Maltese plurals: phonotactics, variation and the structure of the mentallexicon.

第一阶段表明，马耳他语使用者依靠他们的词汇来概括单复数映射到新单词的复杂性。他们的概括依赖于他们所知道的单数词的语音和伴随的复数形式。这种复杂的映射只能通过大词典来实现：对于一个新词的单复数映射，相似词的队列越小，泛化的置信度就越小。另一个重要的发现是，成年人对待正确的复数和破碎的复数是不同的。然而，尚不清楚的是，这是否是一个模型的证据，在这个模型中，关于发音复数的概括是规则应用的结果，而关于断裂复数的概括是模式联想的结果。 为了更好地理解这个问题，我们将研究奇异-复数映射的获取。儿童词汇量的不断增长将为我们提供极好的数据，以测试他们是仅仅依赖类比，还是依赖规则和类比。如果孩子们依赖类比，我们希望在所有年龄段都有频率和语音的影响。如果他们依赖于规则和类比，我们预计会有一些泛化是独立于词汇在他们的词典中的频率和其他依赖于词汇在他们的词典中的频率。 本项目将通过研究3岁、5岁、7岁和9岁儿童的习得来检验这些期望。孩子们关于单复数映射的直觉会随着词汇的增长而变化，换句话说，随着年龄的增长而变化。 我们将通过在不同的计算模型中对我们的习得研究结果进行建模来补充我们的实验，以测试关于概括的性质及其背后的处理的特定假设。

The first phase has shown that Maltese speakers rely on their lexicon to generalize the complexities of the singular-plural mappings to novel words. Their generalizations depend on the phonotactics of singular words they know and the accompanying plural forms. Such complex mappings can only be achieved with a large lexicon: The smaller the cohort of similar words for a singular-plural mapping for a new word, the smaller the degree of confidence for a generalization. Another important finding is that adults treat sound plurals different from broken plurals. What is not clear, though, is whether this is evidence for a model in which generalizations concerning sound plurals are the result of rule application and generalizations concerning broken plurals are the result of pattern associations.  We will study the acquisition of singular-plural mappings in order to better understand this problem. The growing lexicon of children will provide us with excellent data to test whether they rely on analogies alone or on rules and analogies. Should children rely on analogies, we expect there to be effects of frequency and phonotactics across all ages. Should they rely on rules and analogies, we expect there to be some generalizations that are independent of the frequencies of words in their lexicon and others that do rely on the frequency of words in their lexicon.  The present project will test these expectations by studying the acquisition of 3-year-olds, 5-year-olds, 7-year-olds and 9-year-olds. The intuitions of children concerning the singular-plural mappings are then expected to change with a growing lexicon—in other words, with age.  We will complement our experiments by modelling the results of our acquisition studies in different computational models to test specific hypotheses about the nature of the generalizations and the processing that underlies them.