“In-between speech”: the role of disfluencies and non-speech vocalisations in characterising speakers for forensic purposes

Applications are invited for an Open-Oxford-Cambridge AHRC DTP-funded Collaborative Doctoral Award at the University of Cambridge, in partnership with the National Crime Agency (NCA).

Available from October 2023, this fully-funded studentship includes: payment of university fees throughout the funded period; a tax-free maintenance grant for your living costs at the UK Research and Innovation (UKRI) minimum rate (£17,668 for 2022/23); an additional AHRC stipend of £550 per annum to support costs incurred by working with the partner organisation and a further £700 to support travel costs for the one-month attachment to the agency. Further details about the benefits of an Open-Oxford-Cambridge AHRC DTP award are available on the DTP’s studentships page.

Closing date: Thursday 1 December (12:00am, midnight, UK time)

Project overview

To date, the majority of research on speech variables for use in the comparison of speakers on crime-related speech recordings (forensic speaker comparison, FSC) has focused on aspects of vowel/consonant sounds or long-term properties such as pitch. Considerably less research attention has been paid to the vocalisations “in between speech”, specifically disfluencies and non-speech vocalisations (NSVs).

Research into the speaker-characterising properties of disfluencies (e.g. filled and silent pauses, repetitions, sound prolongations and self-interruptions) is in its infancy. Promising findings of consistency across occasions (Braun and Rosin 2015), better levels of speaker discrimination from filled pauses than lexical vowels (Hughes et al. 2016), and speaker differences in disfluency rates and profiles (McDougall and Duckworth 2017, 2018) demonstrate further investigation is warranted. There remain significant unanswered questions regarding the speaker-specificity of disfluencies in larger populations, different languages/dialects, different speaking styles, multilingual contexts, etc.

Research into the speaker-discriminatory potential of NSVs (laughs, coughs, yawns, sighs, screams, breath noises, etc.) is sparse. There are a few promising results available on the speaker-specificity of e.g. laughter (Bacharowski et al. 2001), breath noises (Link 2012) and screams (Engelberg et al. 2019), but detailed research is crucially needed to determine ways in which such individual variation can be exploited for FSC.

The present project offers the opportunity to conduct Ph.D. research into “in-between speech” and its forensic relevance. Candidates should propose a project that investigates an aspect of the relationship between disfluencies or NSVs and speaker identity. The candidate will initially investigate existing speech databases for their suitability for researching the speech feature(s). It is likely they will also develop materials and record speakers to produce a new database specifically designed for research of the target feature(s). Analytical methods will include auditory, acoustic, computational and statistical techniques.

The studentship will include opportunities to collaborate with the NCA in defining and shaping the scope of the study. Beyond this the NCA will provide potential opportunities to develop wider non-academic competencies around client relationship management, cross cultural presentation/communication, working at pace and an understanding of policy making.

Supervision

The student will be jointly supervised by Dr Kirsty McDougall, Theoretical and Applied Linguistics Section, University of Cambridge and the Departmental Head of Biometrics at the NCA. Subject to security and HR procedures, the student will make regular day-visits to the NCA and will undertake a one-month attachment with the agency. Training opportunities relating to speech and biometrics at the NCA will also be made available to the student.

How to apply

To meet the requirements of the CDA partner organisation for those working at their premises and with their data, applications are open only to those who meet UK Civil Service nationality standards, i.e. UK nationals; nationals of the Republic of Ireland; nationals of Commonwealth countries; European Economic Area (EEA) nationals with (or eligible for) status under the European Union Settlement Scheme (EUSS); relevant EEA or Turkish nationals working in the Civil Service; relevant EEA or Turkish nationals who have built up the right to work in the Civil Service; certain family members of the relevant (European Union) EU and Turkish nationals.

The successful candidate will also need to fulfil the requirements of a Security Check at DV Level, which requires residency in the UK for three out of the last seven years (please see the UK government website for further information about security clearance) and enhanced NCA checks against crime records.

Applicants should have a background in experimental phonetics and/or forensic speech science, and hold a Masters qualification in linguistics (phonetics), forensic speech science or a closely related area. Experience of recruiting and testing experimental participants and statistical analysis is desirable. Applicants should meet the eligibility criteria for Open-Oxford-Cambridge AHRC studentships.

For informal enquiries about the opportunity and how you might frame your approach to the project, please contact Dr Kirsty McDougall (kem37@cam.ac.uk).

You should apply to the Ph.D. in Theoretical and Applied Linguistics at the University of Cambridge using the Postgraduate Applicant Portal by Thursday 1 December 2022 (12:00am, midnight, UK time), indicate your interest in being considered for Open-Oxford-Cambridge AHRC DTP studentship funding and upload a completed copy of the OOC DTP Application Form at the same time. For further details on how to apply for this CDA through the University of Cambridge, please see the advert on the Cambridge jobs site.