The purpose of this paper is to analyse a corpus of 185 conference abstracts, written in either English or Spanish, to determine their internal structure and specific linguistic features. Four variables were taken into account (language, research area, moves and style) to establish similarities and differences between the texts, in an attempt to determine possible universal features of all abstracts. Results show that the abstracts analysed share some of the features, but that there are also differences depending on the language of the text and the area of research. These results seem to have pedagogical implications for the teaching of writing of abstracts as well as for reading these kinds of texts.