This data set contains data for three studies (each involving several experiments) looking at the cognitive and interactional mechanisms that underlie regularisation in language. All three studies used the artificial language paradigm to teach adult or child participants miniature languages exhibiting unpredictable variation and observed their learning and use of these languages. Feher et al. (2016) used adult participants and languages with variation in word order. In Exp. 1, participants viewed pictures of four novel objects presented in different numbers (2, 3 or 4) or with different surface characteristics (spotty, blue or furry) along with the descriptions of the images in a semi-artificial language with variable word order. Participants then had to recall the language by providing descriptions of the same images. Exp. 2 used videos of puppet animals performing simple actions on each other and descriptions of these actions in a semi-artificial language exhibiting variation in word order of the verb-argument structure. Participants then used this language in a simple communication game with each other. The authors measured regularisation and investigated the communicative mechanisms in different interactive situations. The results provide evidence for structural priming (reusing the partner's previously used word order) in human-human and human-computer interaction. Priming occurred regardless of behavioral convergence: communication led to shared word order use only in human-human interaction, but priming was observed in all conditions. Furthermore, interaction resulted in the reduction of unpredictable variation in all conditions, suggesting a role for communicative interaction in eliminating unpredictable variation. Samara et al. (2017) used a semi-artificial language learning paradigm to investigate learning of the sociolinguistic cue of speaker identity in 6-year-olds and adults. Participants were trained and tested on an artificial language where nouns were obligatorily followed by one of two meaningless particles and were produced by one of two speakers (one male, one female). Particle usage was conditioned deterministically on speaker identity (Experiment 1), probabilistically (Experiment 2), or not at all (Experiment 3). Participants were given tests of production and comprehension. In Experiments 1 and 2, both children and adults successfully acquired the speaker identity cue, although the effect was stronger for adults and in Experiment 1. In addition, in all three experiments, there was evidence of regularization in participants' productions, although the type of regularization differed with age: children showed regularization by boosting the frequency of one particle at the expense of the other, while adults regularized by conditioning particle usage on lexical items. Overall, results demonstrate that children and adults are sensitive to speaker identity cues, an ability which is fundamental to tracking sociolinguistic variation, and that children's well-established tendency to regularize does not prevent them from learning sociolinguistically conditioned variation. Smith et al. (2017) explored the relationship between learners' biases and common linguistic features such as the scarcity of unpredictable variation. This study presented a Bayesian model of the learning and transmission of linguistic variation along with a closely matched artificial language learning experiment with adult participants. The participants were trained on a miniature language for describing simple scenes involving moving animals, where every scene consisted of one or two animals performing a movement, and the accompanying description consisted of a nonsense verb, a noun, and (for scenes featuring two animals) a post-nominal marker indicating plurality. This plural marker varied unpredictably: sometimes plurality was marked with the marker fip, sometimes with the marker tay. After training on this miniature language, participants labelled the same scenes repeatedly, generating a new miniature language. The language produced by one participant was then used as the training language for the next participant in a chain of transmission, passing the language from person to person. The results show that while the biases of language learners can potentially play a role in shaping linguistic systems, the relationship between biases of learners and the structure of languages is not straightforward. Both the experimental and the computational results imply that biases for regularity in individual learners may not be enough to engender predictability in natural languages: while transmission can amplify weak biases, in some circumstances it can produce the opposite effect, masking learner biases. Furthermore, the use of language during interaction can reshape linguistic systems. The languages of the world are superficially rather different: they employ different sounds and build complex utterances in different ways. However, all languages seem to share a common set of structural properties. Where do these fundamental properties of language come from? One influential hypothesis suggests that they are a reflection of a highly-constraining blueprint for language, which children impose on language during learning. An alternative possibility is that these properties might be due to weaker biases in learners, which have a very small impact for a particular child learning a particular language, but have a strong effect as a result of the transmission of language over thousands of episodes of language learning. This grant seeks to address both these hypotheses. One strand of research applies artificial language learning methods, where children and adults attempt to learn miniature languages, to test for differences between adults and child learners, and the extent to which single learners can re-shape a language. A second strand features experiments where adult participants learn, transmit and interact using artificial languages, to simulate the transmission and use of language in a population and its consequences for linguistic structure.

Child and adult participants were tested individually in the laboratory. Participants interacted with a computer program that automatically trained and tested them and recorded their output data. Every experiment was overseen by an experimenter. The precise methods varied according to the particular study - the publications are listed in Related resources.