Elizabeth Wonnacott

Acquiring and Processing Verb Argument Structure: Distributional Learning in a Miniature Language.

Adult knowledge of a language involves correctly balancing lexically-based and more language-general patterns. For example, verb argument structures may sometimes readily generalize to new verbs, yet with particular verbs may resist generalization. From the perspective of acquisition, this creates significant learnability problems, with some researchers claiming a crucial role for verb semantics in the determination of when generalization may and may not occur. Similarly, there has been debate regarding how verb-specific and more generalized constraints interact in sentence processing and on the role of semantics in this process. The current work explores these issues using artificial language learning. In three experiments using languages without semantic cues to verb distribution, we demonstrate that learners can acquire both verb-specific and verb-general patterns, based on distributional information in the linguistic input regarding each of the verbs as well as across the language as a whole. As with natural languages, these factors are shown to affect production, judgments and real-time processing. We demonstrate that learners apply a rational procedure in determining their usage of these different input statistics and conclude by suggesting that a Bayesian perspective on statistical learning may be an appropriate framework for capturing our findings.

Eliminating unpredictable variation through iterated learning.

Human languages may be shaped not only by the (individual psychological) processes of language acquisition, but also by population-level processes arising from repeated language learning and use. One prevalent feature of natural languages is that they avoid unpredictable variation. The current work explores whether linguistic predictability might result from a process of iterated learning in simple diffusion chains of adults. An iterated artificial language learning methodology was used, in which participants were organised into diffusion chains: the first individual in each chain was exposed to an artificial language which exhibited unpredictability in plural marking, and subsequent learners were exposed to the language produced by the previous learner in their chain. Diffusion chains, but not isolate learners, were found to cumulatively increase predictability of plural marking by lexicalising the choice of plural marker. This suggests that such gradual, cumulative population-level processes offer a possible explanation for regularity in language.

Variability, negative evidence, and the acquisition of verb argument constructions

We present a hierarchical Bayesian framework for modeling the acquisition of verb argument constructions. It embodies a domain-general approach to learning higher-level knowledge in the form of inductive constraints (or overhypotheses), and has been used to explain other aspects of language development such as the shape bias in learning object names. Here, we demonstrate that the same model captures several phenomena in the acquisition of verb constructions. Our model, like adults in a series of artificial language learning experiments, makes inferences about the distributional statistics of verbs on several levels of abstraction simultaneously. It also produces the qualitative learning patterns displayed by children over the time course of acquisition. These results suggest that the patterns of generalization observed in both children and adults could emerge from basic assumptions about the nature of learning. They also provide an example of a broad class of computational approaches that can resolve Bakers Paradox.

Language learning, language use and the evolution of linguistic variation

Linguistic universals arise from the interaction between the processes of language learning and language use. A test case for the relationship between these factors is linguistic variation, which tends to be conditioned on linguistic or sociolinguistic criteria. How can we explain the scarcity of unpredictable variation in natural language, and to what extent is this property of language a straightforward reflection of biases in statistical learning? We review three strands of experimental work exploring these questions, and introduce a Bayesian model of the learning and transmission of linguistic variation along with a closely matched artificial language learning experiment with adult participants. Our 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. Weak biases can have strong effects on language structure as they accumulate over repeated transmission. But the opposite can also be true: strong biases can have weak or no effects. Furthermore, the use of language during interaction can reshape linguistic systems. Combining data and insights from studies of learning, transmission and use is therefore essential if we are to understand how biases in statistical learning interact with language transmission and language use to shape the structural properties of language. This article is part of the themed issue ‘New frontiers for statistical learning in the cognitive sciences’.

Is children's reading "good enough"? Links between online processing and comprehension as children read syntactically ambiguous sentences.

We monitored 8- and 10-year-old childrens eye movements as they read sentences containing a temporary syntactic ambiguity to obtain a detailed record of their online processing. Children showed the classic garden-path effect in online processing. Their reading was disrupted following disambiguation, relative to control sentences containing a comma to block the ambiguity, although the disruption occurred somewhat later than would be expected for mature readers. We also asked children questions to probe their comprehension of the syntactic ambiguity offline. They made more errors following ambiguous sentences than following control sentences, demonstrating that the initial incorrect parse of the garden-path sentence influenced offline comprehension. These findings are consistent with “good enough” processing effects seen in adults. While faster reading times and more regressions were generally associated with better comprehension, spending longer reading the question predicted comprehension success specifically in the ambiguous condition. This suggests that reading the question prompted children to reconstruct the sentence and engage in some form of processing, which in turn increased the likelihood of comprehension success. Older children were more sensitive to the syntactic function of commas, and, overall, they were faster and more accurate than younger children.

High or Low? Comparing high- and low-variability phonetic training in adult and child second language learners

Background High talker variability (i.e., multiple voices in the input) has been found effective in training nonnative phonetic contrasts in adults. A small number of studies suggest that children also benefit from high-variability phonetic training with some evidence that they show greater learning (more plasticity) than adults given matched input, although results are mixed. However, no study has directly compared the effectiveness of high versus low talker variability in children. Methods Native Greek-speaking eight-year-olds (N = 52), and adults (N = 41) were exposed to the English /i/-/ɪ/ contrast in 10 training sessions through a computerized word-learning game. Pre- and post-training tests examined discrimination of the contrast as well as lexical learning. Participants were randomly assigned to high (four talkers) or low (one talker) variability training conditions. Results Both age groups improved during training, and both improved more while trained with a single talker. Results of a three-interval oddity discrimination test did not show the predicted benefit of high-variability training in either age group. Instead, children showed an effect in the reverse direction—i.e., reliably greater improvements in discrimination following single talker training, even for untrained generalization items, although the result is qualified by (accidental) differences between participant groups at pre-test. Adults showed a numeric advantage for high-variability but were inconsistent with respect to voice and word novelty. In addition, no effect of variability was found for lexical learning. There was no evidence of greater plasticity for phonetic learning in child learners. Discussion This paper adds to the handful of studies demonstrating that, like adults, child learners can improve their discrimination of a phonetic contrast via computerized training. There was no evidence of a benefit of training with multiple talkers, either for discrimination or word learning. The results also do not support the findings of greater plasticity in child learners found in a previous paper (Giannakopoulou, Uther & Ylinen, 2013a). We discuss these results in terms of various differences between training and test tasks used in the current work compared with previous literature.

Learning: Statistical Mechanisms in Language Acquisition

The grammatical structure of human languages is extremely complex, yet children master this complexity with apparent ease. One explanation is that we come to the task of acquisition equipped with knowledge about the possible grammatical structures of human languages—so-called “Universal Grammar”. An alternative is that grammatical patterns are abstracted from the input via a process of identifying reoccurring patterns and using that information to form grammatical generalizations. This statistical learning hypothesis receives support from computational research, which has revealed that even low level statistics based on adjacent word co-occurrences yield grammatically relevant information. Moreover, even as adults, our knowledge and usage of grammatical patterns is often graded and probabilistic, and in ways which directly reflect the statistical makeup of the language we experience. The current chapter explores such evidence and concludes that statistical learning mechanisms play a critical role in acquisition, whilst acknowledging holes in our current knowledge, particularly with respect to the learning of ‘higher level’ syntactic behaviours. Throughout, I emphasize that although a statistical approach is traditionally associated with a strongly empiricist position, specific accounts make specific claims about the nature of the learner, both in terms of learning mechanisms and the information that is primitive to the learning system. In particular, working models which construct grammatical generalizations often assume inbuilt semantic abstractions.

Semantic conditioning in interaction and transmission

A central question in language evolution research is how fundamental properties of language have evolved and how that evolutionary process is shaped by human cognition. One property observed in all natural languages is variation. Linguistic variation tends not to be random and fully unpredictable. Rather, it is conditioned on the linguistic or social environment (Givón, 1985): linguistic or social context deterministically or probabilistically predicts the use of linguistic variants. Previous research has shown that when children acquire artificial languages containing unpredictable variation, they often eliminate the variation by overusing one of the variants (e. g. Hudson Kam & Newport, 2009). However, at present there is no satisfying experimental account of why natural languages should contain so much conditioned variation or how conditioning comes about. We investigated the evolution of conditioned variation using an artificial language paradigm that included transmission and interaction. We presented participants with images of objects accompanied by their descriptions in an artificial language. Depending on experimental condition, the objects were drawn from either one semantic category (e.g. all objects were animals) or two semantic categories (a mix of animals and vehicles). Each description consisted of a nonsense verb, a noun for the object and, for scenes involving multiple objects, a variable plural marker. The plural was marked by one of two markers (e.g. dak and fip) which occurred equally frequently in our initial experimenter-designed languages. 111