I will be talking about some recent modelling work I’ve been doing that suggests that “Linguistic structure is an evolutionary trade-off between simplicity and expressivity”. Tuesday 26th, 11am, DSB 1.17.
“Linguistic structure is an evolutionary trade-off between simplicity and expressivity”
Language exhibits structure: a species-unique system for expressing complex meanings using complex forms. We present a review of modelling and experimental literature on the evolution of structure which suggests that structure is a cultural adaptation in response to pressure for expressivity (arising during communication) and compressibility (arising during learning), and test this hypothesis using a new Bayesian iterated learning model. We conclude that linguistic structure can and should be explained as a consequence of cultural evolution in response to these two pressures.
Nic Fay is in town and will give us a bonus LEC talk on “Evidence for the cumulative cultural evolution of human communication systems”. Friday 22nd February, 4pm-5.30pm, DSB 1.17. Abstract below.
Evidence for the cumulative cultural evolution of human communication systems
Biological evolution has endowed humans with a second inheritance system that allows people to socially learn from others. Key benefits of this cultural inheritance system is that it lets people sample cultural traits from many others (not just two parents as per genetic evolution) and it does so on a rapid timescale (relative to genetic change), enabling swift behavioral adaptation in response to changes in the environment. Whereas the literature has primarily focused on the progressive adaptation of technologies such as canoes and stone tools, less attention has been devoted to the cumulative cultural evolution of social artifacts such as customs and language. In this talk I will discuss the results of a recent laboratory experiment that examines the task performance of human participants organized into 8-person interactive and non-interactive transmission chains whose task is to communicate a route on a map to their partner. Progressive adaptation is seen in both conditions (i.e., the routes are reproduced with higher fidelity across generations in both conditions), but the mechanism behind this improvement differs across the interactive and non-interactive conditions. In the interactive condition performance is explained through interactive grounding processes, whereas in the non-interactive condition task performance is consistent with a biased transmission account.
[This was originally scheduled for 12th Feb, but was rescheduled]
Olga Feher will be giving a talk on ”The development and evolution of song culture”. Tuesday 19th, 11am, DSB 1.17. Abstract below.
The development and evolution of song culture
Similarly to humans, songbirds learn their songs by imitating adult individuals and achieve complex vocal behaviour gradually during a sensitive phase in early development. Young birds raised in complete social and acoustic isolation produce abnormal songs. However, when we tutored young birds with these abnormal, isolate songs in an iterated learning experiment, the songs gradually acquired normal, wild-type characteristics. Next, to find the minimum requirements for the emergence of wild-type song culture, we tutored young birds with their own developing songs, removing any external song input. This experiment externalized the song feedback and is similar to iterating the song learning within a single generation. Young birds trained with their own songs in such a manner produced much more normal songs than their isolate brothers. Lastly, I will briefly talk about the development of syllable ordering in zebra finches which seems to share important features with other songbird species and human babies.
Prof. Vera Kempe is visiting from the University of Abertay Dundee, and will give a talk on “Exploring Iterated Language Learning in Children”, detailing some recent work she’s been doing to adapt the method from the Kirby, Cornish & Smith (2008) PNAS paper for kids. Tuesday 5th February, 11am, DSB 1.17.
Rachael Bailes will give a talk about a paper she’s working on: “Levels of analysis in the common ground debate”, Tuesday 29th January, 11am, DSB 1.17.
Olga Fehér has joined the LEC on a prestigious Newton Fellowship. Olga is most famous for her research on zebra finches, in which she implemented for the first time an experimentally-controlled analog of cultural transmission of song from an initial non-typical starting point. In essence, this was experimental iterated learning of bird song. You can read more about this at the Nature blog related to her paper.
At the LEC, Olga hopes to expand her expertise to humans, with the ultimate aim of improving our understanding of the similarities and differences between iterated learning in birds and humans. We’re very excited to welcome her here!
We’ll kick off 2013 with a talk from Vanessa Ferdinand: “How we lose variation through learning: on the domain-generality of linguistic regularization biases”, Tuesday 15th January, 11am, DSB 1.17.