| Abstract [eng] |
The current study aims to establish, if authors idiolect can be recognized in electronic mails language and to determine the features of lexis and graphics, which can be linked to idiolect. The data has been derived from a corpus of 65,000 words consisting of electronic letters written in Lithuanian by six persons. The WordSmith Tools software was used to generate frequency lists of six subcorpora, representing each person’s language. By using the contrastive method the frequency data of six persons language were compared. The lexis and graphics elements, which were used by one person more often or more rarely than by others and were not determined by the topic, were linked to authors idiolect. As a result of the analysis the classification of lexical and graphical elements is given, which can help recognizing idiolect. The study shows that on a lexical level the main differences between idiolects are in the usage of the modality and stance expressing words, and also the words and abbreviations, which are differently chosen from possible variants. On a graphical level idiolects can be recognized from punctuation marks, emoticons and graphic symbols, used at a different frequency. Based on research results the recommendations for authorship attribution examinations are given. |
