Stable multiword unites represent a large class that unites different linguistic phenomena, which include compounds, idioms, colligations, and collocations. The nature of these units is distinguished by lexical (semantic), syntactic or statistical idiosyncrasy. Due to their peculiarities, namely the noncompositional nature of the meaning, they pose a problem when developing various linguistic models for syntactic annotation, grammar modeling, automatic text processing, information extraction, and other practical linguistic tasks.
The research presented in this paper is done in the framework of the large project aimed at large-scale description of the system, the structure and functioning of stable multiword units in everyday Russian speech, which has the following objectives: 1) development of a typology and theoretical description of stable
multiword units; 2) building a database of these units and describing the quantitative characteristics of their use obtained on natural speech material; and 3) the compilation of the materials for the Dictionary of stable multiword units of everyday Russian speech. The paper describes the notion of stable multiword units,
reviews the most well-knowns approaches to their study, and proposes the methodology of their further research based on transcripts of everyday Russian speech.