Tues 11th November, 11am-12:30, DSB 1.17
Yasamin Motamedi
The emergence of systematic structure in artificial gestural communication systems
Languages exhibit systematic structure: signals are not independent of each other but form part of a system. Previous work has shown that the emergence of systematic structure increases learnability of a system, and the pressures of transmission drive this emergence (Kirby, Cornish and Smith, 2008; Cornish, Smith and Kirby, 2013). Additionally, a link has been posited a between systematic structure and the arbitrariness of a systems signals; that is, as signs become more arbitrary, the systematic re-use of signs increases, aiding learnability (Theisen, Oberlander and Kirby,2010; Theisen-White, Kirby and Oberlander, 2011).
In this talk I will present a study that looks at how systematic structure arises in manual communication systems. I will present results that suggest the gradual emergence of systematic structure in the systems participants created. I will present analysis that attempts to understand which aspects of the sign are affected as systematic structure emerges, relating the results to structures found in natural sign languages.
Tuesday 14th October, 11am-12:30, DSB 1.17
Joe Fruehwald
How Phonetic Changes Happen
In this talk, I’ll walk through a careful case study of a change in pronunciation that took place in Philadelphia across the 20th century which is based on acoustic analysis of archival recordings. The goal is to revisit some first principles about how changes like these take place. For people with an interest in language change, but without specific background in sound changes, this should should be informative about the properties of phonetic changes that are distinct from other kinds of language change. For people with a specific background in sound change, I hope to challenge the conventional wisdom that phonetic changes are driven primarily by bottom-up factors, like coarticulation, sampling with noise, etc. Rather, there seem to be important top-down, categorizing factors which play a crucial role from the very onset of this change.
Very best of luck to Vanessa Ferdinand, Catriona Silvey and Bill Thompson, who have finished up their PhDs in the LEC and headed off to fantastic post-doc positions elsewhere.
Vanessa (thesis title: “Inductive Evolution: the cognitive explanation of cultural change and regularity in language”) is off to the Santa Fe Institute, to take up an Omidyar Fellowship.
Cat (thesis title: “The Communicative Emergence and Cultural Evolution ofWord Meanings”) is joining Susan Goldin-Meadow’s lab at the University of Chicago.
Last but not least, Bill (thesis title: “Transmission, Induction, and Evolution”) is moving to Brussels, to work on Bart de Boer’s ERC-funded project at the Vrije Universiteit Brussel.
Don’t forget to come back and visit us from time to time!
Congratulations to Dr James Thomas and Dr Justin Sulik, who were awarded their PhDs this week. James did his PhD with Simon Kirby and Richard Shillcock (Informatics), on “Self-domestication and Language Evolution”; since his PhD he has been working as a post-doc with Richard. Justin was supervised by Kenny Smith and Jim Hurford, and his thesis was titled “Cognition at the Symbolic Threshold: The role of adductive inference in hypothesising the meaning of novel signals”. He is currently a course tutor for psychology and linguistics on the University’s International Foundation and Open Studies Programmes.
Tuesday 7th October, 11am-12:30, DSB 1.17
Jennifer Culbertson
Distinctions among cue types in semantically-based noun class learning
Work on natural language noun class acquisition suggests that phonological information – even if less reliable – may be privileged over semantics (e.g., Gagliardi, 2012). If phonological information is available to the child before meaning, then any privilege for phonological cues could be due to evidence accumulation rather than active down-weighting of semantics. We show that by removing the problem of acquiring meaning, classes in an artificial language can be readily learned based on semantics alone. However, as with phonological cues, some semantic distinctions may be more cognitively accessible than others. We compare shape- and flexibility-based systems and show that the former is learned more readily, in line with typology.
Tuesday 30th September 11.00-12.30, room 1.17 Dugald Stewart Building
The Dynamics of Social Learning on Undirected Networks
Andrew Whalen (University of St. Andrews)
Social networks, and dynamics of processes operating over them have been an exciting area of recent research. However in a wide variety of theoretically interesting processes relating to language and cultural evolution, the structure of the population has minimal effects on the average dynamics of the process. To explore this problem, we examine the influence of a network on three dynamic processes. We find that the learner’s learning rule has a substantial impact on whether or not there exist differences between networks; social structure plays a limited role when learners use a linear learning rule, but may play a larger role (particularly the network’s clustering coefficient) when learners us a non-linear learning rule. We also find that in cases where the difference between networks is minimal, there may be drastic differences in the dynamics of the process depending on where on the network a behavior originated from.
*** Note unusual room ***
Tues 23rd September, 11am – 12:30, Lecture theatre 5, 7 Bristo Square
Olga Feher
Social influences on the regularisation of unpredictable linguistic variation
Languages tend not to exhibit unpredictable variation, and learners receiving variable linguistic input tend to eliminate it, making the language more regular. We explore how this behavior is influenced by social cues, in particular when variability is distributed within and across teachers. We trained participants on an artificial language that contained lexical variability and manipulated how that variability was distributed across teachers: learners either received input from one or three variable teachers, or from three teachers who were individually consistent but exhibited variability collectively. We found that learners were more likely to produce variable output when their input came from (one or multiple) teachers who exhibited variable labeling, and they regularized more when learning from individually consistent teachers. To see whether our finding was due to conformity effects or imitation of the linguistic variability of teachers, we ran a further two conditions, and the results suggest that the effect is indeed driven by matching the variability of speakers, rather than by conformity effects.
*** Note unusual room ***
Tuesday 16th September, 11am – 12:30, Lecture theatre 5, 7 Bristo Square
Marieke Schouwstra
Basic word order: from natural order to convention in silent gesture
Interest in the origins of basic word order in human language has increased dramatically over the last five or six years. I will present my contribution to this area and describe experimental work in which adult participants communicate about simple events using only gesture, both alone and in dyadic interaction.
The alone-condition leads to conditioning of word order on the semantic structure of the event that is described: extensional events give rise to SOV word order, while intensional events result in SVO word order (Schouwstra & de Swart, 2014). These two orders arise independently of the dominant order of the participants’ native language, and I will claim that they represent naturalness: they are cognitively the most intuitive way to impose linear structure on information.
I have also investigated the behaviour of the two natural word orders when they are used in dyadic interaction to observe the effect of naive participants communicating repeatedly using only gesture. I will focus on the emergence of linguistic conventions, and discuss the dynamics of this process by looking at the influence of naturalness, regularisation, and the relative frequency of the events being communicated.
Tues 9th September 11am-12.30, DSB 1.17
Isabelle Dautriche (Laboratoire de Sciences Cognitives et Psycholinguistique)
Title: Dealing with ambiguity in the lexicon: a challenge for language acquisition
Abstract: The lexicon is ambiguous: word forms can have several meanings (homophones) and form a highly connected network in the phonological space with regions of high neighborhood density. While this organization might be functional for a mature language processing system, it poses a number of challenges for the language learner. Using lexical models and experimental methods in toddlers and adults, I investigate the organization of lexicons and how does it affect learning of similar sounding words (neighbors) and homophones. The study of homophones and confusable words reveal crucial details of learners? implicit knowledge and assumptions about how word forms map onto meanings and how the semantic lexicon as a whole is structured.
Tuesday 19th August 11.00-12.30, room 1.17 Dugald Stewart Building
Alex Papadopoulos-Korfiatis
A dynamical systems approach to iterated learning
Both dynamical systems accounts of cognition and the Iterated Learning model of language evolution can be regarded as holistic, systemic approaches that base their explanatory power on the interaction of a number of factors.
A potential combination of the two can be beneficial for both but a number of changes must be made to the standard IL model to make it compatible with the constraints that dynamical accounts of cognition introduce.
In this talk I’ll describe the above constraints and present an experimental set-up based on a coordination task that takes them into account.