This paper presents the findings of an experiment in which 20 Greek Deaf students produced written texts under two different conditions of language input: (1) a translation from a videotaped story in Greek sign language, and (2) a direct composition produced from a picture story--a neutral non-linguistic input. Placing Deaf writing within a bilingual frame, the effect of language input on the quality of written texts was explored, i.e. whether or not the use of sign language facilitates the teaching of written language. In this paper, similarities and differences between Deaf writers and hearing bilingual writers are explored in terms of current theoretical perspectives on bilingual learners: Deaf writing, similar to bilingual writing, is the result of an interaction between two languages, although in the case of Deaf writing, the languages are an unrelated sign language and written language; the role of first language in teaching; and whether sign language qualifies as L1 for Deaf students. This discussion is complemented by the quantitative results in the study, which showed that the use of a language (in the form of translation) in second language writing may facilitate certain features, such as the organisation of text, but not others, such as the grammar of text. The implications of the findings for bilingual education and Deaf education are discussed.