U00521/P00218 (Honours/MSc course)
How is language learned, and how do we process language? What
drives language to change, and how can we predict the time-course of that
change? What are the origins of language and how did it evolve in humans?
These are some of the questions that can be tackled by building working
models of language in computer simulations.
Aims and general description
Although many consider the only interface between computing and linguistics
to be speech technology (where linguistic knowledge is used to solve engineering
problems), increasingly in recent years computers are being used to explore
fundamental problems of theoretical linguistics.
The aim of this course is to give an introduction to some of the techniques
of computational modelling and simulation and show how they have been used
to shed light on debates in: language acquisition, language change, and language
evolution. The focus will be mainly on learning how to assess the literature
in these areas, rather than a practical course on simulation techniques.
However, simulation packages are available and a small practical project
will be an option for part of the course assessment (see below).
There are no official prerequisites for this course. The course is predominantly
theoretical rather than practical, so programming expertise is not
a requirement. That said, a practical project is an option for those
that are interested.
Syllabus (subject to change - details will appear on WebCT)
- Introduction: what is a model? and how does it relate to theory?
What areas of linguistics could benefit from computational models?
- Neural networks: a non-mathematical introduction to connectionist
approaches to modelling learning.
- Language acquisition and connectionism: a survey of how neural
network models have become central in debates about how children learn
- Language change: a brief overview of how simulations are beginning
to be used to understand why historical changes seem to have a particular
"shape", and where language universals come from.
- Language evolution: an introduction to how computational techniques
such as "genetic algorithms" might be combined with models of acquisition
and change to eventually solve the problem of language origins.
The course will be based around readings mostly drawn from the recent
research literature. Each week there will be two lectures introducing the reading
and placing it in a broad context, and one tutorial with more detailed group
discussion of the readings.
There is no set text for this course. Readings for the majority of the
course will be research papers and chapters. These will be available as photocopies
in the filing cabinets in the common room, and will be made available a week
before the tutorial sessions discussing them. Along with each reading will be a
set of questions to be discussed in the tutorial session. More details will be
made available on the WebCT page for the course. However, readings will include
- Bechtel, W. and Abrahamsen, A. (1991). Connectionism
and the Mind. Chapter 1. "Networks versus Symbol Systems: Two
approaches to modeling cognition" pp. 1-20
- Elman, J. et. al. (1996). Rethinking
Innateness. Chapter 2. "Why connectionism?" pp. 47-66
J. et. al. (1996) pp. 66-106
& McClelland 1986 “On learning the past tenses of English Verbs”
- Pinker & Prince (1989)
"Rules and connections in Human Language"
Elman, J. (1993). "Learning
and development in neural networks: the importance of starting small"
Mitchell, M. (1996) "An introduction to
Genetic Algorithms" pp 1-16
Batali, J. (1994) "Innate
biases and critical periods: combining evolution and learning in the
acquisition of syntax".
Christiansen, M. H. and Devlin, J. T.
(1997). "Recursive inconsistencies are hard to learn: A connectionist
perspective on universal word order correlations".
Christiansen, M. H. and
Ellefson, M. R. (2002). "Linguistic Adaptation Without Linguistic Constraints:
The Role of Sequential Learning in Language Evolution".
Batali, J. (1998) "Computational
Simulations of the Emergence of Grammar".
Kirby, S. & Hurford, J.
(2001). "The Emergence of Linguistic Structure: An overview of the Iterated
Learning Model." In Angelo Cangelosi and Domenico Parisi, editors,
Simulating the Evolution of Language, pages 121--148. London: Springer
(2003) "Anchoring of semiotic symbols" Robotics and Autonomous Systems.
Vogt, P. (2000)
"Bootstrapping grounded symbols by minimal autonomous robots" Evolution of
Communication 4(1): 89-118
The course assessment will consist of an exam in May which will combine an essay question and a number of