Available at

3.2 Data transcription 
In corpus-based interpreting studies, as in any corpus investigation that looks at 
spoken language, it is necessary to transcribe oral data in order to turn them into 
an analysable format. However, as Agar (1996: 153) says, “transcription is a 
chore” and transcribing interpreting data is not easy neither. Niemants (2012: 180) 
furthermore says that “interpreting research is not very helpful when it comes to 
determining an adequate transcription system”. When it comes to transcribing 
interpreting data, Niemants talks about three main issues: participants, languages, 
and timing. Any transcription involving more than one participant needs to the 
speaker of each utterance, the beginning and the end. The main problem about 
participants is then to segment utterances and link originals with their renditions, 
knowing that one original segment does not systematically correspond to one 
rendition segment and vice versa. Multilingualism is also a big issue for 
interpreting data as the transcriber must be able to input language-specific 
characters and specify the language used. There are also many phonetic and 
prosodic features to take into consideration: tempo, volume, duration, pitch, 
intonation, and accent, among others. Finally, timing is also a big issue as we 
have to pay attention to the beginning and the end of all utterances, and also to 
how they intertwine (Niemants, 2012). 
For the transcription of my interpreting data, I followed a set of guidelines, which 
are described in Table 2. These guidelines are used for the transcription and for 
the enlargement of EPTIC, and they serve those involved in the process of 
transcription and coding. 
Table 2 – EPTIC guidelines 
Feature 
Example 
Transcription convention 
Punctuation 
Based on the syntax and the 
intonation, no comas 
Truncated words 
propo 
propo- 
Pauses 
filled 
ehm 

---

Data and methodology 
 
page 42 
empty 
… 
Interrupted speech 
… 
Numbers 
Numbers 
Written out in full 
326 000 000 
326 millions 
326 000 000 000 
326 milliards 
30 % 
30 pour cent 
Years 
In numbers 
Fractions 
Written out in full 
Unclear bits 
# 
Language shift 
# 
Invented words 
endavure 
{endavure} 
Capital letters 
To refer to interinstitutional
stylistic guidelines and official 
websites such as: 
http://www.cnrtl.fr/definition/ 
Acronyms 
To refer to interinstitutional
stylistic guidelines and official 
websites 
Titles 
monsieur/madame/mademoiselle 
First, all the words that the students started uttering but did not complete (i.e. 
truncated words) are transcribed with a hyphen.
(29) Le don- don- le donim- le dénominateur commun est simplement ce cette 
inflexion cette intensité 
There are then two types of pauses: empty and filled pauses. An empty pause is 
transcribed using “…” (irrespective of its length) whereas a filled pause is 
transcribed using “ehm”. A filled pause is “a period of articulation of non-
propositional content and fitting a language specific convention” (Rose, 2017 
based on Mahl, 1956; Clark & Fox Tree, 2002). In English, “um” and “uh” are 
typical filled pauses. Then, the symbol # is used for unclear bits. Non-attested 
words in the target language are transcribed between curly braces (either totally 
invented words or English calques that do not exist in French). 

---

Data and methodology 
 
page 43 
(30) {combinations} 
(31) {synapsis} 
(32) {synesthie}