Kolokacijų ir frazeologizmų atpažinimo kriterijai
Towards the Criteria for Identification of Idioms and Collocations

Summary/Abstract: The paper analyses two types of multi-word expressions: collocations and idioms. It provides definitions of these two types of lexical items and reviews their understanding in works of Lithuanian and foreign scholars. A problem of identifying multi-word expressions arose when carrying out a research project Automatic Identification of Lithuanian Multi-word Expressions (PASTOVU). When the analysis of the corpus compiled from delfi.lt news portal texts started, it turned out that there is a need for more precise criteria to identify collocations and idioms in Lithuanian data. Therefore, this paper aims at describing clear criteria that would allow linguists to identify these particular multi-word expressions in texts as objectively as possible. It seems that collocations and idioms have three main features: conventionality, semantic non-transparency, and fixed form. All these features and their strength can be evaluated applying various tests designed based on identification criteria. The paper explains the proposed tests and gives examples for each one of them in order to practically reveal the similarities and differences between collocations and idioms. From the data described, it becomes clear that the classes of collocations and idioms are not homogeneous: both of them have typical and less typical cases. We suggest conventionality and semantic non-transparency as two main features, which could help to separate collocations from idioms by applying particular criteria and tests.

• Details
• Contents