This paper investigates two of the most widely analysed universals in translation research, namely simplification and explicitation. It examines the oral production of bilingual children with different language pairs as available in the CHILDES project (MacWhinney 2000) (i.e. the FerFuLice, Ticio, Deuchar, Vila, Genesee and Pérez-Bazán corpora) as well as in other compilation forms (i.e. Ronjat 1913; Leopold 1939–1949; Swain 1972; Lanza 1988, 1997, 2001; Cossato 2008). It addresess two main issues: whether instances of simplification and explicitation appear in the production of non-instructed interpreters and, if so, how their occurrence relates to the type of data (i.e. spontaneous or experimental) and the language pair involved. The results show that children acquiring two first languages often translate and use simplification and explicitation at varying degrees irrespective of the language pair. It is concluded that the analysis of acquisition data can contribute to shed light on the nature of these translation universals.