BNC AND COCA AS CORPUS RESOURCES OF THE VERBAL EMOTICON IN CONTEMPORARY ENGLISH: REPRESENTATIVENESS, SEARCH TOOLS, AND LIMITS OF INTERPRETATION
The article substantiates the methodological value of the British National Corpus and the Corpus of Contemporary American English for the study of the verbal emoticon in contemporary English. The verbal emoticon is defined as a functional-discursive category: a verbal unit or construction that names, frames, intensifies, evaluates, or performs emotion in a specific context. The category includes lexical nominations, predicative patterns, evaluative exclamatives, interjective formulas, metaphorical expressions, and phraseological units when they function as markers of affective stance or interpersonal alignment. The study argues that BNC and COCA should not be used as interchangeable sources. BNC provides documented British evidence and enables comparison between written and spoken material, while COCA offers a large American corpus with genre-sensitive access to frequency, KWIC, collocation, and diachronic distribution. The article develops a corpus-based procedure that combines inventory building, concordance reading, collocational and constructional analysis, genre comparison, and contextual annotation. It also defines the interpretive limits of corpus evidence: frequency measures textual representation rather than psychological occurrence; collocation indicates recurrent co-selection but does not explain meaning without context; and corpus architecture determines the scope of possible conclusions. The final section outlines the relevance of corpus-grounded semantic profiling for large language models, especially for emotion recognition, stance detection, and affect-sensitive language generation.