Open projects for possible collaboration
In the Whova event page for the meeting in Laboratory Phonology, I started a thread with the title "How are better collaborations created?" The goal of this was to really ask the question of "what works?" with labphon-related projects that involve multiple people and institutions. I suppose there is another kind of guilty reason - I have several things I've worked on but many things that are at various earlier stages of development. It would be great to see people tackle some of these types of projects and also be involved with other things along the way.
I received feedback from eight people: Jen Nycz, Anne Pycha, Valerie Freeman, Paul De Decker, Miao Zhang, Bihua Chen, Ivy Hauser, and Timo Roettger.
Paul started by asking whether it is really clear if people want to collaborate. Is there a mechanism that we can think of to make this known to others? Valerie suggested a "collaborator's corner" at conferences with skills/preferences for a particular project. She also mentioned a resource for ensuring that collaborators are on the same page with regard to goals. Jen mentioned how each person has their strengths/preferences in research projects and that we might try to match along these preferences. This way we would be truly aiming to find not just collaborators, but ideal collaborators where all parties benefit. Ivy mentioned that more intentional networking at conferences might serve some of these goals. Timo's idea involved a special session proposal for a conference (maybe the next LabPhon?).
I like all these ideas. I think there are some separate threads:
(a) Identification. We could identify what we're doing and discuss our project goals with others. Maybe this is the collaborator's corner that become part of the networking process at conferences?
(b) Needs/Wants. We could focus on really identifying what we would like with each of the projects we are working on. Is it in the idea stage? Is the data already collected? Is the data already annotated? Is the data ready for analysis? Where are you stuck and what would you like to collaborate on?
(c) Goals and agreements. As per Ivy's point, each project could have a timeline and set of goals that collaborators agree upon. Is the project part of a larger project? Do you want to submit a paper this year? Next year? What about author order in submission? Will the collaboration continue or end at a certain point? Who is responsible for managing goals?
With these in mind, I'm going to try to identify some of my own projects that are seeking collaboration.
1. Speech rate and lenition in Spanish
Back in 2010, I collected a set of recordings from 9 young Oaxacan Spanish speakers (ages 19 - 26). They produced a short read passage (Sleeping Beauty), a retelling of a narrative after a short video (the pear story), and a free narrative. The initial goals of this project were to examine speech rate variation across speech styles across different dialects of Spanish. The cross-dialectal goal did not work out, but the data remains.
Team: Myself (UB Department of Linguistics, Colleen Balukas (UB Romance Languages and Literature), Jamieson Wezelis (UB Romance Languages and Literature)
The current goals of this project are rather open, but we have considered three topics:
a. An exploratory study on vowel sequences and vowel hiatus patterns across word boundaries. There is a literature on this topic in Spanish phonetics, but not with spontaneous speech data (and certainly not across speech styles).
b. An exploratory study on aspects of vowel reduction in Oaxacan Spanish.
c. An exploratory study on patterns of vowel devoicing in Mexican Spanish.
The eventual goal would be one (or more) papers on the acoustic phonetics of spontaneous speech in Spanish.
The current state: All the recordings have been trasnscribed in ELAN and force-aligned. The read passages have also now been hand-corrected. All recordings have also been syllabified using a custom Praat script. However, Jamieson can no longer be actively involved in the process of hand correction of the data.
An ideal collaborator is (1) interested in helping with the remaining hand-correction of the acoustic recordings (roughly 1.5 - 2 hours worth), (2) is either interested in one of our goals or has their own which we could all pursue once the alignments are corrected, (3) has some knowledge of statistics as it applies to analyzing acoustic phonetic data, (4) is interested in delving into some of the literature in Spanish phonetics (lots of dissertations), and (5) is literate in Spanish.
Timeline: We're kind of stuck right now (no progress for about a year), but we can devote some time to this starting in the next semester. It would be great to see results in 2023 (a talk, a paper, etc).
Bonus: I'm open to data sharing after collaboration.
2. Glottal reduction in Itunyoso Triqui
Throughout the course of my language documentation and phonetic data analysis grant, we collected about 35 hours of spontaneous speech in Itunyoso Triqui, an Otomanguean language spoken in Oaxaca, Mexico. Triqui languages are rather tonally complex and have orthogonal contrasts involving glottal consonants (/ʔ, ɦ/). While there is some description of glottalization in the language (DiCanio 2012), there is an open question as to how much lenition of glottal stops occurs. The goal would be to analyze the acoustic data to examine variation in the production of the glottalization. We are particularly interested in variation in glottalization as a function of word position (VCV vs. VC#) and contrast type (pre-glottalized sonorant vs. glottal stop). This project would tie in nicely with recent work on Hawaiian glottal stops (Davidson 2021).
Team: Myself (UB), Lisa Davidson (NYU), Richard Hatcher (postdoc, Hanyang University - former UB grad student)
The current state: All of the recordings are force aligned with a custom-built aligner for Triqui. The recordings of interest have also been hand-corrected. We have begun some analysis of variation in production of the glottalization using a script I wrote for Praat which allows users to identify glottal reduction types. We presented preliminary results from this work at Haskins Laboratories in Fall 2021. We would like a collaborator to help us analyze more of the existing data.
An ideal collaborator is (1) interested in non-modal phonation type in complex tone languages, (2) has some knowledge of the phonation literature and acoustic phonetics, (3) is familiar with running voice quality scripts in Praat (or at least scripts), (4) has some knowledge of statistics as it applies to analyzing acoustic phonetic data, (4) is interested in judging patterns of glottal reduction in field recordings, and (5) would like to get involved with work on phonetic variation in Itunyoso Triqui.
Timeline: We have not made new progress for about a year, but some of us can devote some time to this starting in the next semester. It would be great to see results in 2023 (a talk, a paper, etc).
Bonus: I'm open to data sharing after collaboration.
3. Triqui clitic phonetics study
Certain Triqui person clitics (speech act participant clitics) condition tonal changes on the right edge of the root they attach to. This is described in the literature on the language (DiCanio 2008, 2016, 2020, 2022). Consider that the 2S clitic /=ɾeʔ¹/ conditions (1) tonal raising on certain roots, e.g. /ɾa³ʔa³/ 'hand / mano' > /ɾa³ʔa⁴=ɾeʔ¹/ 'your hand', (2) leftward, low-tone spreading on others, e.g. /ka⁴ne⁴³/ 'bathed / se bañó' > /ka⁴ne¹=ɾeʔ¹/ `you bathed', and (3) no tonal change on others, e.g. /ki³ɾi¹/ `took out / sacar' > /ki³ɾi¹=ɾeʔ¹/ `you took out'. There are two research questions here. First, there is an empirical question as to what these tonal changes look like for roots containing the 9 lexical tones. Of particular interest is the observation that, in those roots where no tonal changes occur, pre-clitic lengthening may. Second, utterance-final prosodic lengthening takes place for lexical roots (DiCanio & Hatcher 2018, submitted), but the conditions on this are quite limited (almost no lengthening takes place for roots ending with coda /ʔ, ɦ/). Moreover, is prosodic lengthening limited to roots or may it also affect clitics? The study here sought to try to answer these empirical questions for Itunyoso Triqui.
Team: Myself (so far)
The current state: This has been on hold for 4 years now. The recordings that were collected alongside this data has been analyzed (DiCanio & Hatcher 2018, submitted). The relevant stimuli were recorded in 2018, consisting of 224 trials with target words in clitic and non-clitic conditions, in both utterance-final and non-final position, repeated 5 times per speaker, with 10 speakers (11,200 sentences). This data has not yet been transcribed or segmented in Praat, though all the stimuli and their (random) order of presentation are saved in an Excel file, so transcription should be relatively straightforward.
An ideal collaborator is (1) interested in tone production and the phonetics of tone sandhi, (2) has some knowledge of acoustic phonetics and Praat, (3) has some knowledge of statistics as it applies to analyzing acoustic phonetic data, and (4) is interested doing speech segmentation work with this data.
Timeline: No work has taken place on this since the recordings were made. It's a big project given the amount of data and speakers. So, it's completely open. I imagine an analysis of the data alongside segmentation would take at least several months with a few researchers.
No comments:
Post a Comment