Posted on 17 May 2012
The long-running series of documentary films, which has followed the lives of fourteen British people since 1964, is providing an ideal database for Richard Rhodes from the Department of Language and Linguistic Science.
The Seven Up series provides an amazing dataset as it is possible to examine the same people’s speech across several decades
The children featured in the original episodes were selected to represent the range of socio-economic backgrounds in Britain at the time. Programme-makers have returned to interview the people at seven year intervals, with the latest series, 56 Up, currently being broadcast on ITV.
Richard Rhodes’ research involves assessing the reliability and strength of forensic speech evidence when there has been a time lapse since the original ‘criminal’ recording was made. His investigations centre on whether it is possible to analyse speech evidence in criminal cases spanning decades in similar ways to forensic DNA evidence.
As part of the research, he has been analysing the voices of eight people aged from 21 to 49, from the Seven Up series.
Richard Rhodes, a final-year PhD student, said: “The Seven Up series provides an amazing dataset as it is possible to examine the same people’s speech across several decades. The programme also gives biographical details and an insight into people’s life changes, such as social or geographical mobility, and health and lifestyle patterns, providing an ideal opportunity to discuss the potential impact of these changes on a speaker’s voice and the language they use.
“My research aims to shed light on the likely changes in voices encountered in forensic cases where there has been a long delay between evidential recordings. Some changes are easier to predict than others. For example, changes to the vocal apparatus due to ageing are fairly predictable, whereas changes to the voice due to different types of mobility are not so straightforward. Similarly, some speakers change their speech to conform to mainstream accents; this was recently shown to happen in the Queen’s own English.”
DNA evidence is presented as a numerical probability of a match and Richard Rhodes is investigating the reliability of applying a similar statistical model for speech evidence which spans several years or decades.
Richard Rhodes said: “Some features commonly used in forensic voice or speaker analysis remain more stable than others. Some are shown to vary in lots of different contexts for different reasons. This is very different to forensic DNA analysis where a person’s DNA is almost totally stable. It is important to assess what effects changes with age have on this approach.”
The most high-profile case in Britain where long-term non-contemporaneous voice evidence has been used was in the prosecution of the Yorkshire Ripper hoaxer, John Samuel Humble. In 2006 he was convicted for perverting the course of justice after sending hoax messages, including a cassette voice recording, to police in the 1970s.