[Guest post by Noam Scheiber:]

The linguistics nerds over at The Economist have launched an entire blog devoted to the subject--called "Johnson," of course--netting us this fascinating insight:

THE idea of "phonetic symbolism"—specifically a link between sounds and perception of size—appeared to me for the first time in Steven Pinker's Language Instinct.  Some vowels are "high" in phonetic terminology, like [i] (as in "sweet"). Others are "low" (like [o]). High vowels (with the tongue raised in the mouth) are often associated with small or delicate things, low vowels (where the oral cavity is made larger) with big things. In addition, "front" vowels, like [i] again, are seen as smaller than "back" vowels, like [u], which has the same height. As Mr Pinker put it, mice are teeny and squeak, elephants are humongous and roar. English speakers are likely to guess, correctly, that of qing and zhong in Mandarin, qing is the one that means "light", while zhong means "heavy".  When I first read this, I filed it mentally under "fascinating" and didn't do much with it.
But Keith Coulter and Robin Coulter, two professors of marketing, have shown that phonetic symbolism can have real-world consequences. Sibilant consonants (where there is a constant flow of air, like [s]) are also perceived as "small", while stops (where the air is blocked, like [t]) are perceived as large. So a word like "six" is perceived as "small" (sibilant s plus high, front vowel), while "two" is "large" (stop t, back vowel).  Subjects were offered a product (an ice-cream scoop) discounted from $10 to either $7.66 or $7.22.  Those who were told to mentally repeat the price to themselves because they'd need to remember it later were morely likely to rate the scoop a good deal at $7.66 than at $7.22.  (The effect disappeared for those not told to rehearse the sale price.)

Maybe something the Obama OMB should look into when selling its next budget...

P.S. And, yes, I realize we're talking phonetics here, not semiotics. Semiotics just sounded better. Maybe there's a branch of phonetics that could tell me why...