A B S T R A C TThis research explores the role of phonotactic probability in two-yearolds' production of coda consonants. Twenty-nine children were asked to repeat CVC non-words that were used as labels for pictures of imaginary animals. The CVC non-words were controlled for their phonotactic probabilities, neighbourhood densities, word-likelihood ratings, and contained the identical coda across low and high phonotactic probability pairs. This allowed for comparisons of children's productions of the same coda consonant in low and high phonotactic probability environments. Children were significantly more likely to produce the same coda in high phonotactic probability non-words than in low phonotactic probability non-words. These results are consistent with the hypothesis that phonotactic probability is a predictor of coda production in English. Moreover, this finding provides further evidence for the role of the input and distribution of sound patterns in the ambient language as a basis for phonological acquisition.

A common model for question answering (QA) is that a good answer is one that is closely related to the question, where relatedness is often determined using generalpurpose lexical models such as word embeddings. We argue that a better approach is to look for answers that are related to the question in a relevant way, according to the information need of the question, which may be determined through task-specific embeddings. With causality as a use case, we implement this insight in three steps. First, we generate causal embeddings cost-effectively by bootstrapping cause-effect pairs extracted from free text using a small set of seed patterns. Second, we train dedicated embeddings over this data, by using task-specific contexts, i.e., the context of a cause is its effect. Finally, we extend a state-of-the-art reranking approach for QA to incorporate these causal embeddings. We evaluate the causal embedding models both directly with a casual implication task, and indirectly, in a downstream causal QA task using data from Yahoo! Answers. We show that explicitly modeling causality improves performance in both tasks. In the QA task our best model achieves 37.3% P@1, significantly outperforming a strong baseline by 7.7% (relative).

In many languages the issue arises as to whether an onglide patters as part of the syllable onset or forms the first part of a (rising) diphthong with the immediately following vowel. If it is part of the syllable onset, the structure of a CGV syllable would be as in (1), but if it forms the first part of a diphthong the structure of a CGV syllable could either be as in (2a) with a monomoraic diphthong (where the glide is ‘co-moraic’ with the following vowel) or as in (2b) with a bimoraic diphthong (C = consonant, G = glide, V = vowel, μ = mora, and σ = syllable).

No abstract

No abstract