In a paper I’m currently working on, I propose a new analysis of the past tense and plural suffix alternations in English. I think that the analysis is interesting for a variety of reasons, but for the purposes of this post I’d like to focus on a particular prediction that the analysis makes and that I have preliminarily found to be correct (as I outline below). As far as I know, this prediction is a novel one; I hope that readers of this blog will either confirm or rectify my (mis)perception in this regard.
The prediction is that with verb stems ending in a postalveolar sibilant (e.g., mash, match, budge), the past tense suffix will also be postalveolar — i.e., it will assimilate in terms of the coronal subplace of articulation. At first I was concerned about this prediction, but after asking a few other native speakers about it I was reassured that the prediction was probably correct.
Sharon Rose and John McCarthy encouraged me to dig deeper with some palatography. Before getting human subjects approval for a fuller study (which has been approved, by the way, and will be conducted later this summer), I was eager for some preliminary results. So I roped Colin Wilson into allowing me to paint his tongue multiple times with a charcoal-and-olive-oil mixture, to stick an intraoral mirror in his mouth, and to take pictures of his palate. Ably assisted by Karen Shelby, we made palatograms of five key words (mast, mashed, mass, mash, and mat). Though preliminary, I think the results confirm the prediction made by my analysis.
The comparison of immediate interest is between mast and mashed. The central part of the articulation in both cases is the result of the final stop’s closure; compare with mass and mash, in which this central closure is absent. This central closure is clearly further forward in the case of mast, almost directly behind the front teeth, while the central closure in the case of mashed is further back, on the posterior part of the alveolar ridge. The clearly anterior articulation of the stop on its own is shown with the palatogram for mat.
|mast: [st]||mashed: [ʃt]
|mass: [s]||mash: [ʃ]
(The marks on the teeth and lips in these palatograms resulted from the application of the (very messy) charcoal mixture; they not a product of any of the articulations. Mixture-application tips from experienced palatographers welcome.)
I’ll eventually post the results of my larger study, as well as (a draft of) the paper I’m writing. Before ending this post, though, I need to thank Bob Kennedy for first pointing out the relevance of these examples to my analysis. Thanks also to Colin Wilson, who in addition to being my experimental guinea pig helped me to see that the prediction made by my analysis was probably the right one.
Update, June 2005: more results from this project were just presented at the First International Conference on the Linguistics of Contemporary English by my research assistant and collaborator, Cindy Kilpatrick — click here for the .pdf slide presentation.