ABSTRACT The distribution of different grammatical forms of collocations in real usage and its comparison to a pattern of the headword is a useful basis for thorough collocational analysis, which until now has lacked proper attention. In this article, we set out to investigate the relation between lexeme and token collocations, i.e., collocations that either follow the grammatical profile of their headword or differ from it significantly. Based on the [Adjective + Noun] collocations in Russian, we attempt to determine the most suitable way to describe collocations. The goal is twofold. First, we wish to prove that distinction between these two types makes sense in terms of their different distributional preferences across the corpus. Second, we offer a plausible method for differentiating between the two types based on corpus data about the collocations’ usage. The justification for this research is the lack of a single prevalent practice as far as the choice of representation of study object for empirical collocational analysis is concerned.

中文翻译：

---

在俄语搭配中选择词位还是标记

摘要 实际使用中搭配的不同语法形式的分布及其与词条模式的比较是进行彻底搭配分析的有用基础，但迄今为止一直缺乏适当的关注。在本文中，我们着手研究词素和标记搭配之间的关系，即，搭配要么遵循其词条的语法特征，要么与其显着不同。基于俄语中的[形容词+名词]搭配，我们试图确定最合适的方式来描述搭配。目标是双重的。首先，我们希望证明这两种类型之间的区别在它们在整个语料库中的不同分布偏好方面是有意义的。第二，我们提供了一种基于搭配使用的语料库数据来区分这两种类型的合理方法。这项研究的理由是缺乏单一的普遍做法，就经验搭配分析的研究对象表示选择而言。