This paper reports on a set of three eye-tracking-informed process-oriented studies aimed to tap into the (re)translation task in three different language pairs (Spanish-Portuguese, French-Portuguese, and English-Portuguese). Such studies were built on a task layout demanding potential attention to three textual inputs, namely a source text and two previously published translations thereof, with a view to rendering a fourth text, the new translation. Results point to greater transitions from target text to source text, and vice-versa, as the most frequent flow of visual attention, with occasional shifts to previous translations for solving understanding or production problems in rendering the target text. The results of such studies and the pioneering description of the (re)translation process under experimental conditions may contribute to translation process research by providing some evidence on how participants cope with several inputs in their rendition of a target text.