The purpose of this paper is to present the findings of a corpus-based contrastive study on a semantic function — restriction — and their applicability to translation. A number of restrictive elements in the two languages involved, English and Spanish, have been selected as search words and their behaviours have been analyzed on the basis of their textual environments. Apart from an overall description of meanings and uses, the article aims at verifying a linguistic hunch: cognates and also word-forms coming from the same lexeme were intuitively considered to differ more than might be expected, which has altogether been proved by the data found. Finally, a series of correspondences are suggested as translational options into Spanish for the most frequent uses and co-occurrences of the English elements.