Arranging a voice lineup in a foreign language

Abstract

Speaker identification by lay witnesses is an area that is still relatively unexplored. The volume of empirical research in this field contrasts sharply and unfavourably with the wealth of work reported in the literature on human identification in the visual domain. Nor is it clear that the findings of eyewitness research are sufficiently well known in forensic phonetic circles, in spite of their patent relevance to earwitness identification. Many of the issues over which consensus has been reached among those working in the field of eyewitness identification (Wells et al. 1998) continue to be unresolved in the forensic speaker identification community. Attempts to take matters further by taking on board the unquestioned insights gained in the visual field have so far met with a lukewarm reaction at best. After a promising start (Hollien et al. 1995, Broeders 1996, Hollien 1996), little progress has been made in recent years within either IAFP (the International Association for Forensic Phonetics) or the ENFSI (European Network of Forensic Science Institutes) Expert Working Group for Forensic Speech and Audio Analysis in formulating generally accepted procedures for the conduct of voice line-ups. At the same time, debate seems to have come to a virtual standstill, as witness the lack of response to the most recently proposed set of guidelines (Broeders and Van Amelsvoort 1999). Yet, ever since McGehee’s (1937) work following the Lindbergh trial, the need to develop proper procedures in the conduct of earwitness identification has been abundantly clear.