Phonetics/phonology workshop

Also known as 'the P-workshop'. The programme consists of talks, seminars and discussions on subjects relating to phonetics and phonology (both synchonic and diachronic) and to speech technology.

Time and place

We normally meet on Friday, 12:10-13.00 (but not every Friday).

Contact

The P-workshop is the forum of the Phonetics and Phonology Research Group. If you would like to give a talk, suggest a reading, or lead a session (or if you'd like to be added to the P-workshop mailing list), email the organisers: Patrick Honeybone, Jeremy Steffman and Oksana Lebedivna.

Current and recent events

• 30th May 2025 (12:10 - 13:00), room: G32 in 7 George Square. Oli Liu (Informatics) 'On the variability of phoneme encoding in self-supervised learning representations'. Abstract: Phonemes play a central role in the study of human speech, underpinning both theoretical frameworks in speech science and applied models in automatic speech processing. Yet, their status as perceptual units remains contested, due in large part to the lack-of-invariance problem: the acoustic realization of a phoneme can vary considerably depending on speaker characteristics, coarticulatory context, and other factors. While how humans cope with variabilities in speech has long been explored in psycholinguistics, recent advances in self-supervised learning (SSL) models provide new computational tools for studying this problem. These models have been shown to encode rich linguistic information, where phoneme categories are mostly linearly separable, while also encoding other attributes including speaker, gender, and emotion. In this talk, I examine how phoneme identity of a phone is encoded relative to speaker identity, phonetic context, and latency from phone onset in self-supervised learning models. In the first project, I used subspace analyses to probe the structure of phonetic and speaker encoding in the representation space. In the second project, I simulated decoding analyses used in a cognitive neuroscience study to investigate the temporal dynamics and context effects in the neural encoding of continuous speech. For each neural phenomenon observed in human listeners, I try to shed light on the underlying learning mechanisms and domain specificity by comparing the behavior of a few SSL models trained with different learning mechanisms and learning material.
• 23rd May 2025 (12:10 - 13:00), room: G32 in 7 George Square. Almudena Caño Laguna (Universidad Complutense de Madrid) ‘Deaccentuation in Spanish: the relationship between word stress and phrase accent. First steps in the research’.
• 2nd May 2025 (12:10 - 13:00), room: G32 in 7 George Square. Rory Turnbull (Newcastle University) 'Sudoku, recreational programming, and indeterminacy in phonology'. Abstract: In this talk, I argue that all three of the objects in the title are "puzzles" to be solved only insofar as we accept the rules of the game. Classical phonological analysis consists of determining a mapping between the observed surface form and an (abstract, unobserved) underlying form. There are typically many logically possibly ways to effect this mapping, and there are various (often implicit) theoretical rules and principles for determining which mapping is the preferred one. However, there are instances of indeterminacy where these rules fail to give us an unequivocal answer, or where the answers given do not correspond well to our understanding of mental representations of phonology. Indeterminacy is only a problem if we accept the premise of phonological analysis.

Previous events

• 22nd April 2025 (12:10 - 13:00), room: 7.01 in the Dugald Stewart Building. Gilly Marchini ‘Clustering methods in the analysis of intonation: the case of focus marking in Afro-Mexican Spanish’.
• 11th April 2025 (12:10 - 13:00), room: G32 in 7 George Square. Amalia Arvaniti (Radboud University) 'How to tame your intonation: from concepts to methods and back'. Abstract: In this talk I will provide a brief overview of the major findings and conclusions of SPRINT, a five-year project funded by the European Research Council to study intonation in English and Greek. The main objective of SPRINT has been to develop a new approach to intonation based primarily on the investigation of intonation variability and pragmatics. SPRINT started from the position that intonation is not a “half-tamed savage”, as the frequently used (and highly questionable) metaphor of Bolinger’s has it, but part of a language’s phonological component whose phonetic features are as tame as any other aspect of speech production, provided a) we treat them as such and b) we employ suitable methodologies to study them. Starting from this position, in the talk, I cover three topics: a) the main sources of variability in intonation and the methodologies employed in SPRINT to address them, so we can distinguish systematic, linguistically determined variation, from gradience, and noise; b) the role that meaning can play in this process; c) the lessons we learned from researching these topics and the ways they have shaped the main SPRINT objective, determining what we retain from AM, the most widely adopted model of intonation, and what we need to revise. Overall, the findings support SPRINT’s starting point and provide encouraging results on which to build this new foundation.
• 4th April 2025 (12:10 - 13:00), room: G32 in 7 George Square. Fae Hicks ‘Is ‘phonetic erosion’ a phonological concept: an exploration of the nature of phonological processes and causality of change’.
• 14th March 2025 (12:10 - 13:00), room: G32 in 7 George Square. Julian Bradfield 'Metaphony, Umlaut, and ǃXóõ A-Raising'. Joint work with Shanti Ulfsbjorninn, (Memorial University Newfoundland). Abstract: Metaphony and umlaut are well studied assimilations, and metaphony in particular has attracted a range of different analyses. This paper casts both in the framework of element theory, but as quite distinct processes. We then study the complex assimilation pattern of the Khoisan language ǃXóõ, between the first vowel of its bimoraic words and the second vowel, also affected by the consonants in the word. The most recent analysis (Lionnet 2018) argues that probabilistic and cumulative gradient effects are required to account for this process. We show that the process can be seen as formally a combination of umlaut and metaphony. Moreover, our analysis explains other assimilations, and predicts a new assimilation which is confirmed by recent data.
• 28th February 2025 (12:10 - 13:00), room: G32 in 7 George Square. Pia Lehecka ‘Spelling practices of high vowels in Older Scots’.
• 21st January 2025 (12:10-13:00), room: G32 in 7 George Square. Oksana Lebedivna 'Word-final depalatalisation in Hutsul Ukrainian: The case of *[tʲ] > [t]'. Abstract: This article revisits depalatalisation of the word-final *[tj] in the Southwest Ukrainian dialects, such as Hutsul, as compared to the North Ukrainian dialects, such as East Polissja, where [tj] retains its palatal component due to lexicalisation. In Hutsul, however, forms with both depalatalised and palatalised word-final voiceless stops are observed. I argue that palatalisation is preserved in the surface representations whose paradigms exhibit lexicalisation of the palatalised stop. A lack of lexicalisation in forms with [tj] can be explained by the underlying Tendency Towards Intrasyllabic Harmony Constraint and rule scattering from sound patterns of /tj/ when followed by a vowel (so-called dialectal backing). To show the intervocalic position of a word-final palatalised stop, I analyse spectrograms from Hutsul, suggesting that [tj] is followed by a vocoid. This suggests new ways of thinking about the formalisation of syllable structure in Hutsul, proposing that in Hutsul, [t] is associated with the onset and [j] with the V-slot, while an empty nucleus position is preserved by a vocoid.
• 24th January 2025 (12:10-13:00). Venue change: Due to the inclement weather and campus closure we are now meeting on Teams. Mits Ota ‘Learning, communication and the typology of phonotactic constraints’.
• 17th January 2025 (12:10-13:00) NOTE THE UNUSUAL VENUE!, room: 1.20 in the Dugald Stewart Building. Patrick Honeybone ‘English r -sandhi: where doesn’t it occur?’.
• 13th December 2024 (12:10-13:00), room: G32 in 7 George Square. Stefano Coretta & Jeremy Steffman ‘Replication of hemispheric differences in priming specificity effects: a study plan’.
• 29th November 2024 (12:10-13:00), room: G32 in 7 George Square. Benjamin Elie 'Optimal control of speech: A permanent learning approach'. Abstract: A growing number of studies support the Optimal Control Theory (OCT) view that articulatory movements in speech, similarly to non-speech movements, aim to satisfy (potentially conflicting) task requirements, including intelligibility, least articulatory effort, and utterance duration. OCT-based models of speech production assume that speech articulatory movements are planned to minimize a composite objective function defined as the weighted sum of the cost of not satisfying these individual tasks. After an introduction showing results of simulation that provide evidence that OCT can explain various aspects of speech, we will discuss how speakers learn and control optimality in speech.
• 15th November 2024 (12:10-13:00), room: G32 in 7 George Square. Alice Turk 'XT/3C: A Model of Speech Production Based on Symbolic Representations, Phonology-Extrinsic Timing, and 3 Processing Components'. Abstract: XT/3C is a model of speech articulation planning and execution based on symbolic phonological representations, phonology-extrinsic timing and three processing components: 1) Phonological Planning, 2) Phonetic Planning, and 3) Motor-Sensory Implementation. I will provide a brief discussion of the lines of evidence that motivated this model, followed by a general overview of the representations and mechanisms used in its three processing components.
• 25th October 2024 (15:10-16:00), room: G32 in 7 George Square. Matthew Sung 'Recent Developments in Computational Dialectology'. Abstract: Dialectometry is the quantitative branch of dialectology which utilises computational methods to calculate linguistic distances and generate visualisations which allow us to explore relationships between dialects. Although dialectometry is a growing field with an increasing number of new approaches, some corners in dialectal variation are still rather unexplored. Firstly, dialectometry is a popular method in Europe, but not so much in other parts of the world. It is unclear whether our findings of dialectal variation in Europe (e.g. the existence of a dialect continuum, the specific dynamics of dialect spread) are also found in other corners of the world. Secondly, most of the work on phonetic variation is based on segments, while most of the world’s languages are tonal (Yip 2002). It is unclear how dialects vary on the tonal level. Lastly, the outcome of a dialectometric analysis is usually a classification of dialects, but the features that contribute to this classification, i.e. dialect features which are exclusive to certain groups, are not explored in these classifications. In this talk, we would like to address the issues raised above based on our latest work done in Leiden.
• 18th October 2024 (12:10-13:00), room: G32 in 7 George Square. Brandon Kieffer 'Are there Rhotic Affricates in Kilimanjaro?'.
• 4th October 2024 (12:10-13:00), room: G26 in 7 George Square. Stephen Nichols 'Word-initial NC sequences in two varieties of Amuzgo'.
• 20th September 2024 (12:10-13:00), room: G32, 7 George Square. P-Workshop kick-off session – come along and meet other P-people.

For older events, have a look at our archive: