The semantics of derivational morphology

英语（以及其他语言）中的许多衍生过程都是高度多义的，通常表现出一整套的意思。例如，-er编码代理（作家），工具（开瓶器），居民（伦敦人）和更多种类的概念。该项目涉及衍生词法的中心问题，衍生过程的多义性和构词过程的组合性。它旨在描述和模拟衍生词的潜在不同解释。主要研究问题如下：-根据词根和词缀的含义，哪些解释主要是可能的？-是否存在一套有限的语义机制，可以以原则性的方式解释衍生读数？-世界知识的作用是什么？如何将其整合到衍生语义学理论中？该项目延续了之前的一个项目，在这个项目中，我们研究了一些衍生类别（-ment， -er和-al名词化，刻板印象否定和小词）。我们设计了基于框架的词法表示，并将HPSG的元素与词汇规则的继承层次结合起来，建立了一个框架，在这个框架中进行基于框架的构词模式分析。在框架语义方法中，多义性产生于表征中引用不同属性的可能性。不可能引用要么通过不可能继承出现，要么通过词缀特定的约束出现。在新的资助期内，该项目将扩展调查领域，包括(a)与非言语事件相关的名词化和(b)产生标量或基于程度的解释的前缀的形态学类别。这将使我们能够同时关注改变类别和保持类别的词缀，同时处理后缀和前缀。包含这些特征的形态过程对于建立一个全面的构词语义模型至关重要。此外，我们将使用语义模型的计算实现（在XMG和Analogical Modeling中）来测试它们的有效性和生产力。

Many derivational processes in English (but also in other languages) are highly polysemous, often exhibiting a whole range of meanings. For example, -er encodes agents (writer), instruments (opener), inhabitants (Londoner) and many more kinds of concept. The project deals with this central problem of derivational morphology, the polysemy of derivational processes and the compositionality of word-formation processes. It aims at describing and modeling the potentially diverse interpretations of derived words. The main research questions are the following:- Which kinds of interpretation are principally possible given the meaning of the base and the affix? - Is there a restricted set of semantic mechanisms that can account for derivational readings in a principled way? - What is the role of world knowledge and how can it be integrated into a theory of derivational semantics? The project continues a previous project in which we have investigated a number of derivational categories (-ment, -er and -al nominalizations, stereotype negation and diminutives). We devised frame-based semantic representations of the base words undergoing nominalizations and established a framework in which to couch our frame-based analyses of word-formation patterns, combining elements from HPSG with inheritance hierarchies of lexical rules. In a frame-semantic approach polysemy arises from the possibility to reference different attributes in the representations. Impossible reference either emerges through impossible inheritance or via affix-specific constraints. In the new funding period the project will extend the domain of investigation to include the morphological categories of (a) non-deverbal event-related nominalizations and (b) prefixes that give rise to scalar or degree-based interpretations. This will allow us to focus on both category-altering and category-maintaining affixation and at the same time deal with both suffixation and prefixation. The inclusion of morphological processes with these characteristics is crucial in order to develop a comprehensive model for word-formation semantics. Furthermore, we will use computational implementation of our semantic models (in XMG and Analogical Modeling) to test their validity and productivity.