We examine the possible impact of frequency differences between a construction in L1 and its equivalent in L2 on translations. Our case is that of existential there in English and existential il y a in French. Using corpus evidence, we first confirm previous claims that existential there is used more freely in English than existential il y a is in French. Drawing on extensive counts conducted in available corpora and self-compiled samples of translated English and French, intra-language comparisons of translated and non-translated language use show that existential there is under-represented in English translated from French while existential il y a is over-represented in French translated from English. It is suggested that source-language interference is responsible for these differences. In addition, counts of existentials in individual novels and their translations show that inter-language frequency shifts systematically occur in the direction of target-language norms, most clearly so for translations into French, which suggests that the observed usage constraint on il y a still applies to a noticeable extent in translated French. Methodologically, we argue the need for a large corpus of translated French.