Delimiting native and foreign syntax in Early Slavonic by enhancing digital corpora: non-finite clauses as intermediate phenomena

我的研究项目的总体目标是通过增强数字语料库和采用定量方法，促进在旧教会斯拉夫语（OCS），现存最古老的斯拉夫手稿的语言本地和借用的句法元素的划界长期辩论。这场争论源于这样一个事实，即几乎所有现存的OCS文本都是从新约希腊语翻译而来的。这通常使得区分斯拉夫希腊语的calques和真正的斯拉夫语结构变得困难，特别是考虑到这两种语言总体上具有相对相似的语言结构。我的项目将试图评估语言外部因素和语言内部因素在决定OCS中非有限分句分布中的相对重要性：这些结构可以被视为假设的“本土光谱”的本地和外国两端之间的一组中间现象，因为它们通常被认为本质上是在希腊影响下扩大其范围的本地结构。为了从数量上证实这一想法，并更好地了解推动不同类型现象分布的复杂因素，我将使用不同的统计模型和两种不同的数据集。分类树在预测决定希腊-斯拉夫语谱中更极端现象行为的各种因素方面是一个成功的模型，例如组成顺序（最大程度地模仿希腊语）和所有格结构（实际上完全独立于希腊语）。将这种模型应用于非有限子句的评估预计会产生非常不同的结果：希腊语和语言内部预测因子都应该作为预测因子排名相对较高。希腊因素的相对排名可能表明每个结构的不同“本地水平”，而语言内部因素的相对重要性可能反过来激发它们在斯拉夫语语法语义界面中的地位。在我的项目中，关键的一步是将希腊-斯拉夫平行分析的结果与后来的原始文本（即后期教会斯拉夫文学和本土文学）中相同现象的分布进行比较。为了做到这一点，TOROT（选定的语料库）将通过添加相当数量的原始早期斯拉夫语文本（主要是南斯拉夫语）来增强，然后这些文本将被自动预处理（词法和词法的词性和标记），并对相关的句法现象进行战略性注释。

The overarching goal of my research project is to contribute to the long-standing debate on the delimitation of native and borrowed syntactic elements in Old Church Slavonic (OCS), the language of the oldest extant Slavonic manuscripts, by enhancing digital corpora and employing quantitative methods. The debate stems from the fact that virtually all of the extant texts in OCS are translations from New Testament Greek. This often makes it challenging to distinguish slavish Greek calques from genuinely Slavonic constructions, especially considering that the two languages have overall comparatively similar linguistic structures. My project would attempt to assess the relative importance of language-external and language-internal factors in determining the distribution of non-finite clauses in OCS: these constructions could be seen as an intermediate group of phenomena between the native and the foreign ends of a hypothetical 'spectrum of nativeness', as they are generally thought to be essentially native constructions which have expanded their scope under Greek influence. In order to substantiate this thought quantitatively and have a better picture of the complex set of factors motivating the distribution of the different types of phenomena, I would be using different statistical models and two different datasets. Classification trees have been a successful model in predicting the various factors determining the behaviour of more extreme phenomena in the Greek-Slavonic spectrum, such as constituent order (imitating Greek to the greatest extent) and possessive constructions (practically completely independent of Greek). The application of such a model to the assessment of non-finite clauses is expected to produce very different results: both the Greek and the language-internal predictors should rank relatively high as predictors. The relative ranking of the Greek factors might indicate different 'levels of nativeness' of each constructions, whereas the relative importance of the language-internal factors might in turn motivate their status within Slavonic at the syntax-semantic interface. A crucial step in my project would be comparing the results yielded in the Greek-Slavonic parallel analyses with the distribution of the same phenomena in later original texts (i.e. in Later Church Slavonic and vernacular literature).In order to do so, TOROT (the chosen corpus) would be enhanced by adding a considerable amount of original Early Slavonic texts (mostly South Slavonic), which would then be automatically pre-processed (lemmatised and tagged for part of speech and morphology) and strategically annotated for the relevant syntactic phenomena.