Semantic Compositions

More on the sound/spelling interface

This past evening, your host received a call from Oh Brother SC, requesting that he settle a dispute between OBSC and a friend. The question was whether "diminished" contained three syllables, or four.

Now, the cheap and easy answer would be "go check a dictionary" (you can do so here). But it wasn't clear that a dictionary was close at hand, and so SC was forced to try to explain how we count syllables to two people who both have bachelor's degrees from upper-tier American universities. Shades of Mark Liberman's concern that linguistic education for nonlinguists is woeful.

It was probably, on reflection, a mistake to try to explain syllable weight, or the notion of moras. Ditto for onsets, nuclei and codas (if you want definitions of these, or any other linguistic terminology, this is an excellent, albeit slightly dated, resource). Fortunately, your host was able to communicate the essential idea that there were three vowels about which the rest of the phonological material syllabified, and that there weren't any syllabic consonants to mess things up.

Or so he thought. Immediately after going through all that, OBSC's friend piped up, "But there are four vowels in the word!". It turns out that it's rather difficult to convince people -- at least on short notice -- that there isn't necessarily a 1:1 correspondence between sound and spelling, even though they may recognize in the abstract that English spelling isn't phonetic. Although the discussion ended with a consensus that there were in fact only three syllables in "diminished", the reasoning was more of a deferral to authority -- "you've got the master's" (or, from another perspective, "I have the power!") -- than an intellectual breakthrough.

Squawk Box, indeed

Readers who watch CNBC early in the morning -- around 7:00 a.m. Eastern -- are no doubt familiar with the show Squawk Box, a review of stories expected to dominate during the financial trading day. Readers with normal sleeping schedules, real lives, or West Coast addresses may not have heard the phrase before; a squawk box is also the slang term used by stock brokers to refer to the loudspeaker part of an intercom system, generally used to announce new recommendations to the brokers in an office at the beginning of a day.

Yesterday, though, it took on a disturbing new meaning as TheStreet.com's Cody Willard attempt to coin a new word to describe the economics of chicken pricing (sorry, no link available). Mr. Willard normally writes a column on his specialty, telecommunications stocks, under the byline "The Teleconomist". So he couldn't resist making a similar pun in discussing chicken:

Pilgrim Pride citing lower feed costs as it guides earnings higher today. As Scott Rothbort points out on SI, gotta figure those lower costs will likely trickle down the chickenomy (I'm coining the term as the chicken world's equivalent of the teleconomy. Chickonomy? Chickeconomy?)

For his part, Mr. Rothbort contributed an additional pun combining chicken and econometrics: chickenometrics.

There aren't any hard and fast rules about coining blends of this sort, only intuitions (and probably underlying rankings) about what most naturally fits English-specific preferences for syllable weights, stress patterns, and other phonological constraints. Given that none of Mr. Willard's choices stand out as obviously the best to him (or to SC), there's probably also some competing consideration about which blend most naturally presents the combined semantics of "chicken" and "economy".

Of course, if one wants to try to predict what the "right" blend would be for most people, it's not that hard to take a theoretical stab at it; English phonology has been studied well enough that one could simply pick up a good reference like Michael Hammond's Phonology of English, or perhaps a few recent papers on stress in English, adopt a constraint ranking from the sources, and compute the tableau necessary to decide what the optimal form is. Mr. Willard clearly missed out on a career as a phonologist, given that he already thought to come up with most of the likely candidates for this process. Nevertheless, this won't quite get you all the way, since it's not clear that this is a purely phonological process. To SC's ear, "chickonomy" and "chickenonomy" (not actually one of the presented choices) both sound pretty reasonable. But then he considers how well received any of them would be in a normal conversation, and suddenly "the price of chicken" sounds a lot better.

Pharmaceutical phonesthemes

Some time ago, SC wrote about phonesthemes, a theoretical attempt to suggest that the pairing of sound and meaning is not arbitrary. Yesterday, this notion crossed paths with another SC favorite, Pfizer (no, I'm not obsessed, not me), in the form of an investment newsletter piece (no link available) on two upcoming drugs to be released by the company in the next 24 months. One, Exubera, is an inhaled form of insulin, and the other, Lyrica, is a pain reliever and anti-epilepsy drug much like the just-off-patent Neurontin.

Assuming the pronunciations follow what I expect are most English-speakers' intuitions, both drugs end in the same vowel. So does Viagra. A quick look at the Pfizer home page only finds one other drug in this paradigm, though -- Bextra -- which is another pain reliever. Bristol-Myers' product page only adds Sustiva, an anti-HIV drug. Aventis throws in a few more, including Allegra, Apidra, Arava, and Diabeta. Novartis adds Zometa; a little more exhaustive searching of their site yields: Aredia, Femara, and Rescula. There's definitely a market for drugs ending in neutral vowels, but apparently, schwas don't quite scream "drug!".

A rather different sound or sequence might, though. Have a look at this list: Keltek, Lasix, Lovenox, Rilutek, Viskaldix, Starlix, Leponex, Gleevec/Glivec (two variants of the same drug), Esidrex, Dovonex, Erbitux, Ifex, Mesnex, Plavix, Rubex, Videx, Celebrex, Relpax, and Zyrtec (also taken from the drug company links provided above). What is it about words ending in /k/ or /ks/ that says "I'll make you better"?

A solid nuculus

Listening to the unrestrained gloating of Chargers fans on the radio after that truly dominating performance yesterday, SC was amused to discover that for some people, nucular isn't merely a one-off pronunciation error, but part of a larger paradigm. In context:

Caller: It's good to see the team has finally got a solid nuculus of core talent.

Aside from the awful redundancy of that sentence, this was of interest as the first time your host has ever heard the "nucular" pronunciation generalized to other words from the same root. It tends to suggest that the source of the error, at least for this person, isn't merely a misfiring in the phonology/phonetics interface, but a real lexical representation. Of course, it could be another random error, but as the classic phonology joke goes, "once is an accident, twice is a rule". So be on the lookout for a second example!

As for the game itself, it has been recently suggested that SC's analysis of the Chargers' draft pick from earlier this year was off-target. Mr. Roethlisberger has indeed played above expectations. But: 1) nowhere did your host actually advocate for taking a quarterback, and 2) Drew Brees is outplaying every starting quarterback in the state of Pennsylvania. Oddly, so is Brian Griese.

Of course, if Prof. Liberman really wants to put something on a Post-It note for the "SC, stick to your day job" files, compare this late-August post with the present-day price of Google stock. At least I didn't tell anyone to short it.

I say chipotle, you say chipotl

Neal Whitman has a very enjoyable post up about work that Elizabeth Hume is doing on metathesis. SC appreciates the reminder that this is the word he was reaching for in the recent discussion of "Calvary" and "calvary", as the type of phonological error that it could be.

At the end of the post, Dr. Whitman comments:

Aside from the above highlights, though, it seems that Semantic Compositions and I are in good company in enjoying fast food linguistic analysis: On p. 223, Hume has a fun discussion of the alteration of chipotle to chipolte.

At first, your host was moved to comment on his observation that he hasn't heard (perhaps more accurately, hasn't noticed) this alteration. But it certainly is within SC's experience that "chipotle" has different pronunciations among the crowd that likes to flash their knowledge of Spanish phonology versus those that tend to ruthlessly assimilate their pronunciations into English.

Essentially, the difference is a matter of vowel quality. A Spanish speaker, or KABC's traffic reporter (a man whose speech shows no trace whatsoever of a Spanish accent until he signs off and rolls the r's in his last name), pronounces "chipotle" as three distinct syllables. To use a grossly inappropriate non-IPA transcription system, it comes off as "chee-POTE-lay", with three distinct vowels serving as the nuclei of each syllable. On the other hand, people prone to butchering foreign borrowings pronounce it as "chip-POTE-l", with a syllabic "l" at the end just like "poodle" (SC will attest to versions with and without a geminate "p"; some people say "chi-POTE-l").

Webster's indicates four pronunciations as canonical, including a version which renders the last vowel as though it rhymed with the name "Lee". Your host can't say he's ever heard that one before. But there's a larger question that SC has no good answer to: are there single preferred pronunciations for words like this from the standpoint of a monolingual English speaker? Recognizing that there are a wide variety of regional accents, and just plain idiosyncratic features of individual speech, and that these will affect any one speaker's judgment of the "right" prounciation, how would one judge whether a pronunciation is better or worse in accordance with one's understanding of both the foreign word and English phonology?

This is not unlike the gradation in judgment we often make between "nuclear" and "nucular', where one prounciation carries the stigma of sounding uneducated. Your host imagines that, at least in some circles, pronouncing foreign words in an assimilated fashion sounds less educated, but it's also easy to imagine other social contexts where making a show of code-switching is considered merely pretentious. SC finds that he has no intuitions (or at least very weak ones) on the relative prestige of pronunciations of "chipotle".

Hamming it up

Right before taking off for a week, your host went on a bit of a sports-related jag. Since Mark Liberman was kind enough to weigh in with the suggestion that lingua-sports blogging is an area that could be explored further (as he's been up to himself recently), there's something else that caught SC's eye during the Olympics which he would have posted about had circumstances not intervened.

It's impossible to have followed soccer in America over the last decade without being aware of Mrs. Nomar Garciaparra, aka Mia Hamm. And unless one avoids the sports pages altogether, it's also impossible to not have read about the remarkable -- albeit controversial -- gymnastics gold medal won by American competitor Paul Hamm.

Both Ms. Hamm and Hamm's beer have conditioned most American English speakers to pronounce the name "Hamm" identically to the cut of pork, which is to say "ham". The gymnast, however, pronounces it to rhyme with read-only memory, as in "ROM" (or at least the sportscasters discussing him do; SC hasn't seen Mr. Hamm interviewed live). As the linked profiles from ESPN indicate, both Hamms were born in the U.S. Presumably, both are also descended from German immigrants; it would be interesting to know the difference in the number of generations that their families have been in this country, which might be predictive of a shift between using the German and English vowel inventories. There are, of course, plenty of alternate explanations: a desire to assimilate, a desire to reassert the original ethnic pronunciation, etc. Nevertheless, one might expect that given the predominance of the "ham" pronunciation, at least some of the newscasters discussing the story would have conflated Mr. Hamm's name with Ms. Hamm's, and so far as your host could tell, that didn't happen.

And while I'm on a sports kick, this is undoubtedly my favorite baseball story of the year. At least until the Padres win the World Series.

Hark or harp?

Mark Liberman has an interesting post up tonight about people who can't keep their complaints about the supposed decline of English straight; is it harking or harping back to the mythical golden age?

Prof. Liberman's citation of Google and the OED is quite thorough, and there's no doubt that the historically correct form is "harking back". The only reason your host feels obliged to comment is that Prof. Liberman analyzes the error as in part a semantic issue; some of the examples of "harping back" seem to refer explicitly to past complaints.

SC was surprised to see this analysis, if only because he would expect a phonetician like Prof. Liberman to seek a phonetic or phonological explanation for the error. The /p/ at the end of "harp" shares the same place of articulation as the /b/ at the beginning of "back", and this sort of anticipatory assimilation is pretty well attested in other cases, without having to appeal to a semantic explanation. One example which makes it impossible for SC to listen to people without giggling, at least since it was first brought to his attention in a psycholinguistics class, is a frequent error when people say "phone book". Listen to someone speaking quickly, and it will come out as "foam book".

How do you say "Giles"?

It's baseball season, which means that rotisserie league geeks -- and wannabes -- are back to the world's most productive use of time: poring over daily statistics tables.

While working on setting up his fantasy team, SC noticed that ESPN has added something new to their player information pages: pronunciation guides for the names of players with uncommon names. So, for example, those wondering about how to pronounce "Erubiel Durazo" will be happy to find the transliteration "eh-ROO-be-el du-RAH-zo".

Unfortunately, their selection criteria are rather bizarre, and their transcription system leaves something to be desired. For example, while they note how to pronounce pitcher Livan Hernandez' middle name (it's "lee-VAHN"), they don't tell you how to pronounce his first name (that's "Eisler"). Catcher Ramon Hernandez gets the transliteration "ruh-MOWN", but it's not clear how they mean for that to be contrasted with the last name of Expos player Scott Downs. As things stand, they can only distinguish between stressed and unstressed syllables, although many players' names have more than two levels of stress (the aforementioned Mr. Durazo's first name might better be rendered as eh-ROO-be-El, depending on how faithful they're trying to be to the original Spanish).

It's beyond SC's powers of comprehension, though, to deduce how the brothers Giles, Brian and Marcus end up with the transcription "JYLES" for their last name. Is it really two syllables of primary stress? I don't think so -- and I'm not sure that "JY" is a more obvious clue to the proper pronunciation than the original English orthography.

This is, of course, an excellent argument for why IPA transcription needs to be a part of everyone's basic high school education.

Pardon my French

The notion of "phonesthemes" has long been one of the more dubious ideas in the study of language, at least among linguists. Although most people would claim to recognize a certain amount of sound symbolism in onomatopeic words, the idea that sound-meaning pairs are not arbitrary has generally been dismissed.

In spite of the meme's rather dismal reputation, some people have gone ahead and tried to catalogue evidence for some of these alleged correspondences anyways. The late Dwight Bolinger, a linguist of impeccable reputation, wrote a paper speculating about evidence for phonesthemes from an American Indian language (the reference can be found here; see here for a list he was partially responsible for compiling). And it's hard not to at least find the idea mildly seductive: with examples like glimmer, glitter, gleam, glare, and glow, it's hard not to think that there's something shiny about gl-.

This all came to mind for SC when trying to track down yet another piece of music heard in the gym while working out ([work harder -- ed.]). A not entirely obnoxious song involving an accordion and the phrase "you promised me" had been bothering your host. Searching the web for it turned up the fact that it was the English version of a currently popular recording of a French song called "Tu Es Foutu". As written in English, the lyrics describe a woman who is rather irritated at her boyfriend. In spite of not being a French speaker, though, SC was reasonably sure that "Tu Es Foutu" does not translate to "You Promised Me"; he was willing to guess that "Tu Es" translates as "You Are", but that was about it.

So what's "foutu"? Altavista's translation was no help, returning the English word as "foutu". One dictionary found through Yahoo suggested it meant "ruined". But SC couldn't help suspecting that one f-bomb sounds much like another, and it didn't take long to find another, much more frank dictionary which suggests, um, "screwed".

Maybe there's something to this phonestheme idea after all.

I wanna set you up

Claire Bowern follows up on a recent Language Log posting regarding "mondegreens". SC was only introduced to the term late last year, but it pretty well sums up why he completely mistrusts himself in his ability to correctly interpret song lyrics.

The term, as this article mentions, was coined by a writer for The Atlantic due to a song she had misheard herself. Specifically:

The term "mondegreen" was coined by Sylvia Wright in a 1954 Atlantic article. As a child, young Sylvia had listened to a folk song that included the lines "They had slain the Earl of Moray/And Lady Mondegreen." As is customary with misheard lyrics, she didn't realize her mistake for years. The song was not about the tragic fate of Lady Mondegreen, but rather, the continuing plight of the good earl: "They had slain the Earl of Moray/And laid him on the green."

A couple of good sources for these things are the "misheard lyrics" pages maintained by various people. SC likes this one, and also this site, which contains substantial additional parody material.

As for the title of this post, SC was just barely into senior high school when he discovered that one of the mothers who drove in his carpool had misunderstood the lyrics to a then-popular song by a group called Color Me Badd. No, SC did not quote the song. He was just surprised to hear her remark that she liked it when it played on the radio.