Cambridge Language Sciences launch event

This Cambridge-internal event is an opportunity to come and meet colleagues working in language research across the University. Please come along, even (and perhaps especially) if your own particular area of research isn't covered. This event does not define the scope of the Initiative, but aims to open up discussion and promote links between different departments.

Registration is now closed and the conference is full. The presentations will be posted on the website after the conference.

Registration (9.30-10.00)

Introduction and welcome (10.00-10.15)

Lynn Gladden, Senior Pro-Vice-Chancellor for Research

Simon Franklin, Head of School of Arts and Humanities

Session 1: Language Communication and Comprehension (10.15-11.15)

Neurobiological and evolutionary perspectives on language

William Marslen-Wilson, Dept. of Experimental Psychology

Robert Foley, Leverhulme Centre for Human Evolutionary Studies

Language and communication in the modern human is supported by a coalition of neurobiological systems whose properties we need to understand in their evolutionary context. We sketch out some recent hypotheses about the properties of these systems and how these relate both to our primate relatives and to some possible scenarios for the emergence of language in humans.

Prediction and adaptation in speech comprehension: evidence from brain imaging

Matt Davis, MRC Cognition and Brain Sciences Unit

This talk reviews evidence from behavioural and brain imaging studies for two key computational processes which support human speech comprehension: prediction and adaptation. These computational mechanisms play a key role in explaining how word recognition achieves optimal speed and accuracy despite the degraded speech sounds that we hear in everyday communication.

Philosophy of language on language communication

Kasia Jaszczolt, Dept. of Theoretical and Applied Linguistics

Language communication relies on many different vehicles in conveying speaker's meaning. One of the main and still unresolved questions is how to represent the interaction of information that comes from natural language sentences with information coming from widely understood context. The pertinent question that occupies philosophers of language is whether semantics should be minimalistic and confined to the meaning of natural language sentences or should it account for the meaning of the 'inner sentence' of the speaker's thought as it is claimed in some versions of the so-called contextualism. In this talk I discuss a radical contextualist approach to semantics and show how it solves some problems with representing meaning of particularly problematic constructions called "de se" belief reports.

[coffee break]

Session 2: Language Learning across the Lifespan (11.30-12.30)

Introduction

Henriëtte Hendriks, Dept. of Theoretical and Applied Linguistics

Auditory entrainment, phonology and dyslexia

Usha Goswami, Dept. of Experimental Psychology and Centre for Neuroscience in Education

I will provide a brief overview of the different projects at the Centre for Neuroscience in Education on phonological development in dyslexia, rhythmic entrainment and neuronal oscillations. I will focus on the previously neglected role of sensitivity to speech rhythm and prosody in developmental dyslexia, discuss the research on auditory processing in dyslexia that led us to focus on prosody, and our current "temporal sampling" framework for understanding the neural basis of individual differences in phonological development.

The acquisition of quantification across languages

Napoleon Katsos, Dept. of Theoretical and Applied Linguistics

In this talk I investigate the acquisition of words like ‘none’ and ‘most’ by children speaking one of 25 languages representing 10 different language genera. Particular emphasis is given to the emergence of cross-linguistically similar patterns (and to the motivations for such uniformity), as well as lexical and grammatical features which lead to language-specific differentiation.

What is learnable in a second language?

Teresa Parodi, Dept. of Theoretical and Applied Linguistics

The outcome of the first language acquisition process is inevitably the mother tongue. What is learnable in a second language, however, is a matter for ongoing debate. This talk illustrates the problem on the basis of pronominal clitics in Serbo-Cration as a second language, touching upon the issues of categorical versus optional features, as well as type and amount of input.

Session 3: Human Language Technologies (12.30-1.30)

Introduction

Ann Copestake, Computer Laboratory

Analysing scientific literature

Simone Teufel, Computer Laboratory

I will give an overview of several projects at the Natural Language and Information Processing Group concerned with the automatic analysis of scientific papers; in particular, the recognition of rhetorical moves in the scientific argument. The applications of this work include summarisation, cancer risk assessment, and the prediction of emerging topics in a field.

Speech transcription

Phil Woodland, Dept. of Engineering

This talk will briefly describe approaches for automatically producing transcriptions for speech data while dealing with the challenges of a wide variety of acoustic conditions, speakers, styles and languages.

Spoken dialogue systems

Steve Young, Dept. of Engineering

This talk will present a brief review of work on Statistical spoken dialogue systems which maintain an explicit representation of beliefs, and use reinforcement learning to learn how to respond to the user. The result is greater robustness to recognition errors, reduced development costs and the ability to adapt to users on-line.

Statistical machine translation

Bill Byrne, Dept. of Engineering

[Lunch in Sidgwick Hall, 1.30-2.30]

Session 4: Cambridge English (2.30-3.30)

Michael Milanovic, Cambridge ESOL

Nick Saville, Fiona Barker, Angeliki Salamoura, Cambridge ESOL

In recent years a number of interdisciplinary research projects have also been set up involving other departments, including the Department for Theoretical and Applied Linguistics (DTAL), the Computer Laboratory and the Speech Group within the Engineering Department.  This presentation will focus on two areas of collaboration which are already well underway:

the English Profile Programme, and in particular the ways in which researchers in Cambridge having been analysing learner data, including corpora, to understand progression across levels of proficiency (www.englishprofile.org);

the development of automated systems for the assessment of English speech and writing based on computational techniques derived from natural language processing.

[coffee break]

Session 5: Topics in Language Diversity and Change (3.45-4.45)

Ian Roberts, Dept. of Theoretical and Applied linguistics

Geoffrey Khan, Faculty of Asian and Middle Eastern Studies

So many languages are studied from so many different perspectives in Cambridge that it is impossible to give a representative overview in one hour. The three presentations offered here represent three different strands of research on language change and diversity: the study of structural variation, modelling variation, and the pressing issue of language endangerment. These represent different aspects of these large areas of research, and at least give a sense, really just by scratching the surface, of the actual and potential research in these areas.

Introduction

Ian Roberts

Ergativity

Michelle Sheehan Dept. of Theoretical and Applied Linguistics

Modelling change and diversity

Theresa Biberauer and Paula Buttery, Dept. of Theoretical and Applied Linguistics

Endangered languages

Mari Jones, Dept. of French

This talk will present endangered languages as a sub-discipline of Linguistics and will outline the ways in which the field is represented within the University.

Closing discussion (4.45-5.00)

Chaired by William Marslen-Wilson and Henriëtte Hendriks.