The original language of Tobit was argued over for centuries. The book was known in Greek in two recensions of markedly different length, a shorter form in the Vatican and Alexandrian codices, a longer and more circumstantial one in Codex Sinaiticus, and in Latin, Syriac, and later Hebrew and Aramaic versions of uncertain descent. Jerome, preparing the Vulgate, said he worked at speed from an Aramaic text through an interpreter, and his Latin diverges from the Greek at many points, which is why older Catholic Bibles carry details other versions lack.
Qumran settled the central question. Cave 4 produced fragments of five manuscripts of Tobit: four in Aramaic and one in Hebrew. That establishes beyond argument that the book existed in a Semitic language in the Second Temple period rather than being composed in Greek, and it tells against the older view that the Semitic versions were later retranslations. Most scholars now take Aramaic as the original with the Hebrew copy a translation of it, though the reverse has been defended.
The fragments also arbitrate between the Greek recensions: where comparison is possible they side with the longer Sinaiticus text, and modern translations have followed, so a reader comparing an older English version with a recent one will find the recent one noticeably fuller. Dating rests on the same evidence together with the book's own silences. The palaeography of the Qumran copies sets a limit, and the book shows no knowledge of the Maccabean crisis while taking a close interest in the eastern diaspora, which points to composition in the third or early second century BCE, more probably in the east than in Judea.