Text Variants and First Person Domain in Author Identification: Hermeneutic versus Computerized Methods

Since a language variety contains shared variants, and since a complete correlation between author and linguistic features is rarely acquired, it is suggested that linguistic features which fall outside the correlational agreement in a variety belong to the author's First Person Domain (FPD). Advances in computerized vocabulary profiling and readability provide useful characterization of features found in Academic English (AE), but they cannot capture the full range of linguistic features in a text. A corpus of about 38 extracts and texts (111.000 words) from local and international authors is analyzed to determine interpersonal and intrapersonal variations. The results show that language variation determines the features of FPD which are crucial for author identification and that computational methods are not adequately sensitive to insure a hundred percent author identification. Therefore, epistemological author identity profile (AIP) is suggested to plot alleged texts against the socio-physical and epistemological parameters of alleging authors.