This paper reports on a study in which subjects were asked to distinguish translations from originally produced (non-translated) texts. The aim was to identify the linguistic features shared by texts assumed to be translations, as well as those shared by texts assumed to be originally produced. The results show (i) that translations were not readily identifiable, and (ii) that the feature that seemed to guide the subjects’ decisions was the frequency vs. scarcity of target language specific (unique) items in the texts: their frequency led subjects to assume &x2015; correctly or incorrectly &x2015; that a text was original rather than translated. It is concluded that the unique items in non-translations vs. translations deserve further research in respect of their frequency and the impressions they make on readers.