2 May: Olga Fehér

The effect of semantic cues on the regularisation of unpredictable variation

Olga Fehér (Edinburgh)

Tuesday 2 May 2017, 11:00–12:30
1.17 Dugald Stewart Building

Variation in natural language is constrained: languages tend to lose competing variants over time, and where variation persists its use is conditioned on linguistic or sociolinguistic context. When learners acquire languages that exhibit unpredictable variation (unnatural, unconditioned probabilistic variation), they often eliminate the variation by regularising to one of the competing variants or conditioning on context. We previously found that, in addition to individual learning and transmission, interaction can lead to regularisation through convergence and priming between interlocutors. In this experiment, we investigated the influence of semantic cues on regularisation and conditioning during interaction and transmission. We had adult participants learn and communicate with artificial languages which exhibited unconditioned variation in plural marking. The languages described images that belonged to one or two semantic categories. We found that interacting Dyads regularised in the one category condition by eliminating one of the markers. However, in the two category condition, Dyads maintained variation but without conditioning it on semantic categories. Semantic conditioning occurred only in Singles, gradually during episodes of transmission. The lack of conditioning in Dyads was probably due to strong priming between communicating partners that was present within and across semantic categories. This suggests that the pattern of restricted, conditioned variation in natural language reflects the combined influences of biases in learning, recall and interaction.

25 April: Rebecca Morley

Representational considerations in models of sound change

Rebecca Morley (Ohio State)

Tuesday 25 April 2017, 11:00–12:30
1.17 Dugald Stewart Building

One view of phoneme split takes it to be the result of divergent phonetic variants (e.g., Janda and Joseph 2003). Closely tied to this view is the hypothesis of iterativity: socially motivated phonetic exaggeration accumulating over successive generations (e.g., Labov 1972, Guy 1980), or progressive reduction of frequent words over time (Phillips 1984, Bybee 2002). Iterativity is often assumed to be an inherent property of exemplar models. In a typical scenario production starts with the selection of a token from the desired category. The token is then reduced, lenited, or otherwise altered in some way, resulting in a new phonetic token. The new token is added back to the cloud of stored tokens, and the process starts over again (see Pierrehumbert (2001)). Via this production-perception loop words can be reduced two or more times with respect to the originating token. As more frequent words are more often produced, the chances of multiply reduced tokens are higher. However, contrary to expectation, the mechanism described does not consistently result in shorter word lengths for high-frequency vs. low-frequency words. If frequency of occurrence is expressed in number of tokens, and sampling for production is random, then producing a less reduced token is also more likely in high, than low, frequency categories. And regardless of whether tokens decay, or are replaced, the low-frequency category will eventually ‘catch up’ with the higher-frequency category, and all words will achieve some optimal length. In fact, the production side of this model makes even more problematic predictions. If phoneme-level tokens are selected at random from a phonetically detailed exemplar cloud then egregious mismatch is possible; e.g., an [æ] originally followed by an [m] being selected for a pre-[b] context. It is the same at the word level: a word token originally produced in a frequent collocation, selected for a low-frequency context, etc. Indexing exemplar clouds with all the necessary contextual information, however, results in an explosion of categories, and a depletion of category members. In the limit, each category would contain a single member.

Developing a model for the interaction of synchronic variation and diachronic change requires resolving these and other representational issues, some of which only surface when the entire trajectory of change is considered. Thus, while existing models can capture category shift and merger (Pierrehumbert 2001), or contrast stability and dispersion (Garrett and Johnson 2013, Wedel 2004), there are few that can capture both. The model of Sóskuthy (2013) can generate phoneme split, no-change, and no-split with phonetic shift, as the result of vowel lengthening before voiced obstruents. However, these outcomes require a representational structure in which vowel categories contain at least two sub-categories: pre voiced-obstruent, and pre voiceless-obstruent. Crucially, these subcategories are semi-permeable, and greater frequency of occurrence can cause one sub-category to subsume the other. This scenario raises another unresolved question in exemplar modeling: the interaction between higher and lower level categories. Most models work exclusively at one level, and assume the others. But the process by which the necessary categories at the sub-word level are generated from the word level (or vice versa) is non-trivial, and may not be consistent with model assumptions. A category as abstract as “vowels occurring in environments followed by a voiced obstruent” requires a massive amount of generalization over words with different syllable structures, over obstruents at different places of articulation, etc. And if speakers create categories such as this, then they can be expected to create categories such as “vowels before coronals”, etc. It is not at all clear that existing models will be able to ‘scale up’ adequately under this added complexity.

This work gives a formal account of the representational commitments and assumptions of a range of models, and an assessment of their self-consistency. The claim is that the resolution of outstanding problems lies in determining the division between representations and processes. I argue, on the one hand, that phonetic effects such as “vowel lengthening”, or “vowel nasalization” are not processes themselves, but reside at the representational level. On the other hand, speaking rate must be able to apply after exemplar selection to compress or expand tokens as necessary to match speed of production. I consider prosodic effects, such as phrase-final lengthening to be necessarily processual as well. The ramifications of these representational choices are discussed with respect to the necessary constraints on a model deriving categorical sound change from existing synchronic variation.

18 April: Marieke Schouwstra, Simon Kirby & Jenny Culbertson

Word order universals reflect cognitive biases: Evidence from silent gesture and corpus statistics

Marieke Schouwstra, Simon Kirby, & Jenny Culbertson (Edinburgh)

Tuesday 18 April 2017, 11:00–12:30
1.17 Dugald Stewart Building

We investigate a hypothesized cognitive bias for isomorphic mappings between conceptual structure and linear order in the noun phrase. This bias has been proposed as a possible explanation for a striking asymmetry in the typology of the noun phrase – linear orders which place the adjective closest to the noun, then the numeral, then the demonstrative, are over-represented in the world’s languages.

Previous experimental work has provided evidence that an isomorphism bias affects English-speaking learners’ inferences about the relative order of modifiers in an artificial language. Here, we use the silent gesture paradigm to explore whether the isomorphism bias influences spontaneous gestures innovated by participants in a modality with which they have relatively little prior experience. We find that gesture order largely conforms to the same striking pattern found in noun phrase typology, supporting the role of the isomorphism bias in shaping the emergence of language (and language-like) systems.

Having shown that spontaneous production of gestures reveals a bias for isomorphism with conceptual structure we are left with an important question about the source of this bias. Is it “imposed by the nature of the human language capacity” as Adger (2017) suggests, or is this bias a result of a domain general constraint for simplicity interacting with a conceptual structure that is in principle learnable by the child from the structure of the world? To test this, we present preliminary results of an investigation of the relationships between objects, properties, numerosity, and locations using corpus statistics.

29 March: Kazuo Okanoya & Miki Takahasi

Wednesday 29 March 2017, 15:00–16:00
G15, 7 Bristo Square

Domestication and evolution of signal complexity in Bengalese finches

Kazuo Okanoya (University of Tokyo)

Among vocalizations birds make, a class of sounds that are consisted of more than two types of sound patterns arranged in a certain temporal sequence is called as a ‘birdsong’, not only because of the organization of sound patterns, but also because our musical aesthetics intuitively allow such an analogy. Scientific investigations of birdsong to date suggest that certain properties of birdsong extend beyond the musical to the developmental analogies Bengalese finches (BFs) are domesticated strains of wild white-rumped munias (WRMs) imported from China to Japan 250 years ago. BF songs are composed of multiple chunks and each chunk is a combination of 2-4 song notes. Furthermore, chunks are arranged in a finite-state probabilistic automaton. We studied how and why BFs sing such complex songs. We found the following facts. 1) The ancestral strain sing simpler songs. 2) There is high learning specificity in WRMs but not in BFs. 3) BFs have larger song control nuclei and higher level of glutamate receptor gene expressions than WRMs. 4) Both BF and WRM females prefer complex songs as measured by the nest string assay and males with complex songs are physically fitted than the males with simpler songs. These results promoted sexual selection scenario of song complexity in BFs. We further examined factors related with domestication. We examined songs of WRMs in subpopulations of Taiwan. Where there is a sympatric species to WRMs, songs were simpler. This leads to a hypothesis that in the wild songs needed to be simple to secure species identification, but under domestication this constrains was set free. We also examined socio-emotional indexes including neophobic tendency, tameness, behavioral stress reactions, and corticosterone levels. All indexes suggested that WRMs have higher level of stress and social shyness, which should be adaptive under natural environment, but could be limiting opportunities for learning complex songs. Thus, evolution of song complexity involves not only factors related with strengthen of sexual selection and relaxation of species identification, but also socio- emotional factors due to domestication. Furthermore, recent suggestion of ‘neural crest’ hypothesis that might account for the domestication syndrome fits well with the properties of Bengalese finches. These results on Bengalese finches must be useful in discussing possible biological origin of human speech in terms of proximate and ultimate factors.

The plasticity of song learning and tutor choice by cross-fostering and multiple tutor experiments in Bengalese finches.

Miki Takahasi (University of Tokyo)

The Bengalese finch (Lonchura striata var. domestica) is a domesticated strain of wild white-rumped munias (Lonchura striata). Through the domestication, the courtship songs of Bengalese finches have been differentiated from those of wild munias. The songs have tonal phonology and complex syntax in Bengalese finches, although munia songs have noisy sound with stereotyped syntax. The cross-fostering experiment revealed that there was significantly different in the accuracy of song learning between Bengalese finches and white-rumped munias. The proportion of shared elements with tutors was about 90 % regardless of tutor strain in Bengalese finches. In munias, they shared 75% song elements from the foster tutor (Bengalese finches) although they could copied almost 100% from own strain. Bengalese finches lost the accurate song learning, it will lead to increase the plasticity of song learning instead. We set a colony with eleven males and ten females of Bengalese finches and let them bred freely. Under the multiple tutor condition, most out of 32 juveniles learned parts of songs from two tutors. They segmented the songs by transition probability and combined some parts into own song. We also conducted a multiple tutor experiment in munias. 11 munia juveniles learned their father songs and only one juvenile learned from two tutors. While munias learned songs from several tutors, munias tended to learn from their familiar songs (their father). These results suggest that the accuracy of song learning is related to how many tutors juveniles select.

21 March: Tillmann Vierkant

Communication or control? The role of expressive behaviours in Non-Gricean routes to language evolution

Tillmann Vierkant (Edinburgh)

Tuesday 7 February 2017, 11:00–12:00
1.17 Dugald Stewart Building

Orthodoxy has it that language evolution requires Gricean communicative intentions and therefore an understanding of nested metarepresentations. The problem with the orthodoxy is that it is hard to see how non-linguistic creatures could have such a sophisticated understanding of mentality. Philosophers like Bar-On (2013) have therefore recently attempted to develop a non-Gricean account of language acquisition building on the information-rich and subtle communicative powers of expressive behaviours. This paper aims to sketch an alternative (additional) account of why expressive behaviours might be crucial in language evolution. On this account, expressive behaviours are important because not only because of their role in animal communication but because they are enablers of early forms of mindshaping.

17 March: Yasamin Motamedi

Artificial sign language learning: A method for evolutionary linguistics

Yasamin Motamedi (MPI Nijmegen)

Tuesday 17 March 2017, 15:00–15:30
3.10 Dugald Stewart Building

Previous research in evolutionary linguistics has made wide use of artificial language learning (ALL) paradigms, where learners are taught artificial languages in laboratory experiments and are subsequently tested about the language they have learnt. The ALL framework has proved particularly useful in the study of the evolution of language, allowing the manipulation of specific linguistic phenomena that cannot be isolated for study in natural languages. Furthermore, using ALL in populations of learners, for example with iterated learning methods, has highlighted the importance of cultural evolutionary processes in the evolution of linguistic structure.

In my thesis, I present a novel methodology for studying the evolution of language in experimental populations. In the artificial sign language learning (ASLL) methodology I develop, participants learn manual signalling systems that are used to interact with other participants. The ASLL methodology combines features of previous ALL methods as well as silent gesture, where hearing participants must communicate using only gesture and no speech. However, ASLL provides several advantages over previous methods. Firstly, reliance on the manual modality reduces the interference of participants’ native languages, exploiting a modality with linguistic potential that is not normally used linguistically by hearing language users. Secondly, cultural evolutionary research in the manual modality offers comparability with the only current evidence of language emergence and evolution in natural languages: emerging sign languages that have evolved over the last century.

The implementation and development of ASLL in the present work provides an experimental window onto the cultural evolution of language in the manual modality. I detail a set of experiments that manipulate both linguistic features (investigating category structure and verb constructions) and cultural context, to understand precisely how the processes of interaction and transmission shape language structure. The findings from these experiments offer a more precise understanding of the roles that different cultural mechanisms play in the evolution of language, and further builds a bridge between data collected from natural languages in the early stages of their evolution and the more controlled environments of experimental linguistic research.

14 March: Kearsy Cormier

Pronouns, agreement, classifiers and role shift: What sign languages can tell us about linguistic diversity and linguistic universals

Kearsy Cormier (UCL)

Tuesday 14 March 2017, 11:00–12:30
1.17 Dugald Stewart Building

The search for linguistic universals (and understanding universals in the face of diversity) is one of the key issues in linguistics today. Yet the vast majority of the linguistic research has focused only on spoken languages. Sign languages constitute an important test case for theories on universals and diversity, since a language “universal” only deserves this name if it holds both for signed and spoken languages, and languages in a different modality surely have much to teach us about the full range of diversity within human language. In this talk I consider four morphosyntactic/discourse phenomena found in sign languages that have traditionally been assumed to be the same as spoken languages but which, on closer inspection, reveal some fundamental differences relating to particular affordances of the visual-spatial modality. In order to understand these differences in more detail, linguists must consider the multimodal nature of human language (including gesture) rather than just the classic linguistic characteristics which are the exclusive focus of much work in mainstream approaches to the study of language.

7 March: Anna Jon-And

Modeling the role of acquisition in contact-induced language change

Anna Jon-And (Stockholm University)

Tuesday 7 March 2017, 11:00–12:30
1.17 Dugald Stewart Building

Accelerated language change in contact settings, especially language shift, has commonly been attributed to innovations during the second language acquisition process. Negative correlations have also been attested between proportions of non-native speakers and morphosyntactic complexity in cross-linguistic data. At the same time, cultural evolution experiments and computational models have revealed learnability as a general constraint in language evolution, suggesting that more learnable features, such as morphological simplicity, would be favored by all language acquisition and not only by second language acquisition. Here, I use agent-based computational simulations in order to test if diffusion of linguistic innovation in a language shift setting may result from a general acquisition effect reinforced by large proportions of learners, or if special weight needs to be attributed to second language acquisition. The simulations are informed by chronological demographic and linguistic data from the ongoing language shift from Bantu languages to Portuguese in Maputo, Mozambique. Parameters are set to proportions of native and non-native speakers over time and the model’s predictions are compared to variation in verbal morphology and use of locative prepositions in Portuguese.

28 February: Alex Papadopoulos-Korfiatis

An autopoietic approach to cultural transmission chains

Alex Papadopoulos-Korfiatis (Edinburgh)

Tuesday 28 February 2017, 11:00–12:30
1.17 Dugald Stewart Building

One of the problems of autopoiesis as a biological, bottom-up, non-representational theory of cognition is that it struggles with scaling up to high-level cognitive behaviour such as language. The Iterated Learning model, a theory of language evolution based on its transmission from agent to agent in cultural chains, is a promising candidate in providing the first step towards a non-representational account of language; our goal in this work is the combination of these two approaches. In order to do that, we introduce a minimal joint action “left/right dancing” task that can be solved in multiple ways. Through individual episodes of reinforcement learning between simulated robotic agents, we show that an initial expert agent’s behaviour persists in cultural transmission chains; we investigate the conditions under which these chains break down and re-emerge, drawing interesting parallels to existing Iterated Learning research.

21 February: Fredrik Jansson

Modelling the evolution of creoles

Fredrik Jansson (Stockholm University)
Joint work with Mikael Parkvall and Pontus Strimling

Tuesday 21 February 2017, 11:00–12:30
1.17 Dugald Stewart Building

We are interested in the contact situation, where several existing languages converge to one, a Creole. Various theories have been proposed regarding the origin of Creole languages. Describing a process where only the end result is documented involves several methodological difficulties. In this paper we try to address some of the issues by using a novel mathematical model together with detailed empirical data on the origin and structure of Mauritian Creole. Our main focus is on whether Mauritian Creole may have originated only from a mutual desire to communicate, without targeted learning, and we show that a minimal model can generate good predictions. With a confirmation bias towards learning from successful communication, the model predicts Mauritian Creole better than any of the input languages, including the lexifier French, thus providing a compelling and specific hypothetical model of how creoles emerge. The results also show that it may be possible for a creole to develop quickly after first contact, and that it was created mostly from material found in the input languages, but without inheriting their morphology.