| Project Supervisor | Chantal Gratton |
| Institution & Department | Queen Mary University of London – Department of Linguistics, School of the Arts |
| Research Area | RA3: Language, Culture and Education |
| Project Start Date | End of June/ early July 2026 – flexible start date offered. |
| Project Duration | 3 months |
| Application Deadline | 4th June 2026 |
| Working Pattern | Please discuss and agree on Working Patterns with the Project Supervisor. |
| Working Arrangements | Hybrid |
| Inductions and meetings will ideally be in-person, but the majority of work can be remote, which can be discussed. | |
| How to Apply | View Guidance Here |
Project Description
Click to View More
How does our production of linguistic sounds convey our feelings and attitudes about who we are and who we are addressing? Interlocutors rarely verbalise propositions such as “We have opposing views” or “That makes me angry,” yet listeners routinely infer such stances from the sound of a speaker’s voice. This project investigates how vowel space variation indexes embodied emotion in spontaneous interaction. It asks how speakers’ fine-grained phonetic choices reflect affective states such as happiness, frustration, or sadness, problematising the traditional assumption that vowel space realisations primarily index group-level identities or articulatory clarity by showing that such variation is also a moment-to-moment resource for expressing stance, arousal, and motivational orientation in interaction.
The project builds on pilot work using a 67-hour audiovisual corpus of dyadic game-play interactions from pairs of friends playing a cooperative video game. That pilot, which examined a single pair, found a robust relationship between emotion and vowel space size: high-intensity negative affect was associated with expanded vowel space, while high-intensity positive affect was associated with contraction. The current project extends that work across the full corpus in order to test whether the pattern generalises across speakers, dyads, and interactional contexts.
The project sits at the intersection of sociophonetics, interactional linguistics, and the study of embodied emotion. It contributes to current debates in sociophonetics by showing that vowel-space variation may be shaped not only by stable social identities or speech style, but also by transient embodied and affective states. More broadly, it aims to produce a rigorous multimodal model of how affect is enacted in naturally occurring interaction, combining transcription, emotion coding, acoustic analysis, and statistical modelling.
The project is already well developed methodologically and has a clear pipeline in place, making it suitable for a short internship while still offering substantial research value. A PhD-level RA would contribute to the higher-level aspects of the workflow: data management, checking and refining analysis-ready data, validating scripts and pipelines, supporting acoustic analysis, and assisting with the integrity and reproducibility of the project’s output. This would directly strengthen the project while also giving the student experience on a live interdisciplinary study with clear academic outputs.
Internship Details
This internship would be best suited to a student who enjoys detailed technical work and wants experience contributing to an active research project rather than designing an independent study from scratch. The role would involve working with existing data, established coding conventions, and a clearly defined analytical pipeline.
The working environment would be collaborative and supportive. The PhD RA would be part of a wider team that includes the PI and two MA-level RAs, with different roles distributed according to experience and project needs. The PhD RA would not be expected to carry the bulk of transcription or emotion coding, but rather to contribute to the more advanced stages of the workflow where their doctoral-level training can have the greatest impact.
Because the data include audiovisual recordings of naturally occurring interaction, the student should be comfortable working with sensitive material and with the ethical responsibilities that come with such work. They should also be prepared to ask questions when needed and to document their work carefully so that processes are transparent and reproducible.
The internship could be structured flexibly as either a 3-month full-time role or a 6-month part-time equivalent, depending on student preference and availability. The project would be suitable for someone seeking a substantive but bounded research experience that offers technical development without requiring a long-term commitment.
It is also designed to be accessible to a doctoral researcher with strong general research capacity and an interest in learning relevant tools and methods. The internship would include training and close supervision, so the most important qualities are reliability, curiosity, and a willingness to engage with the project’s technical and analytical workflow.
Internship Structure
The PhD RA would support the project’s later-stage data preparation and analysis workflow, with a focus on high-level technical and analytical tasks rather than bulk transcription or emotion coding. The projects current MA-level RAs will undertake most of the transcription and emotion annotation, while the PhD RA will help ensure that the resulting dataset is accurate, well structured, and ready for analysis.
The student’s activities would likely include the following:
First, they would review and quality-check transcription and annotation files, helping to identify inconsistencies in ELAN tiers, timestamps, naming conventions, and file structure. They would support the conversion of ELAN materials into analysis-ready formats, including Praat TextGrids and other derived data files.
Second, they would assist with the management and validation of acoustic processing pipelines. This would include checking existing Praat scripts, confirming that batch-extraction procedures are working correctly, troubleshooting errors, and helping to ensure that output files are consistently generated and stored.
Third, they would support data cleaning and preparation for statistical analysis. This could include reshaping datasets, checking for missingness or duplication, confirming token-level alignment between transcription, coding, and acoustic outputs, and helping prepare tidy analysis files in R.
Fourth, depending on the student’s interests and experience, they could contribute to preliminary statistical checks, exploratory analyses, or visualisation of patterns in the data. This might include generating summary plots, testing code for mixed-effects models, or supporting reproducible analysis workflows.
If useful and appropriate, the student could also receive additional training in transcription or emotion coding so that they understand the annotation system and can contribute to reliability checks or targeted re-coding where needed. However, this would be secondary to their main technical role.
The intended outputs for the internship would be concrete and manageable: a cleaned and checked analysis dataset, validated script pipelines, documented data-management procedures, and a short internal summary of the work completed, including any issues identified and resolved. A strong additional output would be a reproducible workflow document or code notebook that future team members can use. If appropriate, the student could additional contribute to further disseminations (talks/articles) if involved in the late-stage analysis
Anticipated Benefits for the Student
This internship would offer the PhD student meaningful development in both advanced research practice and transferable professional skills. The core development would come from working within a live interdisciplinary project that combines corpus management, acoustic phonetics, multimodal annotation, and statistical analysis.
On the research side, the student would gain experience in handling complex naturalistic data, including the practical challenges of aligning audio, video, transcription, and annotation files across multiple software environments. They would also become familiar with scripting and batch-processing procedures used in phonetic research, including automated acoustic extraction and dataset preparation for statistical modelling.
If they take on some analytic support, they would gain exposure to mixed-method and quantitative research design, including how to structure data for mixed-effects modelling and how to interpret outputs in the context of an empirical sociophonetic question. They may also develop a stronger sense of how theoretical claims are grounded in careful data preparation and verification.
The project also offers significant transferable skill development. These include project organisation, attention to detail, version control and file management, documentation of procedures, problem-solving across software platforms, and collaborative communication within a research team. Because the role involves supporting a project that is already underway, the student will also experience how to work to deadlines, prioritise tasks, and maintain consistency across multiple datasets and research outputs.
Depending on their background, they may also gain familiarity with emotion coding and transcription conventions for audiovisual data, which could broaden their methodological skillset and be valuable for future work in linguistics, psychology, speech sciences, or related fields.
Overall, the internship would give the student experience that is highly relevant to doctoral research: working independently within a defined remit, contributing to a high-quality project, and seeing how technical data work feeds directly into publishable analysis.
Skills, Experience and Knowledge Requirements
Essential Requirements:
- Should be comfortable working carefully with data
- Willing to learn project-specific procedures
- Strong organisational skills, good attention to detail
- Ability to follow and document technical instructions reliably.
- Must be able to work with care, confidentiality and professionalism.
- Showing that they can manage files securely and follow data-handling procedures consistently,
Desirable Requirements:
- Any experience with coding or data cleaning would be an advantage, particularly if the student is interested in the more technical aspects of the role.
- Ability to work independently once trained, and confidence in troubleshooting small technical issues
- Interest in interdisciplinary work across phonetics, sociolinguistics, and emotion research.
- Familiarity with qualitative coding, transcription conventions, or reproducible research practices would also be useful.
- Familiarity with one or more of the following would be beneficial: ELAN, Praat, or R.
Prior experience with transcription, annotation, corpus data, or acoustic analysis would be helpful but not required.