Maurice Varney (originally entitled ‘Forensic linguistics’ (1997). Reprinted from English Today 52, 13.4: 42–4, 46–7)
In recent years, linguists (and in particular phoneticians) have been greatly in demand to assist in crime detection. Maurice Varney discusses the reasons for the increasing importance of forensic linguistics (as this branch of the science is called) and points out a few of the difficulties sometimes encountered. In the original article, Varney discussed three branches, namely handwriting; phonetics and phonology; discourse analysis, but in this extract we have selected only the second of these.
The term forensic linguistics, by analogy with forensic science and forensic medicine, where scientists and pathologists use their skills in connection with problems involved in both criminal and civil law cases, simply means that linguists act as expert witnesses in legal cases where some aspect of written and/or spoken language seems to be significant. Forensic linguistics, though its practitioners are still a small and select group of experts, now stands equally as an accepted branch of linguistics, alongside psycholinguistics, sociolinguistics, etc., and also alongside its older partners forensic medicine, forensic dentistry, forensic chemistry, etc. The speed at which forensic linguistics has been accepted by lawyers, judges and juries is somewhat surprising given that at least one of its branches, discourse analysis, is often rather subjective. Perhaps it is because the application of forensic linguistic techniques has led to swift and apparently reliable solutions to legal questions that the subject has acquired a respectable legitimacy which other branches of linguistics took decades to achieve.
There are three branches: handwriting; phonetics and phonology; and discourse analysis. […]
The expert in pronunciation and other features of speech appeared in literature and folklore, long before the notion of forensic phonetics and phonology developed. George Bernard Shaw, in 1912, in his play Pygmalion (later even more famous as the musical My Fair Lady) satirises Professor Henry Higgins, an expert teacher and student of speech. Higgins was almost certainly based on the phonetician Daniel Jones, who established Received Pronunciation as the English pronunciation standard for use by phoneticians and phonologists.
Jones and his immediate followers were primarily concerned with articulatory phonetics, describing and cataloguing how human speech is produced. Today, the main interest of most career phoneticians is acoustic phonetics, studying and interpreting the physics features of the sounds that human beings emit. Spectrographs and other equipment are used to produce visual analysis of speech to a depth of great sensitivity.
The notion of the ‘voiceprint’ has been put forward, analogous to that of ‘fingerprints’. There is considerable evidence to suggest that no two people, living or dead, will ever have the same acoustic features of voice and that, no matter how well a person disguises the voice superficially, sensitive equipment will be able to show the base features of the original voice, and phoneticians will be able to interpret these.
Voiceprint analysis is used in a great range of legal cases, including blackmail, kidnapping, nuisance calls, confessions, telephone bomb threats, conspiracy, and hoaxes. Of course, the spoken message needs to be recorded before it can be examined and analysed, but this is often easy to do. Voiceprint comparison can only be done if there is a bank of recorded voices similar to fingerprint records. Such banks are being built up from cases where the criminals are apprehended and, as many perpetrators of crimes involving spoken messages reoffend, acoustic features of voices are recognised in the same way as fingerprints. An electronic programme has now been developed called Forensic Speaker Identification (SID) which conducts a ‘voice line-up’ identity parade and acoustically compares the voice patterns of suspects with that of the recorded message.
The Yorkshire Ripper case in northern England in the 1970s and 1980s involved phonetic and phonological analysis in a dramatic way. After a series of murders of women throughout Yorkshire, letters were sent to the detective in charge of the case, mocking him and his investigation and purporting to be from the Ripper himself. The letters were followed by a tape-recording, and it was confirmed that the handwriting of the letters and that on the parcel containing the tape were the same.
The detective in charge was convinced that the tape was genuine, and he called in specialists from the University of Leeds, the dialectologist Stanley Ellis and the phonetician J. Windsor Lewis. They successfully and accurately identified the voice and speech patterns as coming from County Durham, farther north than Yorkshire. A village was pinpointed and the police investigation resources were turned to this area, every man who lived there being interviewed.
Ellis and Windsor Lewis were right in identifying the geographical features of the voice on the tape, but the tape was a hoax. The Ripper killed again and, when he was eventually caught, he was found to come from South Yorkshire. The sender of the hoax letters and tape was never found. The detectives and forensic linguists in this case were vilified and accused of making mistakes which cost more lives and millions of pounds, but the criticism is unjust. In the circumstances of the time, the linguistic evidence appeared genuine and the experts followed the only road open to them.1
Apart from recognition of regional features and voiceprint acoustic features, forensic phoneticians are often called on to make educated guesses about the background, age, education, etc., of the people whose voices they have on tape. This is a tricky process involving much subjectivity, but the track record has been good, and experience and case studies enable practitioners to deduce much from an anonymous voice. Linguistic background, possible profession, and even age can be assessed and often checked by parallel analysis of the vocabulary and sentence structure used.
All phonological features are considered by the forensic phonetician: pronunciation, intonation, stress, pace, etc. In one case, a New Yorker living in California was arrested and indicted for making threatening calls to Pan American Airlines. His defence attorneys called in a voiceprint expert and the dialect expert William Labov, and together they proved the intonation of the anonymous caller was Bostonian and not New York. The suspect was acquitted (Labov 1988: 170–81).
1An interesting follow-up to the Yorkshire Ripper case took place in 2005, nearly thirty years after the murderer’s first attack, when the hoaxer, John Humble, was eventually tracked down through DNA analysis of a drop of saliva he had left on a letter. It turned out that the phoneticians Ellis and Windsor Lewis had indeed traced him to within a very short distance of his home. In 2006, Humble confessed to his guilt, and was sent to prison for eight years.