Asosiy kontentga oʻtish
Maqola

Frequency, Dispersion, and Text Heterogeneity in English Literary Collocation Candidates: A Quantitative-Linguistic Study of Eleven Project Gutenberg Texts

Haydar HamdamovPhD in Philology, Acting Associate Professor, Department of Philology, University of Information Technologies and Management, Qarshi, Uzbekistan
2026
Ochiq kirishYashil

Litsenziya: cc-by

ABI

Annotatsiya

Corpus frequency is indispensable in collocation research, but it can become misleading when many occurrences come from one author, one character or one repeated narrative situation. This article examines whether text dispersion can distinguish broadly attested English lexical combinations from work-specific clusters in a literary corpus. The empirical material comprises eleven Project Gutenberg texts published between 1813 and 1912. After the Gutenberg licence blocks and paratextual headings were removed, the corpus contained 1,747,615 tokens distributed across works of markedly different lengths. A transparent rule-based procedure generated adjacent lexical bigram candidates without lemmatization. It produced 245,983 candidate instances and 171,843 surface-form types; 1,320 types occurred at least ten times. For these repeated candidates, raw frequency was compared with text range, normalized deviation of proportions (DPnorm) and a dispersion-weighted frequency defined as . Frequency correlated positively but only moderately with text range (Spearman's , ) and negatively with concentration (, ). Eleven of the thirty most frequent candidates occurred in only one work, whereas none of the top thirty dispersion-weighted candidates was confined to one text. Work-specific sequences such as said Dorothea, said Anne and Green Gables moved sharply downward, while good deal, next day, take care, much obliged and low voice gained or retained prominence because they occurred across several independent works. Contextual analysis nevertheless showed that wide dispersion does not by itself prove collocational status: temporal frames such as one day and social descriptions such as young man remain less lexically selective than prototypical collocations. The study therefore proposes that frequency, dispersion and lexical-semantic review should be treated as complementary layers of evidence in literary corpus analysis and bilingual lexicography. In quantitative-linguistic terms, the contribution is a source-sensitive refinement of frequency-rank interpretation for a heterogeneous literary corpus rather than a formal test of Zipf's law.

Hali tarjima qilinmagan

Identifikatorlar

Iqtiboslar va manbalar

0 ta iqtibos0 ta foydalanilgan manba