Tuesday 3rd Feb, 11am-12:30, G32, 7 George Square
Jasmeen Kanwal (UCSD)
Zipf’s Principle of Least Effort from a diachronic perspective
“The magnitude of words tends, on the whole, to stand in an inverse (not necessarily proportionate) relationship to the number of occurrences.” This is the Principle of Least Effort (PLE) as stated by George Kingsley Zipf in 1935. This inverse relationship between word length and word frequency (and a closely related inverse relationship between word length and predictability in context) has since been found to hold synchronically for many languages. One explanation for the ubiquity of this pattern is that it results from a pressure on language users to optimise efficiency of communication; by assigning shorter forms to more frequent or predictable meanings, while assigning longer forms to less frequent or predictable meanings, we maximise the chances of successful communication while minimising overall effort (see, e.g., Piantadosi et al. 2011). If communicative pressure indeed plays a role in pushing languages towards alignment with the PLE, then we should expect to observe a diachronic effect: as the frequency or average predictability of a meaning increases with time, the average length of the corresponding form should decrease with time, and vice versa. An ideal testing ground for this hypothesis is provided by synonymous (or near-synonymous) word pairs that differ in length, for example clipped pairs such as ‘info’ and ‘information’. As a meaning becomes more common over time, the relative frequency of the shorter form should increase, and vice versa. I will present the results of a large-scale corpus study–focusing, for the time being, on English–that supports this hypothesis, thus suggesting that communicative efficiency does indeed play some role in driving languages towards greater alignment with the PLE over time.
Tuesday 27th Jan, 11am-12:30, G32, 7 George Square
Mark Atkinson
Adult language learning and sociocultural determination of linguistic complexity
Languages spoken by larger groups have less complex morphology than those of smaller communities, although it is not immediately clear why this would be the case. One explanation considers the proportion of non-native speakers, which is larger in larger groups. As adults find complex morphology challenging to learn, these languages may be under a greater pressure to simplify.
I’ll describe a recent experiment which investigates these proposed effects of group size and non-native speakers on the transmission of morphological complexity. A set of input data was first created by training participants on a morphologically-complex miniature language over 8 rounds of training and testing. The languages at Round 2, exhibiting morphological simplifications as a result of more limited exposure to the target language, made up a set of “non-native” data. The more complex languages at Round 8 then made up a set of “native” speaker data.
This native and non-native speaker data was then used to construct the inputs for a second group of participants, and so allow us to investigate whether different group types affect how complex a language is acquired. 4 different social conditions were considered: a large group of native speakers, a small group of native speakers, a large group of both native and non-native speakers, and a small group of both native and non-native speakers.
Despite large differences in the input data received by the learners in the different groups, there was no difference in the complexity of the languages they acquired. I’ll discuss the reasons for this (arguably surprising) result, what this means for the hypothesis that adult learning causes morphological simplification, and some possible directions for future research.
Tuesday 16th December 11.00am, DSB 1.17
From Item to System in Cultural Evolution Experiments
Simon Kirby (with Andrea Ravignani & Tania Delgado)
A wide range of iterated learning experiments show a strikingly similar pattern of results: over multiple transmission episodes systematicity increases cumulatively. What begin as independently learned behaviours start to form systems of interdependencies. Randomness transforms into structure; complexity yields to simplicity. This result is so common – despite very different framing of these experiments – it is tempting to see it as a universal law of the cultural evolution of sets of complex behaviours.
This matters because it leads us naturally to the hypothesis that the emergence of language in the real world involves a transition from independent signals to increasingly interconnected systems of signals. It leads us to treat the major design features of language as a natural consequence of cultural evolution.
In this talk I’ll address two questions:
1. Does this emerging systematicity depend on the task being one of first learning a set of explicitly language-like behaviours followed later by a production phase?
2. Are there phenomena other than language that we might characterise as involving cultural transmission of sets of complex behaviours?
My answers will be “no” and “yes” respectively, and I hope to demonstrate this by showing you some *very* preliminary results of an iterated drumming experiment.
*** note unusual day, time and venue ***
Friday 12th December 11am, DSB 3.10
Damian Blasi (MPI Leipzig)
Linguistics meets data science
My work stems from classic questions in typological-functional linguistics: what aspects of languages are similar or different across lineages and regions? Do similar patterns of use, social composition and/or ecology shape similar grammars? I will argue, however – uncontroversially I hope – that counting languages in contingency tables is simply not enough to address these questions, and that the methods and research strategies used in the data sciences are those which will most reliably lead to a coherent, empirical and typologically-informed science of language. I will illustrate this with three examples from my research: the existence of robust and widespread sound-meaning associations in basic vocabulary, commonalities underlying creole languages, and causation within the structure of the lexicon.
Tuesday 9th December 11.30am, DSB 1.17
Ashley Micklos (UCLA)
Interaction’s role in emerging communication systems and their conventionalization: Eye gaze, turn-taking, and repair
When considering the emergence and evolution of language, it is important not to disregard how we use language: in a dynamic interaction in which resources work in concert with one another to make complex communication possible. The data and research presented here address the nature and role of discourse features, namely repair, eye gaze, and turn-taking, in an experimental language evolution setting. Thus, we see how an emerging silent gesture system is negotiated, changed, and conventionalized in a dyadic interaction. Using a conversation analytic approach, we can glean how communicative devices themselves emerge and evolve within a language system. For example, the strategies for and frequency of repair may be indicative of the stage of evolution/conventionalization of a given language system. These features are essential to natural conversation, and the data suggest they may also be culturally transmitted.
*** note unusual day, time and venue ***
Friday 5th December 4pm, DSB 3.10
Rick Janssen (MPI Nijmegen)
Anatomical biasing of speech sounds: an empirically grounded agent model
The Darwinian principles of variation, selection and reproduction have had widespread success in explaining the emergence of biological complexity. Similar principles might be at work in other domains. Computer models on cultural evolution of language have explained features such as recursion, compositionality and vowel distributions that mirror real-world counterparts without any invocation of biologically innate language modules.
We draw on past hypotheses that speakers’ anatomy constitutes a biasing factor in cultural evolution. As such, speakers with different vocal tract morphologies can be associated with subtle, yet different costs in producing particular sounds. Larger ethnic groups that share a similar morphology might hereby converge on vowel distributions that differ from others. Our study employs an agent model to investigate such anatomical biases. Notably, our agents are characterized by realistic 3D geometries of the vocal tract, making empirical grounding and subsequent validation feasible.
We hypothesize that the shape of the hard palate is one prime source of anatomical biasing. As such, our study uses Bezier curves to model the hard palate’s mid-sagittal profile. Using five parameters, we constrain the degrees of freedom of our curve. Curve fitting by application of an evolutionary algorithm shows that the standard deviation of a fit to sample is low in every case. As a result, we are now able to automate the integration of e.g., MRI samples into our 3D model.
More job news: Justin Sulik, who recently completed his thesis in the LEC (supervised by Kenny and Jim, thesis title “Cognition at the Symbolic Threshold”), is off to join Gary Lupyan’s group at the University of Wisconsin-Madison, starting January 2015. Good luck Justin!
*** note unusual day, time and venue ***
Friday 28th November 4pm, DSB 3.10
Thom Scott-Phillips (Durham University)
Non-human primate communication, pragmatics, and the origins of language
Comparisons with the cognition and communication of other species have long informed discussion of the origins of human communication and language. This research has often focused on similarities and differences with the linguistic code, but more recently there has been an increased focus on the social cognitive foundations of linguistic communication. However, exactly what these comparisons tell us is not clear, because the theoretical concepts used in the animal communication literature are different to those used in the corresponding literature on human communication, specifically those used in linguistic pragmatics. In this article, I bridge the gap between these two areas, and in doing so specify exactly what great ape communication tells us about the origins of human communication and language. I conclude that great ape communication probably does not share the same social cognitive foundations as linguistic communication, but that it probably does involve the use of metacognitive abilities that, once they evolved to a more sophisticated degree, were exapted for use in what is an evolutionarily novel form of communication: human ostensive communication. This is turn laid the foundations for the emergence of linguistic communication. More generally, I highlight the often neglected importance of pragmatics for the study of language origins.
Tuesday 25th November 11am, DSB 1.17
Alan Nielsen
Motivated vs conventional systematicity: Implications for language learning.
Two complementary streams of research in the past decade have suggested that systematic associations between words and meanings are beneficial for language learners. On the one hand, work into sound symbolism has suggested that the motivated connections between certain characteristics of words and meanings are responsible for this learning benefit. At the same time, artificial language learning work has more generally suggested that systematic languages are easier to learn. In this talk I will present data from an experiment designed to explore the differences between these types of learning, suggesting that the sole benefit for motivated systematic languages over conventional ones is in the earliest stages of learning.