While we seem to be on the topic of the L2 English phonology of our south american kin, I wanted to bring up an observation and see if others had noticed anything similar elsewhere: some speakers of Brazilian Portuguese, who epenthesize [i] in consonant clusters (e.g. ‘gifit’ for ‘gift’) appear to also *drop* [i] word-finally (e.g. ‘unhealth’ for ‘unhealthy’ and ‘Chomsk’ for Chomsky). A possible explanation seemed to be in terms of a hypercorrective deletion rule.
If I’m remembering this correctly, Brazilian Portuguese (and perhaps also European Portuguese?) has a pretty regular vowel devoicing rule in the context [-voi] __ # — the context may be broader than that, and it may only affect high vowels. The result can often be indistinguishable (to nonnatives, anyway) from deletion. So I doubt it’s hypercorrective, but who knows.
I think I might agree with Andrew here. BP also has a productive process of word-final [i] paragoge (and apparently in other places, as Andrew says). When I was learning BP, our teacher instructed us that at least word-final stops are followed by [i]. And this of course leads to palatalization, so a name like Fred (this was a textbook example) becomes [fredʒi].
There’s a reference on this:
Major, R. (1986). Paragoge and degree of foreign accent in Brazilian English. Second Language Research 2, 53-71.
So, since this is such a productive process in BP, it seems plausible that speakers are aware of it, and when speaking English, delete final [i] just to be sure that they didn’t put it there themselves. Or, something like that.
Lisa’s comment is not inconsistent with the doubt I expressed about classifying this as a case of hypercorrection. So long as word-final [i] paragoge (hey, I learned a new word!) precedes the vowel devoicing process I mentioned, invocation of hypercorrection is unnecessary.
Of course, my alternative falls apart if there are any examples of the [i] deletion process in contexts other than those in which vowel devoicing is otherwise expected. The two examples that Andrew cites are both ones in which devoicing is otherwise expected.
Thanks both of ya! The reference, the new word, and the fragments of analysis will lead me to think more about this.
And now, on a more blatant “cry for help”: I have a student that is wildly interested IN L2 phonology and was wondering if anyone knew any locus classicae (or at least, loci interestingae) on the topic, particularly at the segmental level?
Best, AIN
I think the hypoyhesis is quite good, but I would add a twist. Maybe not only the avoidance of final [i] insertion, but also to conform better with more typical english phonotactics having syllables with long codas, which are very uncommon is BP.
About the paragoge, I would say it is an L2/loanword phonology only.
Another pronunciation I have seen for Chomsky: [‘tSjo.mis.ki], this one conforming better to BP phonotactics.
LO
One thing that must be said about BP English is that words like ” gift” are produced with thwo epentheses, one following [f] and the other one following [t]. I haven´t come up with data such as [gifit], with only one epenthesis, once both [f] and [t] are not allowed by our coda inventory (I say ” our” because BP is my L1). Anyway, the idea of hypercorrection is a plausible (and very probable) one – this because structures showing a final vowel would be less marked in terms of syllabic structre. I agree with Leonardo´s idea on conforming better to the English syllables. I have great interest in this issue, once English coda acquisition by BP speakers is the theme of my PhD thesis here in Brazil. My analysis has been developed under Optimality Theory. Let´s keep talking about it!