Transcription Methodology

The corpus is transcribed orthographically and, for the most part, without punctuation. A full-stop is used in the transcription where there is any need for clarity about the position of the end of a clause (this need can arise for syntactic or phonetic/supra-segmental reasons). In the default case of no full-stop between clauses, adverbials appearing to the left of verb phrases can be assumed to be initial adverbials occurring pre-verbally (e.g. in a sequence such as 'c'est ce qu'elle a compris le lendemain matin elle a décidé de partir', 'le lendemain' is at the head of the second clause unless otherwise indicated by a full-stop after 'matin') . A question mark is used to indicate a question, primarily because questions in oral French often take the same syntactic form as declaratives, the distinction being intonational rather than syntactic. The exceptions to orthographic transcription are proper nouns and names, where initial capital letters are used, as well as a small number of very specific contexts where a non-orthographic transcription is necessary to reflect the context in which a negative construction is found, i.e. 't'as pas' for 'tu as pas', 'j'sais pas' for 'je sais pas', 'y a pas' where 'il' is not pronounced. The transcription is literal rather than normative, reflecting the nature of oral discourse, and therefore includes repetitions, ellipses, false starts etc. Errors relative to standard French appear where they occur (e.g. 'il s'est mit à…'). Truncated words are transcribed by giving an approximation to the intial sounds followed by '-', e.g. 'str-'. In the case of proper names, an attempt was made to establish attested spellings and the spelling adopted was the one that best reflected the storyteller's pronunciation. In some cases it was necessary to make a small modification to an attested spelling in order to convey the pronunciation in French. Where it is not absolutely clear whether an item is articulated or not, the item is given in brackets; this is particularly common with 'et'. Where there are two lexical or grammatical possibilities in a given context, either (i) the version considered most likely by the researcher in the light of the pronunciation and context appears in the transcription with the other possibility in brackets after a forward slash 'et (/est)', or (ii) in the case of singular/plural agreement, the second possibility is given in brackets e.g. 'il(s) pense(nt)'. Where a semi-lexical form is used, it is transcribed where possible, e.g. 'euh', 'mmm'. Less clear forms are given as part of a vocal 'seg' with an approximation to spelling if possible. Unclear sections are marked up using the 'unclear' TEI function and where possible, a reason is given and/or a best guess at content. Sung, incanted or whispered passages within the narrative are normally transcribed and marked up using the TEI 'rend' function. Songs at the beginning or end of narratives are not normally transcribed if they do not form part of the narrative itself; these and non-vocal events such as instrumental music are marked up using the TEI 'event' function. Pauses are not normally indicated because they are multiple and are normally short but where there is an exceptionally long pause, the TEI 'pause' tag is used.