3 Colors of Velar Palatalization

This project probes, both, the nature of the consonantal changes involved in palatalization/velar softening and of the causal forces that underlie them. The motivation is to understand exactly when palatalization is actually phonological in nature – i.e. when it involves proper phonological computations in the brain, as opposed to more substantive phenomena. The change can be viewed as a straightforward case of feature insertion/change, but given that more often than not the change is triggered in the context of front vowels a phonetic explanation is almost always possible. Viewed from the perspective of Substance-Free Phonology, then, the primary task is to decide (a) whether any given instance of palatalization is actually phonological, and (b) when they are not, what other accounts can be called upon? Tempting as it is to appeal to co-articulation, one problem inherent to this approach lies in the fact that while there are many known instances of velars becoming apicals in the context of front vowels, there aren’t any known examples of the opposite happening (at least not to my knowledge)!

This project started in 2014 as a purely behavioral investigation of velar palatalization in Malayalam at the MARCS Institute of Brain, Behavior and Development in Sydney, Australia. Malayalam is particularly interesting for two parallel reasons. First, the set of triggering vowels in Malayalam involves [i,e,a]. The inclusion of [a], a phonetically back vowel in Malayalam, raises the possibility of investigating the nature of the phonology-phonetics interface. The standard assumption, found for instance in Mohanan and Mohanan (1984), is to assume language-specific phonetics – [a] is phonologically [-Back] even though its phonetic signatures might be those of back vowels. Alternatively, viewed from the perspective of Substance-Free Phonology, one could also argue that [a] is, in fact, [+Back] vowel, but the Malayalam rule is triggered by the class of [-Round] vowels. In this instance palatalization is not necessarily triggered by front vowels, because phonological rules are not beholden to phonetics. Second, the Malayalam process manifests numerous ideolectal and dialectal aberrations which Mohanan and Mohanan (1984) simply ascribe to “lexical exceptions”. This is not an optimal solution, to say the least. Moreover, viewed from the perspectives of Government Phonology any process with so many exceptions do not pass the phonological litmus test anyway. On the other hand, the same exceptions also make it hard to appeal to a purely substantive/phonetic explanation. A third possibility, then, is fossilized diachrony!

I have been on this trail for almost six years now… and several conferences talks, proceedings, and a (forthcoming) journal article later, I am now exploring possible ways to bring electrophysiological evidence to bear on this issue. As usual, much of this current approach is inspired by Tobias Scheer’s phonolEEGy program!