Arnold Zwicky's Blog

Over in Facebook, in the midst of a rambling discussion of dialect differences, Arne Adolfsen presented a phonological puzzle:

My sister-in-law who’s from here (the capital of Nowhere in northwestern Georgia) has a cat whose name she pronounces “Bayloo”, but when I call it that she has no idea what I’m talking about since the cat’s name is spelled “Betty Lou”. It’s really peculiar. My brother and I have to pronounce the name as three syllables with the TT (sounded like DD) in there or she and her sister and son have no idea that we mean the two-syllabled “Bayloo” they talk about.

Two issues here: where “Bayloo” (roughly [béylù], though the phonetic details will depend on fine details of Arne’s sister-in-law’s dialect) comes from; and why Arne’s sister-in-law doesn’t recognize his reproductions of her pronunciation (I’ll assume that his reproductions are close enough to accurate). The first question is easy; the second has a more complex, and much cooler, answer.

Whence Bayloo? In a response to Arne, Leith Chu got the core of the answer, though again there are some inscrutable details. Betty Lou, for most Americans, has a medial consonant that is not a [t], but an alveolar tap or flap, a voiced approximant that involves a brief ballistic contact with the alveolar ridge; the result is something similar to the pronunciation of the (dental) medial consonant in Spanish pero ‘but’ — that is, something r-like. For most Americans, medial /t/ and /d/ (after an accented syllable and before an unaccented one) are similarly realized in pronunciation: latter and ladder have the same medial consonant. One way to see this is that medial /t/ (in the relevant environment) is voiced (this is an accommodation to the voiced environment it is in), and then that this voiced segment (a [d]) is further simplified to a tap or flap.

That’s step 1. Step 2 is an even further simplification, the elimination of the medial consonant. That moves the [i] of Betty into the first syllable, as an offset; the result is equivalent to [béy]. That is, roughly, [béylù] Bayloo.

All of these simplifications are well attested in (many) American dialects.

The perceptual puzzle. So why would people who say [béylù] not accept this pronunciation as a reproduction of what they say? The answer, somewhat remarkably, is that what they are aiming at, in their mental processing, is something much closer to Arne’s pronunciation of Betty Lou (with a medial [t], or maybe [d], but anyway a tap/flap). In a very real sense, what they *think* they’re saying is much like what Arne says, but it gets automatically converted in production to Bayloo. So Arne’s pronunciation must  be of something different.

Let’s take a trip to language acquisition. Phonological acquisition is very complex, but one piece of it is pretty well known. People mostly think that kids “can’t produce” particular sounds. That’s not entirely false, but there’s a way in which it’s just wrong. Much of the course of phonological acquisition has to do with the way in which phonological competence is folded into the whole complex of highly coordinated language production (and perception). Along the way, kids have to deal with the production of particular sounds in specific contexts. And how to use all of this to convey meaning.

So, look at cases where it’s said that kids “can’t produce” specific sounds. Articulation problem? Sometimes. But then there are all those cases in which the kid is said to be unable to say X (in some context). Then it turns out that the kid

(a) can produce X as a noise or sound (a kid who “can’t” say [r], but produces [w] instead) does it just fine as an imitation of a dog’s growl);

(b) can produce X in favor of something more complex/demanding (a kid who replaces final obstruents with their devoiced counterparts can nevertheless replace final nasal stops by their oral counterparts; a kid who has initial [f] replaced by [d] — attested! — can nevertheless produce [f] when he’s aiming for [θ]);

(c) will reject imitations of their own productions (“NOT [fwi]: [fwi]!” — aiming for three).

Kids have a system of simplifications (starting from something like adult productions) that allows them to coordinate the complexities of speech production, while still doing their best to make meaning distinctions. Eventually, they overcome many of these simplifications and come to a system that lets them communicate (though not necessarily in exactly the same way as others). But of course many of these simplifications will remain, as part of the way people speak.

So adults will maintain their simplifications, while also maintaining their mental targets for particular words. The target is Betty Lou (or Beddy Lou), but Bayloo is just how you say it.

This entry was posted on March 1, 2013 at 10:47 am and is filed under Language acquisition, Phonology. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.