Files
Abstract
This dissertation investigated the patterns of language use in mania using corpus linguistic methodology. The data were drawn from a collection of journals composed over a twenty eight year period by an individual diagnosed with mania, yielding a Master Corpus of 242,589 words which was subjected to an array of analyses, including comparison to patterns of typical language use as represented in the Freiberg-Brown Corpus of American English (FROWN). The Master Corpus was also compared to a reference corpus composed of journals produced by other writers for a within-genre analysis. Multiple intra-individual analyses were conducted in order to test hypotheses concerning differences in language behavior during medicated versus unmedicated conditions, manic versus non-manic writing, and more versus less severe episodes of mania. A final analysis investigated changes in patterns of language use relative to variation in content and frequency of entries across early, middle, and late phases of a single manic episode. The data were analyzed with WordSmith Tools to determine patterns of word frequency and collocation, and for the derivation of keyness statistics between corpora. Results confirm all hypotheses, which postulated there would be significant differences between language as observed during manic episodes and the respective reference corpora, and likewise intra-individual differences in all the previously described experimental conditions. Corpus analysis showed sensitivity in detecting language behavior that correlated with the diagnosis of mania, and also the language effects of treatment with medication. According to the DSM-IV, clinical diagnosis of mania and other mental illnesses relies substantially upon the presence of anomalous patterns of language use in a constellation of psychological, emotional and behavioral manifestations. Therefore, utility of corpus linguistic methodology in clinical applications was proposed, and a model for conceptualizing the constellation of linguistic criteria for diagnosing mania was delineated, in an effort to clarify the terminology used for identification of linguistic behaviors in mania. Areas of future research involving enhanced diagnostic validity and reliability were proffered, as were other directions for using corpus linguistic analysis in the diagnosis and monitoring of mental illness.