Saturday, January 27, 2007

Stress and Confusion

A few weeks back (in two posts) I addressed a topic related to AmE post-stress intervocalic flapping of [t]. There are some complications to this rule. Some pure examples of this flapping are the AmE pronunciation of wattage batting and fated. I use these examples because on each side of the [t] there is a vowel with no complicating possibility of syllabic [ɻ] or [l] after the [t].

Let us consider another source of complication. The rule as commonly formulated specifies a post stress position. This can explain the lack of flapping in words such as meticulous baton and rotund. There are however examples of pre-stress flapping. The relevant contexts must be expanded to include word boundaries.

We consider a post-stress intervocalic [t] at a word boundary in the following sentences and the flapping rule appears to be at work.

"I'll jump when the mat is in place."
"Put it there please."
"Did you see how fat Elvis got?"

The first sentence has a clear primary stress on the word "mat." The second is not so clear. The primary stress is probably on "there" and I would argue that secondary stress is on "put" so the sentence comprises two trochaic feet. The third sentence might have a clear stress on "fat" but it could also have a clear stress on "Elvis". (depending perhaps on the established discussion topic: if it was already about Elvis, "fat" would probably get primary stress; if the discussion was about fat people, "Elvis" would likely get primary stress.)

This last sentence is evidence that even before a stressed vowel a [t] might undergo flapping. Consider a possible use of emphatic stress when the second syllable of "waited" is misunderstood.

"No...not 'wait-ing' I said 'wait-ed'"

The [t] in these examples may have dual citizenship. The words could be pronounced either with or without the flapping and either would fit into AmE flapping pattern; if we suggest that the flapping depends on syllabification. According to my ear (not the best test, but I trust it) a very common word with a flapped [t] subjected to this alternation of emphatic stress would more likely retain the flapping--I would expect "water" or "butter" to retain it more often than "wattage" or "waited" would with the same emphatic iambic stress.

Phrases like "nothing at all" or "write everything down but only once" show pre-stress flapping even when emphatic stress doesn't change the prosody.

The words "retail" "detail" encourage us to consider the gradable analysis of stress. The un-flapped [t] in these words might be a result of spondaic stress. Because there is clearly a lesser stress on the 2nd syllable of those words we could argue that the environment does not provide the stress differential necessary for flapping. We might also go back to look at syllabification. In either case we should go back to the original rule and scratch stress as a clear binary environment.


  1. Are these observations for English only, or do they work for other languages and merely an observation on phonemes regardless of word?

  2. More specifically these are observations for American English. RP in British English doesn't flap like AmE does. These phonological alternations will change from language to language and even from region to region.

    Of course it's a process that you might find in some other languages. Intervocalic assimilation is a very natural phonological rule. But once you get down to the specifics of environment and distinctive features you start finding all sorts of differences.

    Take for instance the voicing of [s] to [z] in some Spanish accents. There is no voice assimilation of a [s] between two vowels but there is regressive voice assimilation before a voiced stop such as [g] or [b].


Thanks for reaching out.

You can also contact me at wishydig[at]gmail[d0t]com.