Introduction Starting in the 1940s American linguists made use of the work of Nida (1946) and Pike (1947) to guide them in their description of unwritten Amerindian languages. Voegelin (1954), however, was the first to establish a formal methodology for the collection, transcription and translation of data, which he called Multiple stage translation and which rendered systematic the relationship between transcription and translation. The transcription method, based on articulatory phonetics, established by the Voegelins (Voegelin & Voegelin 1959) is cited to this day in field manuals (Samarin 1967, Vaux & Cooper 2003). Voegelin's intermediary language, a combination of morphemic glosses and mathematical operations, also designed for computational use, did not however become a standard among linguists. It is only in 1982 that the glossing tier (also known as literal translation or morpheme by morpheme translation) was reworked, with Lehmann proposing a system for the alignment of transcription and glossing tiers, with standardized practices and grammatical gloss abbreviations. The result of Lehmann's proposal, which evolved into the Leipzig Glossing Rules (Bickel et al 2004) and other variations, is still the current standard among field linguists as well as typologists presenting linguistic data. It must be noted, however, that, while the transcription and morphemic levels can be considered to be governed by standards, such systematized practices do not cover the translation level nor the relationship it bears to the glossing tier, resulting in grammatical analysis being interwoven, in many cases, into the translation tier. In this article, we explore various phases of the relationship between transcription and translation in linguistics and typology. In the first part of the article, we describe the processes set up by American structuralists, most notably Voegelin and his students, for the transcription and translation of data on the Amerindian languages they were describing; in the second part, we attempt to trace the Voegelin legacy to the present day, in looking at how a few field manuals present transcription, glossing and translation, and what efforts there are at systematizing data annotation across these three tiers. The questions that arise from the work of descriptive linguists involved in transcription and translation from the 1950s until the present include the following: