Second International Workshop on Forensic Linguistics/Language and the Law:
present dimensions and new perspectives and challenges.
|
|
|
 |
Current Issues in Forensic Phonetics
Paul Foulkes
Department of Language and Linguistics Science
University of York, York, United Kingdom
In this presentation I will address two current issues in forensic phonetics: (1) the use of a likelihood-ratio based framework for the expression of conclusions, and (2) language analysis for the determination of origins (LADO). The aim of the presentation is to raise both issues for wider discussion.
The first issue has been the subject of considerable debate within and beyond phonetics. Arguments in favour of the LR framework in forensic phonetics are presented by Rose & Morrison (2009). While accepting these arguments in principle, I will suggest that in practice the framework is not implementable at present.
LADO has only recently been a subject discussed by forensic phoneticians, yet it is clear that phonetic and phonological analysis can make valuable contributions to the practice. I will present some empirical evidence from a recent study (Wilson 2009) which is a first attempt to assess the relative skills in dialect recognition of LADO analysts, native speaker-listeners, and phoneticians.
References
Bio-note
Paul Foulkes is Reader in Language Variation and Change at the University of York and a consultant with J P French Associates. He is director of courses in Forensic Speech Science and was co-editor of the International Journal of Speech, Language and the Law from 2002-2009. He has research interests in phonetics, phonology, sociolinguistics, and language acquisition as well as forensic phonetics. His publications include the edited volume Urban Voices (Arnold 1999), and articles in Language, Journal of Phonetics, Laboratory Phonology, Language and Speech, Journal of Sociolinguistics, Journal of Linguistics.
The Integration of Automatic Speech Recognition Systems into the Forensic Combined Methodology. A Proposal
Carlos Delgado
Laboratorio de Acústica Forense
Comisaría General de Policía Científica, Madrid, España.
Nowadays, in every scientific discussion field related to Forensic Speech Analysis (F.S.A.) there is no doubt at all concerning what kind of approach is needed to be integrated as a part of a particular F.S.A. methodology. These methods, which combine perceptive, acoustic and phonetic-linguistic studying approaches together with the application of automatic or semi-automatic recognition analyses, are considered to be the most suitable option. However, the integration of automatic recognition systems has reactivated an on-going debatable discussion topic: the way in which forensic scientists have to report their conclusions.
Apart from presenting the State of the Art in Forensic Speaker Identification (F.S.I.), this paper deals with a proposal which fits into balance the results achieved using the Bayesian framework and those expressed when using confidence ranges. Additionally, two research projects carried out in the Spanish Police Forensic Acoustics Laboratory are mentioned briefly. One of them focuses on the categorization of phono-articulatory and acoustic parameters in spoken Spanish by taking into account their identification weight; the other one is related to the assessment of the voice comparison methodology used by the Forensic Acoustics Laboratory of the Spanish Scientific Police.
Bio-note
Dr. Delgado Romero joined the Spanish Police Forces in 1980, becoming a Senior Expert for the Central Government since twenty years ago. In his early years on the professional career, he carried out some different operative tasks, being posted in the Scientific Police Central Department five years later. Since 1989, he's part of the Forensic Acoustic Lab staff, which is under his management from 1995. His PhD thesis, "The speaker identification in the forensic scope" is considered as a pioneer into its investigation field in Europe.
His training as an expert in speaker identification began at the Michigan University, achieving his qualification in 1992 as an international expert in acoustic analysis and voice identification by the International Association for Identification (I.A.I.). Since then he was member of the Qualification Board and Vice-chairman of the VIASS Steering Committee (Voice Identification and Acoustics Analysis Subcommittee) in the I.A.I.
A project for standardization in forensic acoustic was introduced by him in the European Union in 1995. He also organized by the very first time in Spain the meeting of the Working Group for Forensic Speech and Audio Analysis (WGFSAAS) from the European Network Forensic Science Institutes (ENFSI), in 1999. From 2002 until 2008, he acted as Vice-chairman of the mentioned Group. It was in that year that the "Terms of Reference for Forensic Speaker Analysis" were approved by the WGFSAAS after being entrusted and drawn up by Dr. Delgado.
Together with experts from Italy, France, Belgium and Slovakia, he has developed different phases of the SMART Project (Statistical Methods Applied to the Recognition of the Talker) -which was funded by the European Commission.
Under the University Institute for Police Sciences Research management (Universidad de Alcalá de Henares), he concluded a research project named "Evaluating of the Speaker Identification Method used by experts in the Scientific Police General Directorate", publishing the results in the 2009 Institute yearbook. Furthermore, in that year he also finished a long and in-depth detailed research on "Categorization of phono-articulatory and acoustic parameters in spoken Spanish for the Forensic Speaker Identification". This survey, carried out together with some members of his laboratory, has a particular relevance because it is the very first time that several speaking references have could be finally organized by hierarchy, taking into account their identification weight.
Forensic Linguistics: Answers and Questions
Malcolm Coulthard
Centre for Forensic Linguistics
Aston University, Birmingham, United Kingdom
In this paper I will both present some of the major recent findings in each of the three sub-areas of forensic linguistics and indicate areas where research is still needed.
The written language of the law: here I will highlight the dangers of opaque legal documents aimed at lay readers, indicate some of the successes in amelioration and focus on some which still need attention.
Interaction in legal settings: much has been done to investigate interaction from the initial police interview to the final sentencing act in court, but many of the insights derived into problems with police interrogations, the treatment of witnesses in court and the provisions for those who need interpreters have yet to be successfully applied.
Language as evidence: linguists have expressed opinions on a wide variety of texts using differing selections from the linguistic toolbox. One area currently receiving a lot of attention is the attribution of texts to authors. Techniques already in have had major successes in court, but current research is working both to make the techniques even more robust and to extend into new areas like identifying the first language of those writing in a second or foreign language.
Bio-note
Malcolm Coulthard is Professor of Forensic Linguistics and Director of the Centre for Forensic Linguistics at Aston University, where he is also Associate Dean for Research. He was foundation President of the International Association of Forensic Linguists and founding Editor of the journal Forensic Linguistics. He has written expert reports in over 200 cases and given evidence three times in the Courts of Appeal in London and many times in lower courts in England, Germany, Hong Kong and Northern Ireland. He is co-author of An Introduction to Forensic Linguistics (2007) and co-editor of the Routledge Handbook of Forensic Linguistics (2010).
Recent Developments in Forensic Idiolectometry
M. Teresa Turell and the ForensicLab team (Raquel Casesnoves, Jordi Cicres, Montserrat Forcadell, Núria Gavaldà, Fernanda López and Maria Spassova)
ForensicLab, IULA, Universitat Pompeu Fabra, Barcelona, Spain
This paper presents the methodology and preliminary results obtained in a series of exploratory studies in the field of idiolectometry and forensic linguistics carried out as part of a project undertaken by ForensicLab (IULA-UPF). The project, entitled Forensic Idiolectometry and Index of Idiolectal Similitude ((FFI2008-03583/FILO) - PI: Dr. M. Teresa Turell), explores and develops the differences existing between the idiolectal style of different speakers and writers, and establishes an Index of Idiolectal Similitude (IIS), which would allow forensic linguists to accomplish forensic voice and text comparison more reliably and contribute to a base rate knowledge of population distribution at different linguistic levels (phonological, morpho-syntactic and discourse-pragmatic levels and in three different languages (Catalan, Spanish and English).
Results derived from the application of several statistical methods (M1, M2 and M3) allow us to draw still a very provisional conclusion as to the effectiveness of M3. These preliminary analyses also show interesting results that confirm some of the hypotheses which are being validated in forensic idiolectometry, in that there seems to be greater inter than intra-speaker/writer variation, and that an individual's idiolect is quite stable despite the course of time and the use of different genres. Further research will improve the selection of the analysed variables, the final methodological approach for the calculation of the IIS and its application to real forensic cases.
Bio-note
M. Teresa Turell is Professor of English Linguistics at Universitat Pompeu Fabra (Barcelona, Spain), Director of ForensicLab and Academic Director of the first Master's degree in Forensic Linguistics offered in Spanish around the world. In the last 5 five years she has written expert reports on plagiarism detection, authorship attribution and trademark litigation in over 40 cases in Spain and the US, and given evidence many times in lower courts and higher courts in Spain. Turell is editor of Lingüística Forense, Lengua y Derecho. Conceptos, Métodos y Aplicaciones (Barcelona: Publicacions de l'IULA, 2005) and co-editor (with John Gibbons) of Dimensions of Forensic Linguistics (Amsterdam/Philadelphia: John Benjamins Publishing Company, 2008). She is the Principal investigator of the on-going project: Forensic idiolectometry and index of idiolectal similitude (FFI2008-08335), funded by the Spanish Ministry of Science and Technology. Professor Turell is vice-president elect of the International Association of Forensic Linguists (IAFL) and will become president in 2011.
Querying the (Non)importance of Theories of Idiolect for the Practise of Forensic Authorship Analysis
Timothy Grant
Centre for Forensic Linguistics
Aston University, Birmingham, United Kingdom
In this talk I shall briefly review theories of the linguistic individual and of idiolect and discuss how important these theories are to the practise of authorship analysis. The theoretical debate surrounding forensic practise can be skewed by legal admissibility issues and thus become inflated in importance but I shall argue that some of this debate is misplaced and that it is possible to provide sound authorship analysis for investigative and evidential purposes without reference to the theoretical disputes. I shall describe two cases to illustrate this point; the first case is that of a series of more than 60 racially and sexually abusive letters sent throughout the UK for which I provided a sociolinguistic profile and contributed to a successful national media appeal. The second case involved the determination of whether a husband or wife sent a series of text messages around the time of the wife's murder. Both cases I will argue rest more on systematic linguistic description than linguistic theory and although debates concerning idiolect are academically fascinating and important they can have limited relevance for forensic practise.
Bio-note
Dr Tim Grant is Deputy Director of the Centre for Forensic Linguistics at Aston University where he teaches English Language linguistics. He previously lectured in forensic psychology at Leicester University. His consultancy primarily involves authorship analysis of threatening and abusive communications and he has worked in civil and criminal contexts including investigations into sexual assault, murder and terrorist offences. His evidence has been used in cases across the UK as well as in South Africa and Pakistan. He publishes on forensic linguistics and forensic psychology and his research into text messaging analysis was awarded the 2008 Joseph Lister Prize by the British Science Association. UK media work includes involvement in BBC Crimewatch appeals, appearances on the BBC Radio 4 Word of Mouth programme and feature articles in the Independent newspaper.
'I could lose everything': Legitimation and Self-presentation in Telephone Calls for Police Assistance
Frances Rock
Centre for Language and Communication Research
Cardiff University, Cardiff, United Kingdom
Broad definitions of forensic linguistics set the boundaries of the field sufficiently wide to include scholarship on language and communication which occurs within and around legal systems in order to accomplish activities relating to law, crime and justice. One of the main settings which has received scholarly attention here is policing.
Studies of language and policing have typically, though not exclusively, focussed on the needs, agendas and concerns of legal practitioners who may be concerned with such activities as gathering evidence or presenting information, often on behalf of the institution. This paper takes the other possible focus - the needs, agendas and concerns of lay people in their interactions with those acting on behalf of legal institutions. Using data predominantly from telephone calls for police assistance, this paper examines lay peoples' interpersonal agendas.
The paper asks how speakers' orient to the transactional during calls. Specifically examining how they work to elicit a response to their call which goes beyond only talk. For them, interactional resources may be geared towards providing justification and evidence for police responses and explaining and legitimising their position as someone eligible for help.
This paper draws on work from sociolinguistics and discourse analysis. Using literature on legitimation and self-presentation, the paper contributes to the body of work on language in legal systems.
Bio-note
Dr Frances Rock is a Senior Lecturer at the Centre for Language and Communication, in the School of English, Communication and Philosophy at Cardiff University. A great deal of her work to date has been concerned with the language of policing. In particular she has investigated the communication of rights in police custody and the taking of witness statements during police investigations. She is currently working on media representations of the police and policing and on non-emergency calls for police assistance. She is also embarking on research on language in prisons, particularly in restorative justice settings and on access to justice by individuals who are socially excluded. In her work, she takes an interactional sociolinguistic perspective drawing on principles and concepts from ethnographic approaches, discourse analysis and new literacy studies.
Native Language Identification for Intelligence and Forensic Applications
Krzysztof Kredens
Centre for Forensic Linguistics
Aston University, Birmingham, United Kingdom
The identification of individuals' national or cultural background based on the content they author is particularly valuable in the age of Internet anonymity. It can be useful in e.g. the profiling of producers of extremist literature or identifying the background of influencers on communities, as well as understanding social interactions and motivations online.
In this talk I will report on a study investigating the possibility of identifying the first language of individuals from the texts they author when writing in English on the Internet.
The study draws on the concept of interlanguage, which, originally developed in the context of foreign language learning and teaching, is based on the assumption that learners use a system with input from both their native and target languages, a 'third' language with its own grammatical rules, lexicalisation patterns etc. I will discuss the implications of using Interlanguage Theory outside of its natural milieu, present the methodology of extracting interlanguage features from online data, and discuss to what extent automated detection of such features is possible.
Bio-note
Dr Krzysztof Kredens received his MA in English Studies and PhD in English Linguistics from the University of Łódź, Poland. He is a Lecturer in Applied Linguistics and Director of the MA in Applied Linguistics programme in the School of Languages and Social Sciences at Aston University, UK. His academic interests include corpus linguistics, translation studies and social applications of linguistics, particularly in legal and forensic contexts. His main interest lies with the linguistics of the individual speaker and its implications for forensic authorship analysis. He is a practising public service interpreter. Recent publications include the edited volume (with Stanisław Goźdź-Roszkowski) Language and the Law: International Outlooks (Frankfurt am Mein: Peter Lang 2007) and (with Ruth Morris) '"A Shattered Mirror?" Interpreting in legal contexts outside the courtrooom', in M. Coulthard and A. Johnson (eds.), The Routledge Handbook of Forensic Linguistics (2010).
Ethics and Method in Forensic Linguistics
Lawrence M. Solan
Center for the study of Law, Language and Cognition
Brooklyn Law School, New York, USA
Forensic linguists focus attention on ways in which knowledge about the workings of language can assist the legal system. And indeed such knowledge has already assisted the legal system, in areas as diverse as the detection of plagiarism, author identification, speaker identification, the misappropriation of trademarks, and the comprehensibility and meaning of authoritative legal documents. All of this attention is the result of a half century of productive research into the nature of the human language faculty, and more recent efforts to apply the fruits of this research to legally-relevant settings. All of this creates an exciting environment for forensic linguistics for the decades to come. With that excitement, however, comes a conflict of interest that must be recognized and dealt with responsibly. The enthusiasm that accompanies growing attention from the legal system can mask the fact that linguists must always first be committed to the scientific method, and insulate themselves from the biases that accompany work conducted in an effort to support particular results in a legal dispute. Basic ethical practice, endorsed by forensic linguists, demands that linguists remain independent of the goals of the parties that seek their expertise. Yet this independence must manifest itself in more than the linguist's proclamations of integrity. Rather, it must rest in the methods that linguists employ, just as the integrity of medical research must rest in the use of appropriate procedures in that domain. Fortunately, forensic linguistics is beginning to develop methods that are likely to lead to results that can stand up to scrutiny as scientifically valid and reliable. This paper will review some of these developments, and suggest fruitful directions for future research into legally-relevant areas of linguistic knowledge.
Bio-note
Lawrence Solan is the Don Forchelli Professor of Law at Brooklyn Law school. He holds both a law degree and a Ph.D. in linguistics. His scholarly works are largely devoted to exploring interdisciplinary issues related to law, language and psychology, especially in the areas of statutory and contractual interpretation, the attribution of liability and blame, and linguistic evidence. He is director of the Law School's Center for the Study of Law, Language and Cognition. His books include The Language of Judges, and Speaking of Crime (written with Peter Tiersma). His new book, The Language of Statutes: Laws and their Interpretation, will be published by the University of Chicago Press in 2010. He has authored numerous articles and book chapters, and regularly lectures in the United States and in Europe.
© INSTITUT DE LINGÜÍSTICA APLICADA - UNIVERSITAT POMPEU FABRA