Research

Phonetic Expression of Glides & High Vowels via Realtime Auditory Feedback

2020, Dan Burgdorf

dan-ze-figureDan is currently working on experiments examining how glides and high vowels differ phonetically.  He is particularly interested in how to differently prioritize acoustic versus articulatory targets and feedback.  

He uses an altered feedback paradigm to see how these segments respond to altered real-time auditory feedback with unaltered somatosensory feedback, and has found that participants compensate more for vowels than glides either by adjusting formants or, unexpectedly, by lengthening the vowel to be longer than the applied alteration.

Dan also works with the Burmese language.  He has conducted field phonological examination and production experiments with native speakers and found evidence for a tonally-based weight distinction within the major syllables


Bidirectional Syntax-Phonology Interfaces in Italian Word Order

2020, Katherine Blake

A long-standing issue in linguistics is how the various subsystems of language interact. In this work, I examine the syntax-phonology interface more specifically, finding evidence for a bidirectional relationship between these systems rather than a unidirectional influence of syntax on phonology. Using corpus methods, I probe phonological markedness effects on word order in Italian {noun, adjective} pairs. Specifically, I look at effects of stress clash, phonological weight (here length, as in Heavy NP Shift), and vowel hiatus. I predict that the order of {noun, adjective} is manipulated to avoid stress clash within pairs, light-final NPs, and instances of vowel hiatus within pairs. Evidence of these phonological effects on word order in Italian supports my hypothesis that there is bidirectional influence at the interface between syntax and phonology.


Predicting the amplitude envelope with interpretable models: Testing whether stress and sonority manifest as intensity variation

2020, Kaelyn Lamp

amplitude envelope for [zl]Phonological properties like lexical stress and sonority are both said to have an effect on intensity.  However, it is not clear how consistent and predictable these effects are.  This project focuses on training neural networks and more interpretable models to predict the amplitude envelope (a metric of intensity) with stress and sonority information to more comprehensively understand if and what predictable effects lexical stress and sonority have on intensity.

 

 

 

 

 


Stimulus timing effects on SVO sentence generation

2020, Sam Tilsen

sentence-generationIn everyday settings, our attention to events in the world and to the entities which participate in those events is unlikely to be simultaneous or evenly distributed in time. This nonsimultaneity and nonuniformity may have consequences for how speakers organize and produce utterances. By manipulating nonsimultaneity of experience in a tightly controlled experimental task, we can learn something about how speakers organize syntactic and conceptual systems when producing utterances. In experiments for this project, the relative timing of Subject, Verb, and Object stimuli is manipulated and we analyze the effects of stimulus timing on how speakers produce SVO sentences.

 

 

 

 


Localizing speech categories in space and time

2020, Sam Tilsen, Claire Wang

infoscalogramThe speech system and our measurements of its outputs—articulator positions and acoustic signals—are very high-dimensional. In contrast, our theories posit low-dimensional categories like phonemes and gestures. How do we determine, from articulatory and acoustic signals, where in time there is evidence for those categories, and which signal dimensions contain that information? This project uses deep neural networks to estimate the temporal and spatial distributions of category-related information.

 

 

 

 

 

 


Motoric mechanisms for the emergence of non-local phonological patterns

2019, Sam Tilsen

motoric-mechanismsNon-local phonological patterns can be difficult to analyze in the context of speech production models. This project develops a dynamical model of spreading and agreement patterns, in the gestural framework of Articulatory Phonology. The model distinguishes between excitatory and inhibitory articulatory gestures, and encodes gestural targets in planning fields. The model is applied to various empirical phenomena: non-local consonant harmonies, anticipatory posturing before production of a word form, and dissimilatory interactions in distractor-target response paradigms. Publications related to this project:

 

 

 

 


The effects of intonation on the timing of lexical tones

2017, Hao Yi

This dissertation investigates the effects of intonation, particularly boundary tones, on the timing of lexical tones in tone languages. I ask three questions: 1) Do intonation and tone differ in principle in terms of how they are integrated into the coupling which accommodates segments, lexical tones, and intonation? 2) Does the presence of intonation give rise to differences in the coordination among tones and segments? 3) Do intonation and tone interact at the phonological level or are they implemented independent of one another? A cross-linguistic production study is planned to be carried out in Mandarin, Cantonese, Vietnamese, etc. Alignment patterns in both acoustic and articulatory data will be carefully investigated.


Real-time MRI of the vocal tract

2015, Sam Tilsen, Pascal Spincemaille, Peter Doerschuk, Wenming Lu, Yi Wang

This project uses a real-time magnetic resonance imaging to investigate the dynamics of speech articulation. Real-time MRI is the technology of the future for articulatory imaging, offering a unique combination of spatial coverage, high spatial resolution, and high temporal resolution.
mri_example_0

 

 

 


Patterns of Misperception of Arabic Consonants

2015, Chelsea Sanker

A perception experiment looking at patterns of which Arabic consonants are most confusable for listeners with different language backgrounds and whether confusions between pairs of consonants are symmetrical or strongly directional. Participants who were native or non-native speakers of Arabic listened to short nonce words played in background noise and chose the written form which they thought best matched the auditory stimulus. The patterns of identification errors suggest which acoustic features are salient to listeners, based on which sounds are most frequently confused and how identifications are influenced by the primary frequencies of the masking noise.


Mechanisms of selection and coordination in speech production

2014, Sam Tilsen

This research project involves experimental investigation and computational modeling of two key motor processes in speech production: selection and coordination. Selection relates to how speakers control the choice of which articulatory gestures to produce, and coordination relates to the relative timing of articulatory movements.selection_coordination

 


Earmaster: a phonetic transcription game

2013, Sam Tilsen & Sam Hyatt

Supported by a Cornell Center for Teaching Excellence Faculty Grant, we are developing an instructional game for practicing IPA phonetic transcription in html/Javascript. The game allows for modular development of transcription exercises, immediate feedback, gameplay logging, and gameplay analysis.


Syllables and Segments

2012, Adam Cooper and Draga Zec

Our goal is assembling a comprehensive database on the typology of syllables in the world's languages. We detail for each entry language aspects of its phonological system relevant for the structure of its syllables -- phonemic inventory, syllable shapes, including the details of segment distribution across syllable internal positions, syllable weight, tone and stress systems, etc. We are currently developing cross-linguistic typologies of syllabic consonants and syllable codas.


Phonetics and phonology of pitch accent explored on the Štokavian dialects of Serbian/Croatian

2012, Draga Zec and Elizabeth Zsiga (Georgetown University)

In this project we analyze cross-dialectal phonetic variation of pitch accent in the Neo-Štokavian dialects, focusing on its systematic aspects, and on the emerging phonological patterns. In particular, we explore two types of dialects: those in which pitch and stress, the components parts of pitch accent, tend to occur on the same syllable, and those in which the two components tend to occur on distinct, most notably, contiguous syllables.


Partially nasal segments

2011, Abby Cohn & Anastasia Riehl (Queen's University)

Phonological and phonetic investigation of partially nasal segments, with particular focus on Austronesian languages


Relationship between phonology and phonetics

2007, Abby Cohn

Nature of representations and patterning of sounds as often characterized as phonological or phonetic, with a particular focus on distinctive feature theory