singapore English supports a strong substrate, rather than universalist, basis for new usage. However, a closer examination shows that only some of these variable features have stabilised and become deeply embedded across the community. Substrates cannot fully account for this subtler distribution. They over-predict change. To better understand which forms become more entrenched, I turn to a sociohistorical hallmark of postcolonial Englishes: diminishing input from the source variety. Drawing on models of input sensitivity from second language acquisition theory (the Subset Principle; the Interface Hypothesis), I develop a two-dimensional typology to assess the relative role of substrate difference and input demand (the degree to which rich input is needed for the acquisition of a specific syntactic form) in stable outcomes in New Englishes. Both appear to be operative but substrates may be the more powerful force, as certain entrenched forms point to hard limits on learnability due to the substrate grammar, despite low input demand. Modelling contact as dynamic phases of individual learning embedded within a changing linguistic ecology helps to account for selective change over time within wider feature pools of variability. It also uses long-term outcomes of social contact to feed into theoretical questions of featural representations and learnability.