Criteria of Semantic Decomposability of Idioms

Although idioms are typically understood as fixed expressions whose individual components have completely lost their lexical meaning, many idioms in various languages contain elements that retain a certain degree of semantic autonomy. We refer to such expressions as semantically decomposable idioms . Since decomposable idioms have a significantly greater potential for variation than non-decomposable ones, it is clear that the phenomenon of decomposability or analyzability is not only of theoretical interest but also has practical significance. The category of semantic decomposability proves particularly useful in the development of modern linguistic resources. If a project involves designing algorithms for the automatic detection of idioms in a corpus, accounting for the variation potential of each idiom is a necessary prerequisite for successful implementation. This paper focuses on operational criteria to distinguish decomposable idioms from non-decomposable ones. A highly reliable indicator of non-decomposability is the lack of parallelism between an idiom’s syntactic structure and its syntactic function. For example, if an idiom is structurally an imperative sentence but functionally an adverbial or predicative, it is clearly non-decomposable. Three criteria are proposed as indicators of decomposability: (1) the presence of idiom correlates differing in aspectual characteristics - conversives, causatives, inceptives, etc; (2) the presence of paraphrasing elements in the context; (3) the referential use of noun phrases within the idiom, often emphasized by demonstrative, interrogative, relative, or indefinite pronouns, quantifiers, etc. Although none of these criteria guarantees absolute reliability, their combined use significantly improves the accuracy of the analysis.