A Day in the Life of a Forensic Speech Scientist
It isn’t everyday that you meet a Forensic Speech Scientist. Yes, a forensic speech scientist.
Dominic Watt is a senior lecturer of Forensic Speech Science at the University of York in England. Dominic, who is one of the world’s leading professionals in his field, has been working on a project called the Accent and Identity on the Scottish/English Border (AISEB) project, which examined the difference in the English spoken on either side of the Scottish/English border and the link between personal identity and accent.
So, how does accent influence our first impressions of people? How does accent relate to national identity? Does this change along the border? Does the accent border follow the political border? These are just some of the things that Dominic and his incredible team explored.
In today’s blog post we’ll explore what forensic speech science is and the life of a forensic speech scientist. We’ll dive into the AISEB project to discuss the findings and implications of this research and the interesting link between national identity and accent.
What is Forensic Speech Science?
According to the University of York, forensic speech science is “the application of linguistics, phonetics and acoustics to legal investigations and proceedings”. Forensic speech science enables certain linguists and speech scientists to provide evidence in legal proceedings. For example, a forensic speech scientist may be used as an expert witness in a case to determine if two brands names sound too similar. It is a beautiful combination of law, speech technology, linguistics and phonetics; think linguistics meets CSI.
Introducing Dominic Watt and How He Became a Forensic Speech Scientist
After talking with Dominic for only a few minutes I could tell he loves what he does. Dominic had always been interested in language from a very young age. Like his parents, Dominic decided to study linguistics at university. He taught in the UK and Germany before coming back home to start a PhD in Phonetics. Then, Dominic was asked to be a part of a project that would change his career.
“I was passed a case by Peter French, who runs J P French Associates, a private lab across town, which is the UK’s largest and oldest speech and audio lab.”
His first case ended positively and began his career in forensic speech science.
“The next case I took on was a trademark dispute, so civil law. That was really interesting. They said ‘we have these two brands. We think the newer of the two contravenes UK trademark law and we want to argue that, so can you provide a bit of expert backup to support us in this suit?’”
Dominic was intrigued and agreed to take on the case.
“I did a survey among my students. The product in question was acne cream, so I thought well, actually, their age group is the target market. So if there was any potential for confusion in the mind of the consumer as a result of these suspiciously similar brand names for essentially the same product then this is the group that would get confused. This could potentially affect the profit margin of the company that produced the original brand.”
We discussed this case for a little while and in the end, the two brands were deemed too similar and the second brand was ordered to make a change. After about 20 more minutes, he finished nicely with:
“Anyway, that’s a quite roundabout way of saying how I got into forensic speech science.”
The Accent and Identity on the Scottish/English Border (AISEB) Project
One of the main things I wanted to talk to Dominic about was the AISEB Project. This project was primarily run by a group of forensic speech scientists from the University of York, which included Dominic Watt. The AISEB project looked at one of the greatest concentration of linguistic features in the entire English speaking world – the Scottish/English border, which you can see below as a red dotted line.
Along the border, the national identity of inhabitants of towns on either side is rather complex. Many people consider themselves undoubtedly English while others are certain they are Scottish, and there are many who lie somewhere in the middle. Language plays a central role in how these people mark their identities, but before the AISEB project it was unclear which features of pronunciation marked ‘Scottishness’ and ‘Englishness’ and whether people with these features identified themselves with the nationality that their accent and pronunciation suggested.
The project also examined how closely the political border followed the linguistic border, how an accent affects a person’s perception of someone they don’t know, and the difference between the east and western side of the border in terms of strength of an accent. What follows are the main highlights from the interview I had with Dominic about the AISEB project.
How did the AISEB project come about?
“Way before AISEB came into being I was interested in the way that the border functioned as a linguistic border as well as a political one. If you look at literature on the topic, dialectologists and various commentators, such as people who are interested in the Scottish language for example, have identified the border as a really significant linguistic boundary.”
“Before the AISEB project, I had an undergraduate student who was from Berwick-Upon-Tweed, which is known for all sorts of peculiar historical reasons.”
He went on to describe the unusual history of Berwick-Upon-Tweed (often shortened to Berwick), with the following points being my favorite historical peculiarities:
- Even though Berwick belongs to England now, it has changed hands 14 times between England and Scotland.
- Queen Elizabeth the 1st spent more money defending Berwick from the Scottish than she did on defending against the Spanish Armada.
- Although Berwick is in England, their football team plays in the Scottish league.
- They belong to the Church of Scotland, not the Church of England.
“All these peculiarities about the town made us think. We talked to some of the people who came from the town and they speak a really peculiar dialect there, which has been nicknamed MacGeordic – Mac being the Scottish part and Geordic being the Northumbrian part. To some people they sound Scottish and to other people they sound English. There are features of both Scottish and Northumbrian English in their language.”
The AISEB team decided to look into Berwick, along with several other towns along the border and examine how such stark linguistic differences could exist and how they influence the identity of the people from those towns.
What were some of your most fascinating findings?
“We found pretty strong correlations between people’s identities and the way that they talked. Now, that’s not a big surprise but we also wanted to link political opinion and speaker’s home countries, Scotland versus England, and find fine-grain speech patterns, which people just really aren’t aware of at the surface level.”
“We’re talking about the pronunciation of consonants which involve variation at the level of just a few milliseconds. You’ve got things that are more obvious, like, everyone can talk about dialect. But we are interested in things that people don’t know that they’re doing, the things they can’t discuss.”
Dominic explained their methodology of visiting each of the towns along the border and asking the inhabitants a series of questions. These questions were designed to encourage participants to demonstrate their accent and pronunciation of different vowels and consonants to discover these fine-grain speech patterns.
“Different pronunciation of “r” for example – there is a lot of variation in that consonant. We hypothesized that people who professed to feel more strongly Scottish would use more of what we call the Alveolar Tap R. I’ve been using quite a few as I’ve been talking. You know, if I say very very brown, you can hear it, it’s that very Scottish sounding r.”
He had a fantastic accent from my perspective, and although you can’t hear it, the R he is referring to is that typical Scottish R, which is very pronounced and distinctive.
“That R is common on both sides of the border, but we found that people who used the tap pronunciation more often were also the ones who felt more strongly affiliated to Scotland. You might not think that is a very surprising finding but no one has ever shown that before. We demonstrated that.”
While I was impressed at the project’s ability to demonstrate these “unsurprising things” as Dominic called them, it was even more fascinating to hear about the project’s more unpredictable findings. For example, the AISEB project not only looked at differences between the north and south side of the border, but also between towns along the border at the western and eastern ends.
On the North side of the border (in Scotland), they looked at Eyemouth and Gretna and on the South side of the border (In England) they examined Carlisle and Berwick-Upon-Tweed. Here you can see the Scottish/English border as a red dotted line with each of the areas examined highlighted in red.
Were there many differences in the accents of the people in the towns on the eastern side and western side of the border?
“We found that there was a big difference between east and west along the border. We had these 2 communities on the north side of the border – Eyemouth at the eastern end and Gretna at the western end. We found that people who lived in Gretna reported much more variation on how Scottish they felt than those in Eyemouth, despite being on the same side of the border.”
He described how people from Gretna would often get told that they “don’t sound Scottish” or that they couldn’t possibly be from Scotland because of their rather English-sounding accent, despite Gretna being located in Scotland. On the other hand, people from Eyemouth sounded very clearly Scottish.
Then we discussed the western side of the border more thoroughly, with Carlisle on the English side and Gretna on the Scottish side.
“In Carlisle, they are really solidly English sounding; their dialect is Umbrian, end of story. People there report feeling either English or British in a fairly straightforward and uncomplicated way. But if you go slightly north of there, just 9 miles and across the border to Gretna you find a really different picture. People there technically live in Scotland, even if they don’t feel very Scottish. That sort of ambivalence is reflected in their speech patterns. ”
One of your aims was to discover how closely the political border follows the linguistic border, what did you find out? Was it closer or further away than you expected?
“It was quite consolidated at the eastern end and then at the western end it sort of fans out more.”
He discussed two small towns on the border and the intense difference between the accents in each town. The towns are so close that “they are separated literally by about 50 yards by the river”, Dominic said. Everyone said to Dominic’s team that if you were on one side of the river, the accent was very Umbrian but on the other side it was clearly Scottish.
“I was always a bit skeptical about that, I always thought these differences are exaggerated. But the political border appears to firmly represent the linguistic border there. They are very close together and yet their speech patterns are very remarkably different.”
He pointed out that this is particularly interesting when you realize that people are free to move between England and Scotland because they are both part of the United Kingdom. This means there is no legal reason why the Scottish people have to stay on one side and the English on the other, but they choose to do so. The AISEB team was intrigued by this and wanted to find out why this happened.
“What is it that makes the Scots so resistant to sound changes that come in from the south?” Dominic said. “What is it about attitudes or the world view of the people in Scotland that says, we’re not going to do that, we’re not going to participate in that particular language change or associate with the people just over there.”
And what did you find out? Why was there such a dramatic change?
“Well, you can’t ask people that overtly. So what we did was we ran a set of perception experiments that helped us scratch below the surface.”
Dominic and his team chose to run a series of experiments where they invited people to respond negatively or positively to short speech samples. These speech samples were deliberately chosen to evoke “Scottishness” or “Englishness”. Then the team asked the participants to discuss their view of the person who was speaking in the recorded audio.
“We played people these little sound clips and we got them to respond as quickly as they could to categorize them or at least associate them with positive or negative perceived personality traits. So they would hear a sound clip, which was from a Scottish person saying the word “car” or something like that and they would try to classify it according to some trait, like being clever or being deceitful.”
“They responded fairly quickly to the sorts of positive and negative associations that we had anticipated. For example, if you come from Scotland and you seem to have a positive attitude towards Scotland, you’ll hear things that sound Scottish and indicate those as being indicative of friendliness and intelligence and things like that. Whereas, the stimuli that sounded more English and less similar to the people in that group, they felt like it was more socially distant and they were less positively evaluated.”
At one point, they asked the participants to identify whether one particular speech sample sounded Scottish or English and why.
“They couldn’t identify why one speech sample would sound Scottish versus English, they’d say, well it just does. They would say something like, I don’t know, I guess it is the tone of voice. They wouldn’t realize that there was actually some fairly concrete features in there that, as forensic linguists, we could point to and say, Scottish people tend to do that but English people don’t.”
Along the border, how does accent affect perceptions of nationality?
They discovered that accent acted as a gatekeeper to a perception of a nationality. This was shown in first impressions, where a person would automatically assign traits to a new person depending on the accent they heard. If the accent they heard was Scottish, for example, they would expect that person to be a Scotland supporter, despite not knowing much about them at all. And when this perception of a person based on their accent did not match with their activities, people would get confused.
“If you said that you went to a football match and you went to watch England, other people would ask, why are you supporting England when you’ve got a Scottish accent? And they’d say, well I don’t have a Scottish accent I come from Berwick, I speak Berwick English. But the others would say, well you have the wrong accent to be an England supporter. So it just shows you how absolutely central accent is to the acceptability of their identity claims to other people.”
Were there instances where someone’s perception of a feature clashed with their production of it? For example, cases where a strongly self-identifying English person identified a certain feature as indexing Scottishness, but nevertheless used it themselves unknowingly or perhaps for other reasons?
“Yes, well I think people in Berwick are particularly susceptible to that mismatch of perception versus production. They’ll express utter mystification as to why anyone could think they could be Scottish. They say, I feel English, I speak with an English accent, you know, my parents are from England. What does Scotland have to do with me?”
“People ask, “Why are you supporting the English football team or rugby team when you’ve got a Scottish accent?” That’s a direct quote from one of the interviews that the Edinburgh sociologists gathered. They are mystified. They say, I just don’t sound like the people from across the border. I hate the way they speak; I would never want to sound anything like that so why do people keep misidentifying me?”
“I think there are a lot of varieties of English in this country where that happens. So, I think sometimes people just think of the next nearest place and say, well you fit best with that kind of accent so therefore I’m going to assume you’re one of them.”
How is the AISEB data currently being used?
One of Dominic’s students, Georgina Brown, is developing an automatic accent classifier system. One of the uses of this is to be able to automatically discover if a person is from a certain country or region based on their accent.
“In order to design and get this software running effectively, she ran it on the AISEB data with four different varieties of English and it performed astonishingly well. It performed as well as a human listener would do. So we had the AISEB corpus to thank for making that software as robust and high performing as it is.”
Georgina’s automatic accent classifier is an innovative use of speech technology, which could expand the realm of applications of forensic speech science. I am planning on interviewing Georgina in the near future and writing a post to share her interesting innovations with you, so stay tuned.
What advice do you have for those who want to get into forensic speech science?
“Be aware that there aren’t too many jobs in this field. But if you’re really committed to this, I’d recommend coming and doing the MSc in Forensic Speech Science in York. It is the only one in the world; it is the first and only of its type.”
“We provide the training to get you to a pretty high level in one year. If you want to practice in the UK, if you want to do actual case work, these days you’d need a PhD. York is the place to be if you want a career in this field. So, jobs are few and far between but our graduates are in the best position to apply for them. They go on to do things like, working for J P French Associates, working for security services, working for the intelligence agencies in this country, the police sometimes, and setting up labs overseas.”
So, readers, what are YOUR thoughts on forensic speech science? Are you intrigued?
Be honest, you were sold at linguistics meets CSI. I’d love to hear what you think – please email me at email@example.com if you have any questions or start a discussion in the comment section below.
Learn More about Forensic Speech Science and the AISEB Project
The data from the AISEB project can be used to aid in other researchers’ investigations upon application. If you’re interested in working with the AISEB project data, please contact Dominic Watt. If you’d like to learn more about forensic speech science, Dominic Watt, or the AISEB project, feel free to take a look at any of these resources:
- About Dominic Watt
- An Overview of the AISEB Project for Linguists and Speech Scientists
- The non-technical version of AISEB project
- Definition of Forensic Speech Science from the University of York
- The MSc in Forensic Speech Science at the University of York
Learn More about NeoSpeech Text-to-Speech
If you’d like to learn more about NeoSpeech and our speech technology, feel free to check out the NeoSpeech Text-to-Speech website here or fill out our Sales Inquiry form and one of our friendly team members will be happy to help.