Theoretical and methodological foundations for corpus-based analysis of formulaic expression units

Abstract EN

The description of the formulaic status of the linguistic behaviuor of the formulaic expression units (FEUS) arises from a set of universal principles underlying the mental organization and representation of language and conventionalized patterns of language use.

the nature and gradeability of formulaicity can be clued-up by the statistical distribution of such units at all levels of linguistic analysis including the syntactic, semantic, and discourse levels.

In addition to the existing discursive approach, the FEUs' formulaic status must be quantitively approached and verified by a corpus-based statistical analysis of distributional frequency.

appropriate theoretical frameworks, including distributional semantics and cognitive semantics, should undoubtedly unveil formalized semantic and cognitive parameters which could better fit for the distribution frequency statistical analysis of the linguistic data.

the corpus-based method allows retrieving sets of expression units to determine their formulaic status based on the frequency of occurrences in documents/domains collection.

in this research, the model of corpus-based statistical analysis of distribution, proposed here, adopts two query-based information retrieval methods : (i) -gram corpus maximum frequent sequences (n-gram-MFS) for the representation for weighing FEUs' formulaic status per n-gram corpus and (ii) maximum frequent sequences (D-MFS) for weighing FEUS' formulaic status per document/domain.

the use of the proposed model offers a syetematic verification tool to weigh and evaluate formulaicity status of FEUs.