Posted by Neal on April 8, 2008
I’ve been reading Steven Pinker’s The Language Instinct, and in the chapter on phonetics he reveals an interesting pattern. Here’s what he says:
Why do we say razzle-dazzle instead of dazzle-razzle? Why super-duper, helter-skelter, harum-scarum, hocus-pocus, willy-nilly, hully-gully, roly-poly, holy moly, herky-jerky, walkie-talkie, namby-pamby, mumbo-jumbo, loosey-goosey, wing-ding, wham-bam, hobnob, razza-matazz, and rub-a-dub-dub? I thought you’d never ask. Consonants differ in “obstruency” — the degree to which they impede the flow of air, ranging from merely making it resonate, to forcing it noisily past an obstruction, to stopping it up altogether. The word beginning with the less obstruent consonant always comes before the word beginning with the more obstruent consonant. (p. 166)
All his examples, naturally, fit this pattern perfectly. Taking the least obstruent consonants first, we have the glides: [y] and [w]. Pinker doesn’t have any items starting with [y], but he does have a few starting with [w]: willy-nilly, walkie-talkie, wing-ding, wham-bam (assuming that for my readers, wh is pronounced as [w] and not [hw]). These all fit his pattern with no problem, since no consonants are less obstruent than a glide.
A bit more on the obstruent side are the liquids ([l, r]), and nasals ([m, n], and for completeness, [ŋ], even though no words in English start with “ng”). I don’t know which of these two classes is the more obstruent, but if Pinker’s claim is true, it would be the nasals, since the element beginning with the liquid comes before the one beginning with a nasal in his item razza-matazz. I’ll give the benefit of the doubt here. The other examples starting with liquids fall into line, since their second elements start with consonants other than glides or nasals: razzle-dazzle, rub-a-dub-dub, roly-poly, loosey-goosey.
Likewise, the second elements for the items starting with nasals all begin with consonants other than glides or liquids: mumbo-jumbo, namby-pamby.
More obstruent still are the fricatives, the consonants that force the air “noisily past an obstruction”: [f, v, ð, θ, s, z, ʃ, ʒ, h]. All but two of Pinker’s items have second elements that begin with a consonant that is not a glide, liquid, or nasal: super-duper, herky-jerky, hocus-pocus, hully-gully, helter skelter, harum scarum. (And my own additions: hanky panky, hokey pokey, higgledy piggledy, Humpty Dumpty, hickory dickory dock, hunky dory.) Now helter skelter and harum scarum have both elements starting with another fricative, [s], but this [s] is part of an [sk] cluster, and [k] is not a glide, liquid, nasal, or fricative, so we’re still good. But I was confused by Pinker’s citing of holy moly and hobnob, since I thought nasals were less obstruent than a fricative. OTOH, [h] is different from other fricatives. It doesn’t involve the tongue obstructing the airflow at all — just the air flowing through an open pair of vocal folds. So maybe this fricative isn’t so obstruent after all. I’ll give him these items as well, and add my own hugger mugger.
Next most obstruent are the affricates, [ʧ, ʤ]. Each consists of a stop and a fricative smushed together into one consonant. Pinker didn’t have any items starting with affricates, but I do: jelly belly and jiggery pokery, whose second elements do not start with glides, liquids, nasals, or fricatives.
I was so fascinated with Pinker’s observation that for the next few days I tried to think of other quasi-reduplicative word pairs and see if they fit the pattern, too. Some did, and I listed them above. But some didn’t. I’ll start with some items whose first element begins with a stop: [p, t, k, b, d, g]. If the second element begins with anything other than another stop, we have a counterexample. With that in mind, I call your attention to: boogie woogie, tutti frutti, ding-a-ling, Ducky Lucky, and Turkey Lurkey. And, of course, bowly holy. That’s a family term for a biscuit into which you dig a hole that you then fill with honey. I’m not making this up! It looks like there’s also at least one person out there who uses this term to refer to navels.
Next, a few that start with affricates: Chicken Licken, chug-a-lug, Chumbawamba. Each second element begins with a less obstruent liquid or glide. Now for a pair beginning with a fricative: fuzzy wuzzy. (Pinker even mentions this one elsewhere in the book to make a different point, but doesn’t mention it in this part.) OK, what about one starting with a nasal? Mellow yellow, with the second element starting with a glide.
What do you think? Are my counterexamples enough to trash Pinker’s rule, or do they have some commonality that we can use to formulate a class of exceptions? What examples can you think of?
UPDATE: I wrote to Steven Pinker and asked about these counterexamples. He was kind enough to provide some further information and references:
The discussion of order in conjunctions is based on an old paper written with Prof. David Birdsong, now at the University of Texas at Austin:
Pinker, S., & Birdsong, D. (1979). Speakers’ sensitivity to rules of frozen word order. Journal of Verbal Learning and Verbal Behavior, 18, 497-508.
It was based on a theory proposed by two of our teachers at the time:
Cooper, W. E., & Ross, J. R. (1975). World order. Paper presented at the Papers from the parasession on functionalism, Chicago Linguistics Society.
In our paper we listed a number of principles, and tested their strengths by eliciting ratings. Number of initial consonants (more in the second conjunct) was pretty powerful (as in helter-skelter). We didn’t test compounds of the piggly-wiggly type, but they would fall under that generalization if, as many phonologists have proposed, “w” is mentally represented as a pair of phonemes.
The conjunctions whose second member begins with “l” does indeed violate the principle that the more obstruent consonant comes second. We are not sure why these examples systematically flout it. It is interesting that the consonant-obstruency principle was one of the weakest in our data, that there is little evidence that it applies across languages, and that speakers don’t apply it to new items in their second language. So it seems to be a marginal principle at best.