In a post on Language Log yesterday, Mark Liberman linked to this audio clip of John Kerry from the second presidential debate. Mark writes:

There’s […] an extra schwa between [p] and [l] in paraplegic and quadraplegic, similar to the extra schwa in Bush’s much-discussed “nucular” pronunciation of nuclear.

When it comes to matters phonological, we here at phonoloblog take such claims like “phono-fact x is similar to phono-fact y” seriously. Perhaps too seriously. In any event, what you are about to read (should you choose to do so) is not at all serious in the all-important sense of “well researched”, but it is serious in the lesser sense of “I’m actually interested in this, but I hope someone who knows more about this than I do will pick it up and run with it.” We’ll see how things turn out.

Geoff Nunberg has written the most accessible (and I think most probably correct) assessment of the whole “nucular”-pronunciation-of-nuclear, from his appropriately-titled collection of essays Going Nucular. Geoff distinguishes two types of individuals who (may sometimes) say “nucular” for nuclear.

First, there are those (let’s call them “members of the first group”) whose pronunciation is due to unfamiliarity and lexical analogy. Geoff writes:

Phonetically, in fact, nuclear is pretty much the same as likelier, and nobody ever gets that one wrong. (“The first outcome was likular than the second”? ) That “nucular” pronunciation is really what linguists call a folk etymology, where the unfamiliar word nuclear is treated as if it had the same suffix as words like molecular and particular.

Second, there are those (let’s call them “members of the second group”) whose pronunciation is more likely affected:

The interesting thing is that these people are perfectly capable of saying “nuclear families” or “nuclear medicine.” I once asked a weapons specialist at a federal agency about this, and he told me, “Oh, I only say ‘nucular’ when I’m talking about nukes.”

In the mouths of those people, “nucular” is a choice, not an inadvertent mistake — a thinko, not a typo.

In a Language Log post earlier this year Arnold Zwicky argues that in the case of the second group, this “choice” is probably not conscious. Arnold writes:

I don’t doubt that some people sometimes consciously re-shape their behavior in certain respects. But I think that most accommodations to social varieties and most constructions of personas via behavior (linguistic and otherwise) happen below the level of consciousness, usually with very little awareness of what features are being chosen or why. (In a sense, this *has* to be true. There are just too many bits of behavior for choices among them to be under conscious control. This is especially true for bits of linguistic behavior, which have to be produced in tiny amounts of time, many at the same time.)

Under this view, if I understand it correctly, it all comes down to whatever your lexical representations are: do you have nuclear stored as “nuclear” or as “nucular”? Note that if you are one of those folks who say “nucular war” but “nuclear family”, you have two different phonological representations, one for each meaning; people who say only “nuclear” or only “nucular” in both cases have one phonological representation (or two identical ones, depending on your theory).

Neither Geoff nor Arnold concludes whether Bush distinguishes two (or more) senses of nuclear — Geoff ends his piece with the promise, “I’ll be keeping my ears peeled” — but both seem pretty confident that Bush is a member of the second group. Geoff bases his argument on Bush’s background:

He must have heard it said correctly thousands of times when he was growing up — not just at Andover, Yale, and Harvard, but from his own father, who never seems to have had any trouble with the word.

Arnold is more sweeping about it:

[V]irtually everybody who says “nucular” is in the second group; though the support of other -ular words helps to make “nucular” sound right, these people are saying it because other people say it.

This brings us back to Mark’s claim:

There’s […] an extra schwa between [p] and [l] in paraplegic and quadraplegic, similar to the extra schwa in Bush’s much-discussed “nucular” pronunciation of nuclear.

For the sake of argument, let’s assume that the jury is still out on whether Bush is a member of the first group or of the second group – or, at least, that lexical analogy reinforces lexical representations and thus that both probably play a role in Bush’s “nucular” pronunciation of nuclear. What Mark is claiming, then, is that Kerry’s extra schwa in the (bound?*) root -plegic is (a) due to a difference in Kerry’s phonological representation of this root, and (b) probably reinforced by lexical analogy.

The trisyllabic “palegic” pronunciation, which is an amphibrach (“pălégĭc”), could be due to analogy with a number of words with largely similar forms (strătégĭc, ăllérgĭc, etc.) – the more flexible you’re willing to be on what counts as similar, of course, the more examples you will probably find. Does the disyllabic “plegic” pronunciation, which is a trochee (“plégĭc”) and which we are considering to be “correct”, have fewer close (and/or more distant) lexical neighbors? True, I can’t think (off the top of my head) of anything as good as strătégĭc, but there doesn’t seem to be far to go to get to things like mágĭc and Régĭs. (But, as I sort of warned you at the outset of this post, I don’t really know what I’m talking about here.)

The overall forms of the full (prefixed) words paraplegic and quadraplegic may be relevant. Kerry’s forms are pentasyllabic and consist of an initial dactyl (“pàrăpă”) and a final trochee (“légĭc”), which I will henceforth refer to as “DT form”. (I could of course say that it was an iamb followed by an amphibrach; honestly, I’m not assuming very much at all about how these metrical forms are represented.) The “correct” forms are quadrisyllabic and consist of two trochees (“pàră”/”quàdră” and “plégĭc”), which I will henceforth refer to as “TT form”.

It’s hard to find any close phonemic matches at this level of analysis, but metrical matches can be found in almost any paper on English stress. In the opening pages of Pater 1995 (published in Phonology 17.2, pp. 237-274) we find the following typical examples of (1) DT form and (2) TT form (see also Schane 2003):

(1)

DT stress Kìlimanjáro Nèbuchadnézzar Tàtamagóuchi àbracadábra Wìnnepesáukee Wàpakonéta dèlicatéssen

(2)

TT stress Pènnsylvánia rèpercússion ìnformátion làmentátion cònservátion fràgmentátion trànsportátion

My (admittedly naive) intuition tells me that the words in (2) are far and away more frequent words for most if not all speakers of English – I don’t care how many times you’ve been to Lòllăpălóoza. This seems like plenty of analogical pressure for the “correct” forms.

Sorry that I can’t follow through here, but that’s as far as I can go without devoting more serious time.

---

Note

* Note: There are no other salient/relevant senses of paraplegic or quadraplegic (at least, not that I am aware of), so we don’t have to worry whether or not Kerry has more than one phonological representation for these words. Although several other words with the root plegic can be found on just the first page of 2,210 Google hits, I’m fairly confident that these two are the only ones really common enough to be in anyone’s lexicon – including Kerry’s, a fact which I think is supported by the delay in lexical access that Mark also commented on. (back)