At times, I find I must choose whether to write a blog post or work on manuscripts that are in various stages of development. For obvious reasons, I usually choose to work on manuscripts.
Typically, once the manuscripts that I am first author on are submitted for review, I post them on my CV page (e.g., this one). Today, I decided to post a manuscript (with a supplement) that is not yet under review. It’s currently in the hands of some co-authors, so I will post a revised version when I get their comments and prepare it for anonymous peer review.
This new paper is on the relationships between abstract phonological consonant categories (e.g., /p/, /b/, /t/, and /d/, as in ‘pin’, ‘bin’, ‘tin’, and ‘din’) and the measurable acoustic properties of produced consonants. The take-home points are:
- The mapping between categories and acoustics is many-to-many and very complex (more complex than one might gather from reading acoustic phonetics papers, even).
- The common assumptions of homogeneity of variance and statistical independence between acoustic measures are usually wrong.
- Syllable position (e.g., ‘pea’ vs. ‘eep’) modulates the acoustic structure of phonological categories in a number of complicated ways.
I’ll have another new manuscript up soon, this one on various theoretical and statistical issues with GRT model identifiability. I also plan to write a long(ish) post about the first paper in the new issue of Language (TOC pdf) before too long, since it’s a compelling mix of an interesting empirical pattern and either sloppy or nefarious statistical treatment, which is turning out to be the kind of thing that gets me going, blog-wise.
