The addition of /l/ to the waveform model of vowels
Michael Stokes · The Journal of the Acoustical Society of America · 2020
Since its introduction, the Waveform Model of Vowel Perception and Production (Stokes, 2009) has achieved human perceptual performance on the data from two of the most cited publications in the literature (Stokes, 2011, 2014). The algorithm achieving human performance dramatically supports the Waveform Model (WM) logic, but there are phonemes not included in the Peterson and Barney (1952) data. Because of the familiarity, the Peterson and Barney (1952) data are used to illustrate the WM categories, but the WM must be able to accommodate phonemes such as /l/ to continue being a successful cognitive model. There are acoustic measurements reported for /l/, and these measurements fit perfectly into category 3 of the WM making it a natural categorical partner to /r/. Importantly, F3 measurements for /l/ are much higher than surrounding categorical vowels making it easily distinguishable from its neighbors and category 3 partner. This presentation will focus on /l/, but other phonemes appear to fit comfortably within the WM categorical pair framework as well. An algorithm is objective, and a perfect demonstration and defense of logic. Incorporating additional phonemes strengthens and extends the WM logic and programming.