Field Project Report 3 for Ling 600.01 Phonetic Theory
Report due by the beginning of class on Tuesday of Week 5.
This report on "The Vowel Space" should be about two or (at most) three single-spaced pages, exclusive of references and appendices.
Part 1a: Vowel chart table. Make a table that is organized along the lines of a traditional vowel chart, with rows for contrasts along the dimension of features such as [high] versus [mid] versus [low], and columns for contrasts along the dimensions [front] versus [central] versus [back] and (where relevant) [spread] versus [rounded] as left and right "sub-columns" dividing up the relevant front-back column(s). Fill in each cell of the table with transcribed wordforms (and glosses) exemplifying the vowels in contrast, so that the reader can see at a glance the justification for your claims about the vowel inventory.
(If you decide to record major contextually conditioned allophonic variation as well as the inventory of robustly contrasting vowels, decide on a convention for indicating sets of vowels that are in such an allophonic relationship. For example, you might decide to designate one member of each set as "the prototypical value for the phoneme" and then put parentheses around the symbols for every other member, along with a footnote about the set and the conditioning environment.)
This table will serve as the "Materials" section of your methods.
Part 1b: Acoustic analysis. Get good clean recordings of the language consultant producing the words in the vowel chart table that you provided in Part 1. Measure the first and second formants in each production. Write a sentence or two describing your criteria for where you chose to take these measurements. (You may want to also make a figure illustrating any concepts such as "steady state" or "vowel mid point" if these were part of your criteria.)
If you want to record your measurement points in a TextGrid file, so that you can consistently extract the formants from the same point, and so on, you might want to ask in class about models for how to do this and also about how to write scripts to extract the formant values at the point automatically, so that we can write the scripts together and save them in the scripts directory under this URL.
Part 2a: The acoustic vowel chart. Plot the vowels in a hertz space. If you put your measurements and an ASCII representation of the vowel in a tab-separated plain txt file called vowels.txt, you can use the plotVowels.R script to make this chart very easily in R.
Part 2b: The psycho-acoustic vowel chart. Convert your hertz measurements into ERB and plot the vowels in ERB space. (The plotVowels.R script includes a command with the formula for making the conversion, which is very easy to apply if you put your measurements and an ASCII representation of the vowel in a tab-separated plain txt file called vowels.txt and read it into R to do the plotting.)
Part 3: Discussion and conclusion. Write a few sentences summarizing your observations about the language's vowel inventory, and comparing the relative usefulness of the two phonetic representations that you provided in Part 2.