The Evolution of Complex Text Patterns:Development and Application of a corpus-linguistic Approach for Analysis of Diachronic Change of Text Patterns in its Multidimensionality

这里提出的项目的目标是根据最近语言史上两种高度相关的体裁(报纸和宗教文学)，为文本模式的历时变化确定适当的指标。尽管语篇模式在语言史上具有相当重要的意义，但到目前为止，只对很小的语篇样本，特别是这里讨论的语篇类型，研究了语篇模式的历时变化。因此，在前期工作中，我们开发了一个描述文本模式历时变化的模型，现在我们建议系统地测试、调整和进一步发展基于语料库的方法。这个潜在的多维模型借鉴了现有的关于语言指标(首选语法结构等)的研究成果。用于描述文本类型。这些具体指标将在拟议项目过程中以大量文本为基础进行核实。在这样做的过程中，所提出的模型将以自动和手动文本分析以及定量和定性检查相互交织和相互影响的方式进行操作。具体地说，我们建议在这里系统地研究文本的结构、功能、社会和文体维度的模式。由于与这些维度相关的模式可以部分从文本表面获得，因此我们可以借鉴传统计算语言分析工具的结果。此外，我们计划在还不能用自动方法可靠地提取特征的地方应用手动标注(特别是对于历史文本)。提取的特征将经过统计分析，这种量化方法的结果将在实际文本来源的基础上以定性的方法进行验证和解释。相反，作为定性分析结果的人工注释将以定量的方法加以利用。这项工作将一方面以报纸为基础，另一方面以虔诚的文学为基础，从而包括两种体裁，这两种体裁在他们那个时代的广泛接受者的日常生活中具有稳固的地位。正在审议的语料库的多样性预计将导致关于总体文本模式历时变化的表现的深远结果，从而确保项目过程中开发的技术适用于进一步的文本类型分析。

The project proposed here targets the identification of suitable indicators for diachronic change of text patterns based on two highly relevant genres in recent language history (newspapers and devotional literature). Despite its considerable importance for the history of language, diachronic change of text patterns has thus far only been examined with respect to very small text samples, particularly for the text types under consideration here. Therefore, in the course of preliminary work we developed a model for the description of diachronic change of text patterns, which we now propose to systematically test, adapt, and further develop in a corpus-based approach.This underlying multi-dimensional model draws on existing research results regarding linguistic indicators (prefered grammatical structures etc.) for the characterisation of text types. These specific indicators are to be verified with respect to a large text basis in the course of the proposed project. In doing so, the proposed model will be operationalised in such a way that automatic and manual text analysis as well as quantitative and qualitative examination intertwine and mutually inform one another.Concretely, we propose here to systematically investigate patterns of the structural, functional, social, and stylistic dimensions of texts. Since patterns associated with these dimensions can partly be acquired from the textual surface, we can draw on the results of conventional computational linguistic analysis tools. In addition, we plan to apply manual annotation wherever features cannot yet be reliably extracted with automatic methods (especially for historical texts). The extracted features will undergo statistical analysis and the results of this quantitative method will then be verified and interpreted in a qualitative approach based on the actual text sources. Conversely, the manual annotations being results of qualitative analyses will be exploited with quantitative methods.The work will be based on newspapers on the one hand, and devotional literature on the other hand, thus including two genres which had a firmly established place in everyday life for a wide range of recipients in their time. The variety of the corpora under consideration is expected to lead to far-reaching results regarding the manifestations of diachronic change of text patterns in general, thus ensuring the applicability of the techniques developed in the course of the project to the analysis of further text types.