Discussed in HoC in 2016:
<a class="break-all" href="https://web.archive.org/web/20200807215406/hansard.parliament.uk/commons/2016-04-20/debates/16042035000001/ForensicLinguistics(Standards" rel="nofollow" target="_blank">https://web.archive.org/web/20200807215406/hansard.parliament.uk/commons/2016-04-20/debates/16042035000001/ForensicLinguistics(Standards
And actually it's really interesting, so posting the whole move:
Roger Mullin (Kirkcaldy and Cowdenbeath) (SNP)
I beg to move,
'That leave be given to bring in a Bill to place a duty on the forensic science regulator to establish a code of practice and conduct for the providers and practitioners of forensic linguistics in the criminal justice system; to make provision about the required scientific quality standards for the discipline; and for connected purposes.
Our children and young people face an enormous threat from being groomed by radical extremists and paedophiles, facilitated by the internet, social media and mobile technology. The Bill is therefore about the protection of vulnerable people, and it is about the monitoring and analysis of communications between people about whom we need to be really concerned: people who plot and scheme to do others harm; people such as paedophiles and extremists; and people who use modern technology as a tool in their evil business.'
Last October, I led a Westminster Hall debate on the use of children as suicide bombers. We know that many of the techniques used in recruiting and grooming such children are the same as those used by paedophiles. We also know that there is software available that will identify the messages and language of groomers, and that, using a variety of tools, security agencies can match those to the voice and language patterns of known individuals. Forensic linguistics is a complicated and relatively new field. Linguistic evidence can involve science, social science and interpretation, and forensic linguistic analyses require a complex set of knowledge and skills. Presently, however, anyone—including you or me, Mr Speaker—can proclaim themselves an expert in forensic linguistics. Consequently, there is a considerable danger of substandard analysis being offered in a court of law.
We need a standardised qualification for analysts and a standardised set of techniques to give the courts confidence that such evidence can be accepted as more than just interesting background. The Bill does not represent sophisticated legislation, as compiling a statutory register would be relatively straightforward. The register called for by the Bill would not need its own regulator, as one already exists: the forensic science regulator. She is already working on including speech and audio analysis as a recognised speciality area, but as textual linguistic analysis draws on interpretative as well as scientific methods, it falls outside her current remit.
I would also draw attention to current codes of practice and conduct for forensic science providers and practitioners, and more generally for expert witnesses in the criminal justice system, that could be adapted to include the practice of forensic linguistics. For setting the accredited qualifications, there are academic institutions with evident authority in this area, such as the centre for forensic linguistics at Aston University. I personally thank the director of the centre, Professor Tim Grant, for his help in developing the Bill. I am also grateful for encouragement from the president of the Chartered Society of Forensic Sciences and the director of forensic services in Scotland, Mr Tom Nelson. The standard of specialist witnesses and forensic scientists themselves is inherently protean—I know of some people who call themselves forensic scientists, but cannot tell the difference between the sensitivity of a test and the specificity of a test, let alone calculate its predictive value.
I have already said that speech and audio analysis and textual analysis are two different things. The problem for textual forensic linguistics is that many aspects of the work—the determination of meanings in messages, profiling the background of a writer and so on—are a long way from the laboratory-based paradigm. The closest we get to laboratory-based science is in comparative authorship analysis, for which methods are published and tested. The diversity of questions that forensic linguists address and the approaches that they take to those questions means that the forensic science regulator does not cover their work, so there is no way for high-quality practitioners to be identified and used and low-quality practitioners avoided.
There is a need for a mechanism to recognise what should count as quality work in textual forensic linguistics. That could be a register of individuals or methods, or both. The obvious person to hold that would be the forensic science regulator, but that would definitely represent an extension of her current role, hence the need for the Bill.
Where is the proof, however, that forensic linguistic analysis can work? In those cases in which forensic linguistic evidence has been allowed in court, it has proved particularly valuable. For example, it was used in the appeals of Derek Bentley and the Birmingham Six. In many instances across the UK, it has been used to determine the authorship of SMS text messages in murder cases. It has been used to extract the meaning of coded texts and slang terms used in internet chatrooms, often involving conspiracies to murder and child sex abuse conversations. Good forensic linguistic evidence has withstood appeal, yet this excellent work could be undermined due to substandard analysis by poorly qualified and unqualified practitioners.
Although it has strong roots in the UK, textual forensic linguistic evidence is increasingly accepted internationally. Examples of its use include successful appeals against murder convictions in Australia and cases of disputed wills in South Africa. In 1996 in the United States, textual forensic linguistic analysis was used to identify the writer of the Unabomber’s manifesto as Ted Kaczynski, and he was subsequently convicted of running a bombing campaign across the country.
In the United Kingdom, too, textual forensic linguistics has been used in investigations of serious counter-terrorism cases. In 2004, for example, Dhiren Barot was arrested in London and charged on the basis of linguistic evidence linking him to the writing of a conspiracy document. He later admitted to plotting to bomb the New York stock exchange, the International Monetary Fund headquarters and the World Bank, among other targets.
The United Kingdom’s forensic science regulator role was created in 2007 by the hon. Member for Hackney South and Shoreditch (Meg Hillier). It is good that some progress has been made, but on this issue, Mr Speaker, it is time to put the regulator to work. The Bill would enable the statutory agencies to use information and evidence that they gather through the medium of forensic linguistics to protect more children from predatory adults, and to protect the British public from the likelihood of events such as those that happened recently in Brussels, Paris, Istanbul, Kabul and Pakistan. I commend the Bill to the House.'