Episode 69 - Natalie Braber & Alice Paver on accent stereotypes

Welcome to Lexus, the podcast all about language and linguistics. I'm Lisa Casey and Jackie Glancy, I'm Rattrana, and I'm Dan Glater. For this episode, just like to give you a content warning, we will be making reference to research that explored attitudes to criminal offences, including sexual offences. So if this is inappropriate for you to listen to or for your students to listen to, we're just giving you that warning in advance. On this episode of Lexus, we're really pleased to welcome not just to interview ease, but also a guest interviewer. So welcome to Amanda Cole from University of Essex, who's going to be asking some of the questions, and it's going to be offering some expert ideas about accent attitudes, and welcome also to Natalie Braber and Alice Paver. Natalie is a professor in linguistics at Nottingham Trent University, and Alice Paver is a research assistant at a phonetics laboratory University of Cambridge, so welcome to Lexus. Thank you, thanks for having us. So we're talking mostly about your recent paper that you've done with other contributors, which is Stereotyped Accent Judgments in Forensic Context, listener perceptions of social traits and types of behaviour. So your paper looks at a fairly kind of well-established area already, doesn't it, about accent attitudes, but you're offering something with a kind of legal forensic twist to this. Before we talk about your paper in a bit more detail, can you tell us a bit more about some of the sort of existing work that's been done on accent prejudice, attitudes to different accents, and what it's revealed? Yeah, I mean, I can start off by telling you a little bit about, about sort of general accent, and I don't know, Alice maybe wants to talk specifically more about sort of forensic work that's been done. So there's been a lot of research over really the last pretty 60 to 80 years looking at perception of accents, and what we see are patterns in the way that people respond to particular accents. We tend to find that people look at what we call status traits and solidarity traits. So status is things like how intelligent is this person, how wealthy is this person? And solidarity traits are more like how nice is this person, how trustworthy would you want to be their friend? And what we tend to find is that certain accents like RP, so received pronunciation, or for the south east of England, they tend to read very highly on status traits. So this person is intelligent, educated, wealthy, but not as well on the solidarity traits. Other accents, such as maybe like Scottish, Irish, the northeast, they tend to read highly on solidarity traits, so people like those accents, but maybe they don't still have very high status. And then the certain urban accents that tend to do badly on both people, don't read them highly on status or solidarity. And of course, that's real world implications. Yeah, yeah, absolutely. And what I would add about the sort of existing work in the forensic domain is that what people tend to have looked at is whether an accent is a standard variety like the RP, athlete mentioned, or a non-standard variety, which is kind of any other in the UK. And whether that aligns with how likely somebody is to commit a crime or how guilty they sound, or how harsh a sentence they should be given, and broadly speaking, we see standard accents are perceived positively in that domain, and non-standard accents are perceived negatively. Although some work is looked a little bit more at blue collar versus white collar crime, and we see a kind of like expected pattern there where non-standard accents are perceived as more likely to have committed a blue collar crime, and standard accents more likely to have committed the white collar crime type. And so that sort of distinction is probably something to do with sort of class-based things. Yeah. So sort of blue collar is in sort of manual labor, white collar is in sort of administrative kind of middle-class, professional sort of stuff. That's exactly it. Yeah. So we see this kind of binary distinction of those two different crime types and the two different accent types patterning in that way. Yeah. And so previous research is tended to show, I think as Natalie was saying, about certain urban accents rate quite low historically. So is that bromine scarcer too, or the kind of lowest generally? They're generally the two, yeah, the two that get the hardest hit by these sort of stereotypes and perceived ideas that people have. And then just on the subject of the sort of standard accents, I mean we've heard a fair bit about RP in the past, haven't we received pronunciation, and what that is. But I think is there now SSB is being talked about a bit more by linguists? Can you tell us a bit more about what that is and how you might define that? Yeah. So SSB is what we might call modern RP, although it's a very broad gradient. You can have people who speak probably closer to what we think of as traditional RP, and that tends to be associated with the upper classes or people, or older people, or people who are on like the news or something like that. And then the other end of the scale, we might have someone who speaks quite like not totally dissimilarly to that, but maybe has fewer of those kind of posh accentures that we associate with RP. But yeah, it's a very big gradient. And for the purposes of the public, RP and SSB are very similar. Typically the way linguists talk about it, they are actually separate varieties, and both of them are also not region-specific. So although they are associated with the southeast of the UK, actually people all over the UK speak like this, and that is because it has been traditionally seen as this standard variety, and it's often what is taught in schools. People who work in media have traditionally been trained to speak in that way as well. So we hear this accent all the time all over the place, and it really has its kind of pervasive perception as the way English should be spoken in the UK, at least in education kind of settings as well. What's really interesting is although you find in what we call in group loyalty that sometimes people read their own accents more positively, overall these sort of stereotyped ideas seem to hold even farther people who have those regional accents are still reading the standard accents more positively, even though they don't speak it themselves. So it's a really interesting thing that people have these ideas about how people speak and what that can tell us about them. Yeah, yeah. I can see SSB was a really obvious accent to choose from what you've said, and the other accents you chose, how did you kind of decide on which ones you would use for the guises? The question, we really wanted to have a good geographical range, and we knew it was also important to include accents that weren't just from England. It's quite tricky because there are so many regional accents. So like I say, we wanted to make sure it was really balanced geographically, we wanted to have a few Northern accents, Northern English accents, we wanted to have at least one Scottish and one Northern Irish and one Welsh, and then we also wanted to make sure that the south of England was represented as well, because quite often when we talk about the south of England, it's quite easy to forget that SSB is not the only variety at an RP and not the only varieties that are spoken in the south, in the south story, there's also estuary English, there's also southwestern varieties of English. So yeah, we were just trying to be as balanced as possible as basically the answer. Yeah, in a sense, considering the guises, you have to be limited in what you had, I think you had a huge range of different kind of class based and regional accents. But there's loads more things that we could have looked at, are more areas or more speakers, but the other thing you've got to be reminded is when you're doing an experimental study like this, you can't ask someone to take part in an experiment that's hours long and they'd be exhausted, so it's having to really balance time with trying to ensure coverage. I mean, one of the issues, of course, is having maybe just that one voice from an area can be seen as problematic, but otherwise, you know, we'd be asking an awful lot of our participants. So you've spoken a bit about the previous research in this area and you kind of like very comprehensively address looking at lounge assitudes, but also with that sort of forensic, forensic twist and also looking at criminal offences. So you say you looked at the attitudes reported in previous studies, but used a wider range of accents, which we've spoken about, but also behaviors and criminal offences. So I was wondering if you could talk just a little bit more about what you did around that. Well, as I mentioned, some previous studies have found this kind of binary distinction between blue color and white color crime and standard and non-standard accents, how they're associated with those particular criminal behaviors. One of our motivations was that we wanted to look past this kind of binary distinction. We wanted to look at a variety of different criminal offences. So we chose some that are associated with this kind of blue color stereotype of the criminal vandalism and theft, but we also specifically chose a driving offense because we kind of predicted that that didn't really necessarily have as much of an association with social class. And then again, similarly, we picked a sexual offence because we thought that that would be interesting to understand what kind of stereotype people have about somebody who might commit that kind of offense. In terms of the behaviors, some previous researchers looked at an association sort of stereotypes people project about the kinds of behavior somebody is likely to enact. And we did take some from previous research, including returning a lost wallet to its owner. And we just wanted to see whether if somebody is likely to commit a crime, are they also unlikely to commit these pro-social behaviors as well as anti-social behaviors that are not criminal? And we did see some really interesting problems there. Yeah, we also tried to include some sort of different of these, what we caused our moral behaviors of some that were seen as sort of positive or negative, but some that were maybe a bit more ambiguous as well. So one of the things we asked was, do you think this person would be likely to report a relative for a minor offense, which can be seen positively or negatively? So trying to just test out some different theories that maybe haven't been looked at specifically before? That's quite tricky when, isn't it? Because you might just be able to grasp, right? It might not be seen as a kind of morally, as you say, a kind of morally ambiguous, isn't it? Yeah. And she's judging you. Yeah. That's what crime they did. Yeah, well, exactly. Well, and we specifically didn't see that. So I guess the big question then, if what did you find what were the sort of main headlines for your results? So there were some findings that we expected to see and some that were a bit different. So that's always quite interesting when there's different things coming up. So we did find that the SSBE was rated highly for the status traits, which we expected to see. But we find that solidarity was less straightforward than it had been in other studies before. For example, Newcastle, which normally rates really highly for solidarity, didn't rate as well and Birmingham rated more positively than it has in other studies. That was one that really stood out for me. Maybe you come back to that in a little bit because I think that was something that was a real surprise. So sort of unsurprisingly, based on previous research, we did see that the non-standard accents tend to be rated more likely overall with committing crimes. And kind of the most important thing here is that it was really based on how working class the listeners perceived the voice to be. So basically, it was much more based on perceived status than it was on perceived, say, trustworthyness or honesty in the voice. So basically, the more working class this person's voice sounded, the more likely they were to commit crimes in general. However, something that came out that was really interesting and following on from what we were saying about how we chose the different criminal offences. Broadly speaking, the crimes were more associated with the non-standard working class accents. However, although the SSBE speaker was the least likely to commit most of the crimes, he was actually amongst the most likely to have committed the sexual offence. And that was quite a surprising finding. And really, I think, says something about the projected image that somebody has in their head when they think about this particular type of criminal. I think probably a lot of high profile cases in the media might have contributed to people's stereotypes of a kind of person who's likely to commit that kind of offence. It's an offence of a power dynamic. And so the kind of social status that you reject when you speak and you have a particular type of accent that projects your high status, your prestige, it's not a million miles away from the idea that a crime of power dynamics could be associated with that kind of image or power that you are projecting when you speak, basically. Yeah, so I'm, I guess, as well, it's that sort of idea that you know, they have the most power and it's easy to abuse perhaps. Yeah. It's, you know, quite shocking to see really that, I mean, these science snippets that people are hearing are very short, but they're quite happy to say, yep, this person after hearing them speak for, you know, 15 seconds, I do think that they're more likely to commit a sexual offence than someone else. It shows that these judgments are so fast. Yes. And I mean, I guess that's one of the sort of implications, isn't it from the sort of legal dimensions to this? It's when you think about that in terms of juries, for example, you know, that can have a massive impact on a snap decision that's made about someone's guilt or otherwise. Yeah, especially because a lot of people probably aren't aware that they have, you know, they're not consciously thinking, oh, this person sounds like this, therefore, it just seems to be, you know, they're, of course, we were asking people to make these, you know, sort of our fault, we were asking them to make these decisions, but we thought, because obviously people could be neutral and say, I don't think that they're, you know, I'm not sure they would, but they were quite happy to see you likely to do this. Yeah. And it does form sort of a broader picture of social prejudice that's based on a lot of other characteristics about somebody, you know, when you're in court and you're on the jury, you aren't just forming a judgment based on their voice, but what's kind of key here is that like voice and accent, they're not protected characteristics and they're also not something that, as Nassali mentioned, people are generally aware of having a prejudice against and they're not worn against that type of prejudice either. So that's what's quite worrying is that, you know, we have seen in the past that one of the most important things for pushing back against this kind of stereotyping prejudice is just awareness. Yeah. And there hasn't really been the awareness of the implications that these kind of prejudices will have in settings like the criminal justice system, you know, we've seen a lot of research, you know, looking into raising awareness, maybe more in the domain of jobs hiring or education or the media and all of those things are really important. And this is why we wanted to kind of shine spotlight on the implications in this particularly fraught setting that has quite severe consequences. I think it's so important that you've done that because I think so many studies, and I'm clear, you know, my own who as well, just look at does this person sound intelligent or friendly and then it gets picked up by the media, but in this kind of like, fluffy, kind of flippant way, you know, look, please don't sound intelligent, but actually I think you've shown how important it is and really shown that sort of real world consequence. That's really important for people to take these issues seriously. Yeah, because if you say things about accent discrimination, people might make a joke like, oh, you know, like you're saying, oh, brumels are excuses. But then if you say to somebody, well, actually the fact that people have these ideas mean that somebody might be less likely to be hired for a job, maybe more likely to be found guilty of sentences and given stronger sentences, is that right? And then this is if there's a sort, there's a realisation of hang on, yeah, it does really have real world effects. It's not just a joke. Yeah, and I think most people have got a kind of inbuilt sort of fairness, haven't they, with around these things and just think, well, that's not fair, even if they kind of, you know, crack to gag about it before. Yeah, I was just thinking when you were talking about some of the other, Alice was saying about the, you know, awareness of these things, consciousness of accent bias. That sort of reminded me a little bit of the work that accent bias Britain did about some of the work, you know, Deviyani Sharma and her team were doing about it. And that sounded like a really interesting way to kind of draw attention to it. I guess it's a bit, maybe there's that sort of sense that, you know, raising consciousness and awareness is a bit sort of hippie-dippy and not necessarily the most sharp-edged thing to do. But it does seem like if people are made aware of the serious implications of these things, then they, they're behavior can change. And so it's, you know, a positive, a positive kind of shift. Yeah, and it was interesting about that study, because like you say, that one was looking at, if people are told before a pretend job interview that accent bias exists, they're less likely to be discriminatory when they make these decisions. But we don't know how long that lasts. Like how often do you need to, is it a one-off you've been told therefore that stays in your head? Or is it something that you need to be regularly reminded of? Yeah. That'd be interesting to see how long somebody who had been given that warning kind of bears that in mind. Yeah, absolutely. You know, if like, you know, you told someone was and you were like, "It's all for life." Yeah, work is done. And we couldn't know this well, don't we? From, I think we talked to Robert McKenzie a little while ago about implicit bias in accents. And even people who would say, "Oh, I'm not biased against a certain accent," that they would be measured as having a reaction that was very different to one that, you know, they they they didn't have a bias towards. So that that kind of difference between implicit and explicit attitudes is quite a profound one, isn't it? Absolutely. When you were saying about the assumption that the SSBE speaker was more likely to commit the sexual offense, I mean, I think that's just an absolutely fascinating finding that we've spoken about. But it made me wonder a bit, and this is obviously speculative, but what we would find if it was a female speaker? And I know that this is complete speculation, but do you have any thoughts on ideas from doing the research, what you think we might find? Regarding the sexual offense, I really think listeners would probably be very reluctant to label a woman as likely to commit that kind of offense, and that completely makes sense, based on what we know about that particular type of offense. But also criminal offenses more broadly. I mean, part of the reason we chose male voices is because the majority of at least people in prison who have been committed for crimes are men, like the vast, vast majority. And so that's kind of why we needed those male voices. It's just more relevant for this particular field. And I do think it would be really interesting to see how these attitudes would change if we had women instead. I mean, we know that women tend to be more stigmatised for having non-standard accents as well. So it would be really interesting to see how that kind of effect of the standard versus non-standard or working class accents, whether that might be even more exaggerated in female voices. I don't know. You're thinking about that power dynamic that we were saying about whether that is the same degree with a female speaker, whether people would still see her as holding any sort of power over the reverse, she's by GSB, and what crime that could be associated with. Because we know that women who have been, you know, a sentence for very serious crimes are seen as being highly unusual, because it goes very much against what a woman should be like, especially, you know, we find if it's like crimes against children, for example, there's this real sort of backlash against this. So it'd be absolutely fascinating to do this kind of study with different types of voices as well. So just going back to some of the results, they seem to sort of indicate that people are operating on fairly kind of basic social stereotypes of certain accents. Is that a fair picture? Do you think or is it a bit more complicated than that? Is there more work behind the scenes? I mean, especially because people are given such a short time to make decisions they're having to make really sort of snap, snap judgments on this. But we know, of course, there's correlations that people make links between certain types of, you know, working class, race, ethnicity. There's all kinds of connections that are being made there. So there's probably a lot more going on, and that's one of the things that we've been talking about in different projects. I've been working on about protecting accent as a characteristic. Yeah. Actually, accent connects with so many other things. It's a really interesting feature to look at. Yeah. And of course, you know, one of the difficulties with using voice stimuli rather than accent labels, which have traditionally been used in this kind of research, you know, we wanted to use voice stimuli specifically because we wanted to mimic the kind of experience that a member of the jury or whoever would actually have in a courtroom scenario. But as I mentioned, one of the sort of drawbacks or complications that that brings is the fact that voice is a multifaceted accent is one feature of a voice, and of course, it's actually not just one feature. It's a very multi-dimension feature. You know, there's also there's pitch, there's the speed of speech, there's voice quality, like the tambourist one's voice. And lots of lots of other things that might be projecting an image into a listener's mind when they hear a voice. And we tried our best to control for these things by sort of manipulating the voice samples to be as similar as possible in the other ways. But yeah, there's there's there's an awful lot of, I don't want to say baggage, but that's the first way that I brought when you hear somebody's voice. Yeah. The other thing that is we know that people aren't always very good at recognizing accents. So it might be that somebody's making a judgment and then they're actually in their own place. Yes. Yeah. And I was sort of wondering about that as well, because I mean, you used voice stimuli, didn't you? As he said, but other other studies in the past, if he's like voice labels, haven't they? So that's quite different, isn't it? So it's not like you're saying, what do you think of the cockney accent? And it might be, well, what do they bring to mind with the cockney accent? It's a literal voice that they've got there, you know, to respond to, yeah. Yeah. Well, I think what it does is it kind of validates the methodology of using labels in a way because we did see extremely similar patterns. So it's not to say that it's more or less valid to use either of them, but they just bring different implications. But something we did actually do just to make sure that there wasn't too much of kind of, you know, as Natalie mentioned, some people are particularly bad at placing an accent and people use very broad labels sometimes. But having looked at after after our listeners, you know, made all of these judgments at the very end of the experiment, we asked them to identify where they thought the speaker came from. And we weren't able to report these results in the paper because there was already so much to be said. But just for the purposes of checking that they weren't getting it wildly wrong, we did look over those responses and make sure, and yeah, you know, they were kind of varying in terms of how specific they were. But broadly speaking, people did get it right. That's right. Yeah, we want to make sure of that. Yeah, yeah. It's guys an idiot. He kind of should recognize anything. Yeah, yeah, yeah. And just going back to one particular act and the Birmingham one seemed to have kind of improved in its kind of rating, if you like. What's going on there? Is this, is this Peaky Blinders? Is this like Alison Hammond, Joe Lysert? Honestly, yeah. I really do think that that has a huge impact having, you know, we've seen them increase in the number of presenters, for example, or newsreaders with regional accents, perhaps they're not particularly strong regional accents, but nonetheless, not everybody on TV is speaking RP anymore. And I do think that that has probably had a positive impact. And Peaky Blinders was a cultural phenomenon. It's entirely possible that that has changed some people's views at least, not everybody watched it, but you know, it was definitely in the cultural zeitgeist at one point as something that was very, you know, sort of emblematic of a particular era and of this area of the UK. And it kind of elevated this profile of like, yeah, you know, they're gangsters, they're their, their criminals, they're, they're extremely fun people, the people in this show, but they're also lovable. And, you know, it's entirely possible that this kind of thing has changed people's perceptions. Yeah, I don't know, not only if you have anything. Yeah, I mean, if it makes you feel any better or a strong worse on that, another study I've done recently that looks at what people expect to hear from baristas or from lawyers, Birmingham was absolutely lowest in everything. So I'm sorry. Cyprus and Birmingham were not positively received as that's how a barrier to should sound. So maybe even though people are becoming more used to hearing different accents in certain positions of power, they still expect to hear plosh people. And I'm pretty sure you mentioned DeFiani Sharma's work, they, there was a recent paper sort of evaluating how things have changed over the last 50 years since a particular survey was done about accent attitudes and I think Birmingham performed pretty poorly as well. So, you know, and maybe it's a case of in that particular example that the voice stimulus was responded to differently than the label of the accent. Perhaps our speaker just had a particularly nice voice, which is one of those situations where we can't tell you exactly what's going on. But, you know, that's maybe one of the reasons why we see some slightly different patterns is because we've used a different method for eliciting responses from our listeners. It's also quite a long way of removing it, and it's sounding like you wouldn't commit a heinous crime to sound it like you could be a lawyer or a politician. Absolutely. Yeah, exactly. It's probably quite a big gap there. Birmingham's a scale. So, you might be a nice person, but don't, you know, don't do anything important. Yeah, it's quite kind of disturbing still isn't it, that there are at least kind of attitudes about these things. Maybe I've been too optimistic there. I mean, just a sort of quick question on the media coverage of it. Have you looked at how it's kind of being picked up in the media? What did you make of that? Because there's some interesting stuff wasn't there? They came out around it. Yeah. I mean, I have to say that my family, who are actually off from Liverpool, weren't very happy about some of their headlines. You know, your classics sort of like, oh, all scousers sound like criminals. You know, they had a good laugh that it's not really a joke. One of my uncles is actually an interpreter. And when he's speaking French or German or Dutch, you know, I think probably the other people can pick up on the fact that he's maybe not a later speaker, but he is fluent. But when he's translating the interpreting the other way and speaking in English, he does get some negative reactions based on his accent. And he has commented before, you know, this is just anecdotal, but he's in a relatively high prestige job. And he still faces some stigma because of the fact when he speaks in English, he speaks with a scous accent. And he's had comments, you know, jokes, but they're not fun to hear about, you know, oh, you can't trust him because you can't trust a scouser kind of thing. So just personally, I know that some of the headlines, of course, you know, they're picking up on what we observed. And that's fine. And we don't want to entrench these attitudes any further. It's important to shine a light on them, though. And the thing that was so interesting about them, because so much academic work that's done is great, but it doesn't, it doesn't always reach the public. And this one seemed to really strike a card. I mean, I have a 15 year old son and he came home. And of course, having an unusual surname must help in that case as well. But we're actually students and teachers have said to him, it's out your mum who's done this study because it was in papers that people were reading. It was, it was really reaching a wide audience. It was, I mean, I listed a massive world when two of media appearances. So it is, people are interested in this. Yeah, I guess around either, but you've had a bit of that with some of your, your studies, haven't you? Kotly's dead, it's been killed. Well, sometimes, sometimes less. You must have had a busy week, Alice. It was a busy, busy time. It was quite intense. But you know what, like Natalie was saying, like a lot of this kind of stuff doesn't normally get picked up on by the media. And it's really brilliant to have this kind of research, publicize. If we can just reach like a few people who might start thinking about how they enapse, it's kind of accent prejudice in their daily life, then great. That's, you know, just just a few more people who might change the way they judge somebody based on how they speak. I think it's got the jury service in the post. Yeah, yeah. Just on that last thing, I mean, what sort of insights do you think it might offer to courts, police, legal system, more generally? Is it, is there any kind of way in which you can see this being picked up and used a bit more practically by by people working in those fields or even sort of institutions, organisations? Yeah, it's just the awareness side of it. So this, obviously our strand was just one of those four strands on the project who are sort of looking at, you know, all other factors of this. You know, we've been working with how the police and how legal services deal with these kinds of issues. And as Alice was saying earlier, sort of putting something like this into, you know, if somebody's going to be in a jury or if somebody is listening to a voice like this, just reminding them that there are these stereotypes to ensure that they maybe aren't influenced by that. It's going to be helpful, I think. And, as Natalie mentioned, you know, that this is part of a broader project, which is actually looking at, in particular, looking at voice parades, which is, yeah, yeah. Yeah, it's a type of sort of evidence gathering that can happen in the criminal justice system whereby if somebody was a witness or a victim, which is too or a victim of a crime, and they heard a voice but didn't see a face and they are later asked to recall that voice from amongst a group of others, just like a visual parade, but the voices. So the broader project was looking at sort of improving that procedure because it's currently very expensive and time-consuming and difficult to implement and consistently used as well, exactly. And so part of the motivation for doing this research and some other studies that we've also conducted was to understand why a voice might stand out in that particular voice parade scenario as sounding particularly guilty, you know, and maybe erroneously being chosen. And that definitely puts the the innocent oil voice at risk of perhaps being incorrectly chosen or even even a suspect who is innocent as well. So it's important to kind of understand a little bit more about what makes a voice sound guilty at stand out in that particular scenario. Yeah, interesting, yeah. And I guess that's, you know, as you say, that's going to have huge implications in the courts, isn't it? Absolutely. I mean, you know, voice evidence never is the one, you know, nail in the coffin that convicts somebody, but it can form, as we were saying before, a part of the broader kind of picture of prejudices, but also a broader part of, you know, some evidence that may come to lead somebody to be convicted. So it can have serious implications. So how would how would a voice parade be used then? What's the kind of normal procedure for that? The normal procedure is that if somebody was to hear a voice in the scenario of a crime and they don't see the face, they are brought back in later just as they would be with a visual parade and are asked to listen to an audio parade consisting of the guidelines currently say nine voices, including the suspect. The reason this is quite complicated and expensive and how I'm consuming to implement is because unlike with visual parades, we don't have a bank of sort of automatically assigned voices that kind of can fill the rest of the parade, but also you need an expert phonetician who help with constructing that parade because voices are a lot more complicated than faces and also because we can't reliably sort of synthesise, you know, a voice in the way that you would with like an e-fit of the person's face. So you have to have an expert phonetician assisting with this, you have to have a sufficiently long in terms of duration sample of the voices in the parade to ensure that you hear them properly. Whereas with faces, you see them in an instant and that's it, you don't need a particularly long duration to hear all of these things. The listener will also have an opportunity to hear all the voices several times and then they can make a decision and say whether they thought the voice that they originally heard a scene of the crime was in the parade or not and if so, which one it was. And the other thing is if you imagine, sorry, Alice, that you have to get voices that sound similar enough to the suspect's voice that, you know, but also not too similar that becomes impossible to tell them apart because then you're not going to have, you know, if you have eight voices that sound identical, then that's not going to be very effective if you have eight voices that sound one way and one voice sound specifically different. And yeah, one of the aspects that we haven't finished work on quite yet is that they do seem to, these what we call the foil voices, so the voices that aren't the suspect, there are some that seem to be picked more regularly than others and not because they sound more similar to the target voice. So one of the things we were interested in is what Alice was saying earlier are there some voices that just seem to be guilty sounding, there are then a higher risk of being picked. So I guess if it's face is right, you see a face and then you have to match it to a face again, but with hearing someone speak, beyond kind of maybe more super segmental features, you might hear them say a certain linguistic feature or very, but they don't do it again. Well, they produce one variant the first one, but they don't the second. Or where you heard the voice where the whispering shouting, you know, do you have a mask on? A study that I mean done that look at replicating that, you know, you can make your voice sound quite different as well. It's really, really complicated and much more complicated than a visual identity parade. Of course, you know, one of the reasons an expert needs to be involved is precisely because you need to consider exactly that kind of thing. So maybe the phonetician might be told, oh, this ear witness heard a person saying three words and they'd say, well, I don't think there's any point doing a voice parade in this scenario because it's just so unlikely that the person is going to make a correct identification. But on the other hand, if they spent hours hearing this voice, you know, perhaps they were kidnapped and blindfolded and they interacted a lot with with the with the perpetrator, that's the kind of scenario in which we might be more confident that we call an ear witness might make a correct identification. So there's lots of factors that go into it. And as I mentioned already, you know, this is precisely why we need experts to be involved in the process because as we've seen in this study, lay listeners, regular people have their own ideas of what voices of accents of speech and what that says about somebody. But also they have, you know, maybe slightly misinformed ideas about how well we can remember voices or how we can translate our memory of voices across different scenarios. So yeah, it's really important to have an expert involved and like Natalie was saying, also understanding why a voice might stand out, you know, some voices do just seem to be picked for one reason or another and understanding a bit more about that helps us implement these kinds of voice parades as evidence more accurately and carefully. Yeah, cheers for expanding on that. I think that's really interesting to mention to it, but you know, wasn't really sort of picked up initially in lots of the reporting. Absolutely. I mean, they are complicated. Yeah, well, that's brilliant. No, thanks very much. It's really interesting. And thanks for talking us through it and the various findings and approaches that you used. Thanks for having us and for all of you. Thank you for coming back by this as well.