a corpus based approach to creativity in non-professional subtitles

Subtitlers' translation strategies and any possible traces of creativity in the produced subtitles. we designed a corpus based study with a population delimited to 269 comedy movies along with their persian and english captions converted to plain texts later on, which count up to 6012457 words overall. to assess the predetermined factor of creativity, 13654 words were detected by antconc as hapax legomena with an occurrence frequency of one. the list went through some revisions regarding an omission of the words with spelling errors, proper nouns, and the ones which could be supplied with a pre-defined dictionary meaning, which had finally provided this research with a body of 4139 words, among which 354 were randomly selected by a random generator software. our results demonstrated the most and the least frequent strategies occupied by subtitlers. transference and deletion took the largest shares of this corpus by 30% and 28% respectively, while literal translation and enlarging a scene accounted for merely 1% of all the elements captured.