Author Archives: charlesantaki

Guest Blog: The Conversation Analyst as an Expert Witness in the Courtroom

Some court cases rely on the interpretation of verbal exchanges – recorded conversations, interviews, police interrogations. The expert may be asked to say whether something is evidence of a bribe, a threat, a confession, and so on. Gary C. David PhD CCS is one of the very few conversation analysts who have been called on to help the court. His report makes for fascinating reading.

Gary C David, Bentley University

When is an interrogation and interview, or an interview an interrogation? How much does it matter how such encounters are characterized versus what they look like? How can we tell the difference between the two? These are questions that I had to address as I was asked to examine a police encounter with a person suspected of murder. My expert opinion could make a difference in whether evidence is admitted or excluded, and whether the suspect goes to trial or goes free.

For the past approximately seven years, I have been working on legal cases in which police interviews and interrogations have been called into question. I originally started this work with retired Detective Jim Trainum, a nationally recognized expert on false confessions. I got to know of Jim’s work through hearing him on an episode of This American Life, where he was discussing his role in unknowingly producing a false confession. After reviewing the video recording of the police interrogation, he was able to identify how he and his partner generated a false confession. As he retold this story, I immediately saw the potential for conversation analysis to contribute to this work. After contacting him and introducing myself, he agreed, and we have been working together ever since.

Coercion and contamination

Typically, I work on cases in which the police/suspect encounter is called into question, focusing on whether coercion and/or contamination produced a false confession. Coercion would involve law enforcement creating duress for the purposes of getting a confession, such as promising (or even strongly implying) a reduced sentence in exchange for a confession. Contamination involves law enforcement providing details of the crime to such an extent that it can’t be known whether the suspect was really there, or just repeating what they were told.

When working on these cases, I position CA under the category of forensic linguistics, which broadly includes approaches focusing on the nature and impact of oral and written communication/language occurring in legal contexts. Conversation analysis has been identified as part of that field, especially for the purposes of research and scholarship. However, I don’t believe CA has been introduced as a scientific area of analysis in criminal proceedings. Until now.

A typical US police interrogation scene (Source: Youtube)

An example case: interview, or interrogation?

For one particular case, I was asked by defense counsel to evaluate a police encounter with a suspect to determine whether it was a non-custodial interview or a custodial interrogation. This determination is important because it determines whether or not the suspect’s Miranda rights need be given. I was asked to use my training as a conversation analyst, my participation in police interrogation training programs, and my previous work on other cases to render an expert judgment on whether the encounter was an interview, interrogation, or both. To do so, I reviewed the following materials:

  • The police encounter with the suspect, both audio recording and official transcript
  • Transcripts from pre-trial hearings
    • Police reports from the investigation

In the United States, the primary method of interrogation is known as an accusatorial approach, meaning that the goal of the encounter is to get the suspect to confess to the crime. This can be compared to the information-seeking interview where the investigators are looking for information regarding the crime. Each is identifiable by the conversation structures present, as well as the tactics used. Thus, determination of interrogation or interview is not a matter of opinion, but a matter of examining the structures and practices

Working with the court

In my role as expert, I had the opportunity to be part of a pre-trial motion hearing where I shared my analysis with the judge, the prosecutor, and the defense counsel (with whom I was working). Part of that process included trying to explain what conversation analysis is and how it was applicable to this case. In doing so, I was able to establish that forensic linguistics and conversation analysis are scientific methods that can be used for the examination of police encounters with suspects. As a result, the court deemed that CA is an appropriate tool for rendering judgments on the nature and structure of these interactions.

A number of CA’s features were important as part of testifying.

First, that we rely on analysis and description rather than interpretation and assessment of mental states which can raise objections in courts in terms of admissibility.

Second, the ability to go “back to the data” in the form of the transcript as forms of evidence to support our analysis. It is not my opinion that this is happening; I can point out where and how it is happening in the transcript and recordings.

Third, focus on audio (or video) recordings to explore parts of the encounters that are not rendered on the official transcript. This focus on listening also is important in terms of checking the accuracy of the official transcript. Fourth, the corpus of conversation analytic work and concepts that have come before which can be used in the examination of police/suspect encounters.

My judgement to the court

I was able to come to a firm conclusion for the court. My assessment was that the encounter was, in fact, an interrogation: it had the conversational structures and interrogation tactics associated with an interrogation. This wasn’t my expert opinion, but rather my expert analysis.

While I don’t currently know the outcome of my testimony, I do know that the three hours that I spent testifying allowed me to highlight the importance of conversation analysis in the examination of police/suspect encounters. Defense counsel admitted that he was skeptical of what I could contribute going in, but then realized the importance of CA in the examination of such encounters. As more CA work is done as expert testimony, we can expect more people to see the value that CA can contribute to legal cases and courtroom proceedings.

Gary David is a Professor of Sociology and Experience Design at Bentley University (Waltham, MA). You can read more about his work and thoughts at  You also can learn more about his professional and consulting work at

Guest Blog: Doing CA on hospital wards with front-line healthcare professionals

Conversation analysis offers a great deal to those trying to improve how to communicate with people with disorders of language. It’s not always easy, and practical obstacles keep getting in the way: Isabel Windeatt from Nottingham University gives us a lively account of what it’s like to collect and analyse data on a ward for older people .

Isabel Windeatt

I’ve been working closely with front-line healthcare professionals in my role on the VOICE2 study, a conversation analytic study of communication between staff and people living with dementia who are in hospital. I want to share the benefits of collecting data and sharing preliminary CA analyses with healthcare professionals, as well as some of the challenges, in the hope that others will find solutions to collecting data in challenging environments, and be encouraged to involve healthcare staff during the analysis.

The healthcare professionals I refer to are ward-based staff who are unfamiliar with CA and don’t undertake research. I also work with clinical academics who are familiar with CA and recognise its value in healthcare research – they’re the reason I have a job.

Recruiting healthcare professionals on acute hospital wards

The data I’ve been collecting involve naturally occurring interactions on UK acute hospital wards between healthcare professionals and patients living with dementia who were prone to distress, recorded at times when distress was anticipated to occur or had previously occurred.

First, I had to get familiar with the scene. This required a traditional ethnographic immersion in the data collection process. From March to September 2022, myself and 3 other researchers spent 6 months on acute older persons’ hospital wards to collect almost 10 hours of video and audio data.

You have to become familiar with the everyday realities of the scene

Much of that time was spent finding patients for the study and recruiting willing healthcare professionals from amongst the healthcare assistants, doctors, nurses and therapists working on the wards. I had to try to build the trust of staff by talking with them, getting to understand their roles, and generally becoming a bit of a ward fixture so that staff could get to know the person behind the camera. Staff were always busy. I was careful to introduce myself to as many staff as possible and explain what we were doing on the ward – it’s a bit intimidating having someone unfamiliar turning up at your workplace with a large folder and recording equipment, watching and taking notes on what you’re doing!

Start at local manager level. Initially, we presented at a Ward Managers’ team meeting so that those running the wards knew who we were and what we were doing on their ward. We would also approach Ward Managers when first arriving on a ward, recruiting them to take part in the study as this set an example for other ward staff, encouraging them to sign up as well. The Ward Managers recommended staff who might be comfortable in front of the camera. We also presented at Medics’ team meetings and attended board round meetings on the wards. This networking offered our first recruits on each ward and, having broken the ice, recruitment would snowball as talk about the study passed between healthcare professionals.

Recruiting people on the fly

That’s not to say we didn’t have to do our fair share of ‘cold recruitment’. Many healthcare professionals who signed up were those we had to pitch the study to from scratch. For this, having a polished ‘elevator pitch’ was useful, and helped with any nerves about distracting a healthcare professional at work, especially as reading a 10 page long study information sheet didn’t make signing up all that appealing. This pitch would often get interrupted by staff having to dash off to attend to a patient, or a patient might instead stop me – either thinking I was their relative, falling asleep on me, or just wanting to chat. I’d find the healthcare professional later, if possible, to resume our talk when there was a brief moment of calm between patient-focussed tasks, such as when they were watching over their assigned bay.

Not everyone wanted to be part of the study. Many healthcare professionals didn’t feel comfortable being filmed and I even received one unmitigated ‘no’ without any delay or account. Fascinating interactionally, but also mildly disheartening. However, I found many staff were still happy to chat about patients who might be suitable for the study, giving me valuable information on distress patterns of patients I was going to speak to. They would also share their own experiences of dealing with patient distress, and cajole other healthcare professionals to take part.

Blending in

We were careful to emphasise that staff did not have to take part, that they could stop at any time, or say ‘not today’ to us. For those that did take part, our two stage consent process, in which we took consent for future use of video material after they had viewed the recording, meant that they could decide after the recording how they would like it to be used and who would see it. 96 staff allowed us to take up their already limited time by agreeing to take part, a testament to their interest in improving care and demonstrating a selfless nature.

Getting immersed in data collection

A lot of the time, the patient being filmed would wander about the ward, or staff would shifting positions to do their work. So I had to navigate around the scene, trying to keep the interaction on screen, but also keeping in mind patient dignity by covering the camera lens if the situation demanded it.

Seeing how staff handle difficult moments

It was very illuminating collecting data in this way. I got to see first-hand what life on the wards was really like for staff. In my first week, a patient I was assessing the capacity of, switched from being quiet and calmly spoken, to yelling at me that he wanted to murder someone. Staff suggested that it was best just to ignore this behaviour and I had many pleasant chats with this patient after this occurrence. In the few months I was there, I’ve witnessed patients throwing walkers and anything else to hand at staff, fighting each other, and trying break the ward doors open to escape. One of my most memorable occurrences of distress, was a patient who, at the time of his distress, believed staff on the ward were a gang who had taken his young children. This gentleman had taken another patient’s shoes thinking they were his, then used them to hit a ward window, resulting in a resounding crack and broken window. Staff were amazing in all of these circumstances, responding calmly, and with understanding, to patients who couldn’t always help their disinhibited behaviour.

The biggest benefit of being so immersed in the data collection, was that I was able to
learn about the context of the interactions that I would otherwise have missed out on. As the focus of our research – distress – could be fleeting, being on the ward for long periods allowed us to collect field notes to be used alongside the recorded data, something not always done with CA (Mondada, 2012:33; ten Have, 2007:88). Much of our data is better understood in the light of the context of the patients’ prior distressed episodes which these field notes provide.

Hit and miss with data collection

Data collection was time consuming: on some days no data collected, and others more time was spent standing around observing, or waiting for a healthcare professional to undertake a task with a recruited patient. On these long days I was very grateful to the wards that had sofas in their bays.

It was a tiring but valuable process, and being on the wards for these extended periods allowed me to learn some of the healthcare jargon and contextual information that I lacked as non-clinical analyst. It’s also given me a better appreciation of healthcare professions’ concerns, an understanding of their roles, and of what our research is being done in support of.

Sharing analyses with other healthcare professionals

I’ve since shared some work-in-progress with other healthcare professionals at monthly project management group meetings and with a community based team in another NHS Trust. Online meetings make doing this sort of dissemination much easier. The analysis I presented wasn’t finished (is it ever?), and I didn’t have any concrete findings to show them. They understood that and didn’t push me for any. What they did do, was offer their own perspective on the data as healthcare professionals. Presenting here offered valuable insights on the analyses and new perspectives into the data that I otherwise may have missed. It reminded me to take a step back from the data and consider why we are doing this research.

Final thoughts (for the moment). A commendable goal of any research is to have the findings from it put into practice and travel beyond academia. The point of research is to learn, not just for the sake of learning (which is still a valuable endeavour), but to also apply that learning to the real world. Working and sharing work at each stage with healthcare professionals is an essential part of the research process to ensure that we have a practical impact with our findings. Building a network of relationships with front-line healthcare professionals and having ongoing discussions with them helps to ensure that our work is relevant and applicable and we should strive to ensure this becomes an essential part of the research process whenever possible. Personally, I’ve found it by turns exciting, frustrating, and occasionally emotionally challenging – and utterly fascinating throughout.


Mondada, L. (2012). The Conversation Analytic Approach to Data Collection. In The Handbook of Conversation Analysis (eds J. Sidnell and T. Stivers). 

ten, H. P. (2007). Doing conversation analysis. SAGE Publications, Limited.

Guest Blog: Being Smart About Artificial Intelligence

There’s a lot going on at the interface of AI and speech – both recognition and production – and some of it draws on ideas from ethnomethodology and conversation analysis. But is it any good? Stuart Reeves runs the rule over some of the issues.

Stuart Reeves, Nottingham

Artificial Intelligence is a big deal now. We’re told that AI systems are reaching and even exceeding human performance at things like playing games, hearing and transcribing words (Xiong et al., 2017), translating between languages, or recognising faces and emotions (Lu and Tang, 2014). This means we are entering a world where it’s possible for people to have conversations with AI agents (a Google researcher recently claimed a chatbot as “sentient”), or get computers to understand what’s in a picture and even generate their own art when prompted.

Continue reading

Guest blog: Should we share qualitative data?

Conversation analysts soon accumulate many hours of tapes and transcripts; usually these have been collected on the understanding that they are for the researcher’s own use, with permission only to publish extracts anonymously. But should such data be open to other researchers? Jack B. Joyce, Catrin S. Rhys, Bethan Benwell, Adrian Kerrison, Ruth Parry summarise here the arguments examined in a recent paper.

Jack B Joyce, Catrin S Hughes, Bethan Benwell, Adrian Kerrison and Ruth Parry
Continue reading

Guest Blog: The EMCA Doctoral Network Meetings restart, November 2021

Among the many formal and informal networks that support postgraduates and early-career researchers in CA and ethnomethodology, the EMCA is perhaps the most venerable and global. I’m delighted that two active members, Felicity Slocombe and Andrea Bruun have sent a report on the most recent meeting, November 2021.

The EMCA (Ethnomethodology and Conversation Analysis) Doctoral event ran every six months, pre-pandemic, with universities taking it in turn to host the event.

Continue reading

Guest Blog: SPAC, an online space for doing CA in Spanish

CA is well established in a number of Spanish-speaking countries, but there is always room for more initiatives, and for ways for sometimes isolated researchers to meet together. I’m delighted that Luis Manuel Olguín has sent in a report on the Seminario Permanente de Análisis de la Conversación, a lively and inclusive forum for Spanish-speaking CA researchers.

Luis Manuel Olguín, UCLA

Although CA is well-known across Spanish-speaking academia, resources for learning and teaching CA in Spanish are still significantly scant, especially if compared to other approaches to language use and social interaction with established traditions in Spain and Latin America. Similarly, the Spanish-speaking CA community is still relatively small and largely scattered across countries, making it difficult for CA to take root in Language and Social Science programs and departments. 

Continue reading

Guest Blog: Promoting CA in Brazil

As Conversation Analysis is increasingly taken up by researchers across the world, we are seeing efforts to bring the approach to their wider local communities. There are several initiatives in Brazil, and I’m delighted that Fabio Ferraz de Almeida, currently working in Finland, has sent in this report of an inaugural workshop in Sao Paulo.

This image has an empty alt attribute; its file name is screenshot-2020-12-21-at-11.23.30.png
Fabio Ferraz de Almeida

The idea of organising an introductory CA workshop in Brazil began to take shape last year, while I was talking to a colleague, Bruna Gisi, professor of Sociology at the University of São Paulo (USP). Bruna was developing a postgraduate course on EM and Goffman and invited me to participate in one of the lectures. According to her, several sociologists in Brazil often talk about ethnomethodology but they rarely show how to put  it to use. Her suggestion was that we  discuss a particular EM concept and show how to ‘apply’ it in empirical research. 

Continue reading

Guest Blog: Using inclusive language when transcribing French data

Every language poses it own problems when transcribing from the spoken to the written word. In this guest blog, I’m delighted that Dennis Dressel, Wifek Bouaziz and Marie Klatt, all at the University of Freiburg, take us through the dilemma of transcribing French, with its convention of treating plurals as masculine, while also trying to respect inclusivity. How does the transcriber acknowledge references to females when the language conventionally refers to them as male?

Continue reading

Guest Blog:  Building an EMCA community at CADSS

Groups of EM/CA analysts have sprung up all over the world to share expertise, pore over data together, bounce ideas off each other and provide a sense of shared community. Here Simon Stewart gives an enthusiastic account of recent developments of the group based on the south coast of England.

Simon Stewart, Southampton

This post is intended to share with the CA community some of the resources and learning that have come from our group, CA Data Sessions South (CADSS), in its first 18 months.

Continue reading