Language-specific morphological and metrical structure in speech processing: Comparative psycho- and neurolinguistic investigations of French and German

这个项目着眼于心理语言学中的两个主要问题。第一个问题涉及词识别的语音流的分割，第二个问题是形态学在访问和组织心理词典中的作用。鉴于口语中没有明确的单词边界，许多研究致力于了解语音处理系统如何设法隔离单词。另一方面，大量关于形态学的研究认为，在单词识别过程中，形态复杂的单词被分解为语素。然而，以前的研究形态处理，主要是处理视觉上呈现孤立的单词。本项目提出将单词识别的两个方面联合收割机结合起来，即，- 在词边界处分割语音流，并将复杂词分割成组成语素，以查看是否分解了在形态上复杂的词，以及该分解是否遵循连接语音分割的一般规则。计划中的行为和脑电图实验将更具体地研究形态加工如何与语音加工相互作用，以及词干和词缀实现的语音变化是否影响分解过程。特别是，我们将看到，如果在语素边界的语音变化，如重新音节化或段删除，防止形态分割。此外，进一步的研究将着眼于在单词识别过程中激活的心理词典中的单位的状态。事实上，即使在单词识别过程中提取了词素并激活了它们的表征，仍然有必要确定它们是否被表征为单词，或者它们是否对应于另一个表征水平，正如各种理论方法所建议的那样。在一组行为和脑电图实验中，我们将研究语素在心理词典中的地位，并与自由词的地位进行比较。德国和法国团队的这种合作将使两个合作伙伴在一般研究问题方面具有共同兴趣，另一方面具有互补的方法论专业知识，法国合作伙伴在心理语言学方面具有更强的背景，德国合作伙伴在神经语言学和语言学方面具有更强的背景。

This project looks at two major issues in psycholinguistics. The first one concerns segmentation of the speech stream for word recognition, and the second one the role of morphology in accessing and organizing the mental lexicon. Given that there are no clear word boundaries in spoken language, much research has been dedicated to understanding how the speech processing system manages to isolate words. On the other hand, a large body of research on morphology has argued that morphologically complex words are decomposed into morphemes during word recognition. Previous research on morphological processing, however, has mainly dealt with visually presented isolated words. The present project proposes to combine the two aspects of word recognition, i.e., segmentation of the speech stream at word boundaries and segmentation of complex words into component morphemes, to see if auditorily presented morphologically complex words are decomposed and if this decomposition follows general rules of connected speech segmentation. Planned behavioral and EEG experiments will more specifically look at how morphological processing interacts with phonological processing, and if phonological change of the realisation of the stem and the affix influence the decomposition process. In particular, we will see if phonological alternations at morpheme boundaries, such as resyllabification or segment deletion, prevent morphological segmentation. Moreover, further investigations will look at the status of the units activated in the mental lexicon during word recognition. Indeed, even if morphemes are extracted and their representations activated during word identification, it is still necessary to establish if they are represented as words or if they corresponded to another level of representation, as suggested in various theoretical approaches. In a set of behavioural and EEG experiments we will look at the status of morphemes in the mental lexicon, as compared to the status of free words. This collaboration of the German and the French teams will bring together two partners with joint interests in the general research questions on the one hand and complementary methodological expertise on the other, the French partner having a stronger background in psycholinguistics, and the German partner, in neurolinguistics and linguistics.