Semantic Compositions

Compounds with interest

Courtesy of Mrs. SC, an article on pollution from Mt. St. Helens in Washington. The linguistically interesting bit:

Italy's Mount Etna can produce 100 times more sulfur dioxide than Mount St. Helens — and sits in the middle of a heavily populated area. The volcano spawns acid rain and a type of bluish smog that volcanologists call vog, which can affect large swaths of Europe, said Terry Gerlach, a U.S. Geological Survey scientist who studies volcanic gases.

Since Mrs. SC already took the trouble to unpack the derivation, I'll quote her:

Apparently, vog = volcano + smog. As you and I both know, smog = smoke + fog. => vog = volcano + (smoke + fog)

That's about right, and raises the question of how many times you can iterate this sort of process before the derivation becomes hopelessly opaque. "vog" could just as easily be derived from "volcano + fog", but fog isn't really scary enough to get a new word on its own, regardless of the source (not that fog could come from a volcano anyway). It has to be poisonous!

So what else might we compound into an existing compound? Could fast-food signmakers save a few letters by combining "bacon cheeseburger" into "beeseburger"? Perhaps editorial writers could stitch together the derisive uses of "pollyanna" and "Mickey Mouse" into "pollymouse"? Maybe the good folks at Louisville Slugger will decide that instead of "baseball bats", they can now sell us "basebats". Yikes! SC is all for new words, but every now and then, maybe it's a good idea to step back and enjoy the ones we've got.

Right on! (Now where's my funding?)

Geoff Pullum comments on an article from the Chronicle of Higher Education (which SC considers a must-read journal even if you're not in Higher Education), which argues about conservative representation on campus. Your host has no argument with Prof. Pullum's general views of the article, but there's a bit of niggling about frequency statistics which can't be allowed to pass.

The trouble starts with an ill-advised sentence like so:

Hence, references to "right-wing think tanks" are always accompanied by the qualifier "well-funded."

Prof. Pullum, being an excellent grammarian but perhaps less sensitive to the relevant frequency effects than desirable, interprets this to mean that this complaint is always expressed with strings of the form: "well-funded right wing think tanks". He duly provides evidence that this is not so, and chides the author, Mark Bauerlein, for making an unnecessarily broad and erroneous statement. Here's his table:

"right wing think tanks"	14,600
"well funded right wing think tanks"	50

The problem for Prof. Pullum is that Prof. Bauerlein is closer to correct, not to say 100% so, than this experiment indicates. Prof. Bauerlein didn't say that the cooccurrence of the phrases in discourse is always a collocation, and his intended point is grounded in this alternate understanding.

Here is a rather different way of looking at the data (all searches done with Google):

+"right wing" +"think tanks"	86,000
+"left wing" +"think tanks"	21,300
+"right wing" +"left wing" +"think tanks"	12,000
+"right wing" -"left wing" +"think tanks"	70,500
-"right wing" +"left wing" +"think tanks"	9,080
+"right wing" +"think tanks" +"well-funded"	934
+"left wing" +"think tanks" +"well-funded"	403
+"right wing" -"left wing" +"think tanks" +"well-funded"	764
-"right wing" +"left wing" +"think tanks" +"well-funded"	81

What we see by looking at these searches is that: 1) there is considerably more discussion of "right wing" "think tanks" instead of "left wing", 2) that the articles discussing "left wing" think tanks are likely to mention "right wing" organizations about 50% of the time (suggesting that no more than half of the documents discussing "left wing" institutions are likely to be conservative polemics), and 3) that there is an indisputable effect that "right wing" occurs between 2x and 8x as often in documents calling something/someone "well-funded" as "left-wing" (the minus operators are used to isolate the documents that discuss one side alone from the ones that discuss both). I am rather skeptical that this occurs merely by chance, especially since the effect holds in both the unrestricted case and when one tries to suppress one wing or the other from the counts.

Here's one more interesting bit of data:

+conservative +liberal +"think tanks" +"well funded"	965
+conservative -liberal +"think tanks" +"well funded"	1,330
-conservative +liberal +"think tanks" +"well funded"	361

Funny thing about that isolation technique; every time you isolate talk about "conservatives" or "right wingers", there seems to be a large multiple of references to "well funded"ness over the counts found when the talk swings the other way.

Now I'm just going to be gratuitous:

"well-funded conservative think tanks"	118
"well-funded liberal think tanks"	0

A reasonable objection from Prof. Pullum might be that there simply is no way to tell algorithmically how many of these references specifically identify right-wing think tanks as well-funded, since we don't have the benefit of collocations to make that immediately indisputable.  Simply inspecting the data will have to suffice to illustrate the point.  So here are the contexts for "well-funded" from the first 10 (working) Google hits for the "plus left/minus right" well-funded search (with the caveat that this was true when I searched the first time; different pages come up now, but the point appears to still stand):

(link) The far right's effort to create an intellectual conservative movement is truly comprehensive, well-funded and well-organized. They are not only building think tanks at the national level, but the state level as well, since congressional conservatives are trying to devolve power to the states.

(link) In terms of political process, the existence of powerful and well-funded conservative "counter-institutions" raises the specter of what some have called "supply-side' politics.

(link) To be so branded opened a journalist to relentless attack by well-funded right-wing media 'watchdog' groups and other conservative operatives...The conservative funding also finances right-wing think tanks such as the American Enterprise Institute, the Heritage Foundation, and the Cato Institute, which provide easy jobs for conservatives who produce the sound bites and op-eds to fill up the mainstream news stories and editorial pages.

(link) Matter of fact, we need, like the right has, half a dozen or a dozen equally strong, well funded progressive organizations putting out policy promotion, fighting the political battles with equal "idea weaponry" to what the right has been using for years.

(link) As many as 50 of these well-funded private institutions, who coordinated as one big right-wing beehive, were effective at training and marketing operatives who would infiltrate major news outlets and take political office.

(link) They may have their little magazines and well-funded think tanks, but they are outnumbered by billions of people.

(link) Daniel Pipes has accused scholars like Foner and Gilmore of hating America...So what Foner and Gilmore encountered is a well-funded campaign to pursue an ideological agenda.

(link) The concept has at least started to permeate American right-wing think tanks (well-funded think tanks at that), so it's just matter of time before it starts to make it's way into the Republican party as a hole and into conservative foreign policy circles.

(link) What I like most about Canason's book is its basic thesis: there exists an enormous, right wing propaganda machine that includes everything from well funded, academic think tanks to party-line pundits, and this machine has produced a spate of "big lies" that have seeped into the public consciousness.

(link) Other than those convinced of a left-wing liberal conspiracy in the media, the book's breakdown of a movement rooted back in the 60's and advanced thru a calculated, highly organized and well funded effort thru various think tanks, deregulation and an increasingly strict media culture will give show a very real movement in convincing fashion leaving the reader with much to think about when he gets his news.

These are 9 of the 10 on the first page, and the first one from the second page.  I could go on pasting text and links but it's quite tedious, and at this point I feel comfortable inviting Prof. Pullum to conduct the same search and see for himself that the usage repeats over and over and over again.  The blunt truth is that, among a certain crowd, the dark insinuation of being "well-funded" is a popular stick to beat opponents with, and I think it is reasonable to intepret the above table as demonstrating that the "well-funded" shot, when it is taken, is taken between 60-90% of the time at conservatives. Admittedly, it's not so common as Prof. Bauerlein indicated; neither, however, is it so nonexistent as Prof. Pullum thinks. If Prof. Pullum wishes to continue defending the point that:

Claims as patently ridiculous as that liberal professors always qualify "right-wing think-tank" with the attributive modifier "well-funded" discredit the people who make them.

on the grounds that the truth conditions of universal quantification aren't met, he's welcome to do so, but the statistics are hardly as lop-sided as he thinks. In one case (albeit not after Google indexes this post), "always" even turns out to be true.

UPDATE: Prof. Pullum updated his post shortly after this went up based on comments from Maryellen MacDonald which addressed a related point. He came across the same figure about "well-funded conservative think tanks" and dismissed it, but did not contrast it with "liberal". Which, in fairness, he had not set out to do originally. But the point being made above was not that he was wrong about "well-funded" always being attached to "right-wing", but rather that as a rhetorical trope, it happens to be rather one-sided, and sufficiently frequent that the people it's directed at would pick up on it.

(Edited on 11/12/04 at 5:05 p.m.)

Coordination is a tricky thing

SC is not a professional musician, but he frequently makes use of pro music suppliers to buy test equipment and spare parts. As a result of this, he's on a couple of mailing lists, including a rather enjoyable one from a company called Musician's Friend. Every week, the company sends out an e-mail with a list of weekly specials, which always includes a link to a list of "on this date"-type trivia. This week's e-mail contained the teaser:

Join us as our Week in Review reveals what femme rocker's concert was the target of a chainsaw protest in the Northwest and who swore off rock 'n' roll after getting religion.

Your host interpreted this to mean that there existed a single person who had three properties: 1) was female, 2) gave a concert that was protested, and 3) swore off rock'n'roll. As it happens, that was the wrong parse of this sentence. The correct answers (available here) are:

1957, in Sidney, Australia, Little Richard announces his intention to give up rock and roll and "live for the Lord" ... he flies to Los Angeles the following day and is baptized a Seventh Day Adventist ... he will abide by his decision for five years before resuming his musical career...

and

1996, it is a musical chainsaw massacre at an environmental benefit concert in Jacksonville, Oregon, when Bonnie Raitt and band are drowned out by protesting loggers who rev chainsaws and light firecrackers to show their opposition to saving the redwoods...

Every now and then, your host idly wonders whether or not grammar would be meaningfully improved if people adopted the semanticist practice of dropping subscripts next to noun phrases to indicate whether or not different entities are being referred to.

I protest!

Geoff Nunberg comments on a headline that reads "Vietnam Veterans Protest Kerry", saying that "most people reserve the verb for objects that denote events or states of affairs". While SC loves the Rambo allusion that this turns out to be an opportunity to reprint (someday, your host will find a way to work the very best movie speech of all time into a post), he's not so sure about the empirical syntactic claim. In particular, it strikes your host as fairly common for "protest" to be followed by the name of a person. As usual, off to Google, 'cause this punk is feeling lucky:

"protest Ashcroft": 327 hits
"protest Bush": 54,700 hits
"protest Cheney": 655 hits
"protest Clinton": 906 hits
"protest Kerry": 1,350 hits

Now, a number of prominent political names failed to register more than a few dozen hits, which were largely syntactically irrelevant. Plugging Dole and Edwards into the phrase got virtually nothing, and trying Reno and Kennedy produced more hits, but almost all of which met the conditions laid out by Prof. Nunberg (i.e. "Reno's speech" or "Kennedy appearance").

And plenty of the hits in the searches linked above also meet Prof. Nunberg's criteria. Google will happily return any page where the string is found, even when you don't want anything to follow it in a sentence (which is why we can't get the Linguist's Search Engine to replace it quickly enough). But browsing the results turns up plenty of additional examples of a person as the object of "protest", at least in headline contexts:

(link) "Tens of Thousands Protest Bush in NYC"
(link) "200,000 in NY Protest Bush"
(link) "Campus Conservatives Protest Clinton"
(link) "Hundreds Protest Clinton at Florida Fundraiser"
(link) "Sierra Club to Protest Cheney"
(link) "Libertarians Protest Kerry"

Of course, in the absence of a more linguistically sophisticated search engine, there's no way to conveniently answer the question of whether "most people" (or perhaps better, "most instances") combine "protest" with event objects. But as the data above illustrate, it's not hard to find contemporary examples, so we can answer the (quite possibly rhetorical) question that ends his post: it's not likely that the headline writers are alluding specifically to Stallone's speech in First Blood, since the examples come from a wide spectrum of publications and political views. Then again, they're also all headlines, so perhaps these uses are elliptical references to the events that would normally accompany "protest", and Prof. Nunberg's analysis of the verb stands.

Warning: Meltup Imminent

That title is, remarkably, not a typo. Following up with recent talk about stocks and language, and also the asymmetry of certain affixes, here's a word that seems to work in financial contexts, but not otherwise.

Doug Kass, a frequent favorite of SC's to write about, paraphrased a fictional anecdote written by a market analyst about 10 days ago by writing: "This time -- despite the dour sentiment and poor action - Benny the Trader expects a meltup in the market." Anyone following the Dow Jones average this past week might conclude this was an accurate expectation. This isn't the only time SC has seen that word in a service of TheStreet.com; although also unlinkable, in the same publication, a search of the last 4 years' worth of articles reveals "An unexpectedly positive resolution could easily produce a meltup scenario", and "A couple of hours into the day, chip stocks went positive, and from there, it was in essence a mini-meltup". That search also reveals this:

We often hear the term "meltdown" when a market seems to fall apart completely. We have a "meltup" today, where the market moves steadily higher without any clear reason or catalyst.

A cursory inspection of Google hits suggests that other instances of the word also are generally financial. With about 1,000 Ghits as a single word, and twice as many as a two-word phrase, most of the uses are like this:

(link) Melt-up? It can't be a well thought-out market if it is represented by so unlikely an image.
(link) Lately, though, I have come to believe that the single most unlikely scenario, the one nobody’s predicting, might be about to occur: the massive melt-up, the upside blow-off that lifts the Dow a thousand points between here and year-end.
(link) This resistance has now been breached and the remaining shorts, faced with the imminent prospect of runaway losses, are likely to generate a self-feeding Meltup.
(link) If the market is going to get up there and keep on going, this would have to be a meltup type of rally.

A number of the hits for "melt up" as a bigram have nothing to do with finance, but the syntax isn't such that it should be interpreted as a phrasal unit anyways. For example:

(link) That allows the snow to melt up on the roof over the attic, and then refreeze when the water reaches the eaves.
(link) In the first, called passive flow, convection of the solid mantle creates a broad region of melt, up to 100 km across.

It makes sense that such an odd construction, which doesn't have any intuitive physical meaning, should be limited to a domain where it doesn't really have to correspond to anything literal. It's just a case of back-formation, where it's quite logical that something fairly abstract (a change in the value of a statistic) should have a pair of antonyms to describe changes in opposite directions. The fact that this particular statistic, the price of an equity, is prone to being discussed in rather colorful language only means that it will produce more fanciful metaphors than, say, changes in the flow rate of a sprinkler system.

As for the recent financial punditry appearing on these pages, readers expecting SC to eat crow on account of Google's failure to crash shouldn't look for him to be breaking out the fork and knife anytime soon. While the stock has gone up since it opened, Google ended up resorting to what is known to economists as a "Vickery auction", where the highest bidder ends up paying the second-highest bid price. Effectively, they made a move to give back some of what they could have gotten for the stock to give the buyers an incentive to actually buy it, but only after limiting the last round of bidding to people who had been willing to pay the higher prices (and had registered at that time). Even at $108, it's still trading at only about 80% of what the company originally hoped to sell it at, and the amount they actually received was a good 25% lower than where it is now. It was a reasonable move to make to keep the stock from acquiring a bad reputation, but it made the IPO notably less profitable for the company than it could have been.

At least three ways to say two

This past Sunday, your host and Mrs. SC went to see the current remake of The Manchurian Candidate. Since neither of us has seen the original, I have no grounds for offering an opinion about the relative merits of the remake, but suffice it to say that it was enjoyable on its own terms (provided that several major plot holes were ignored). That's not what this is about, though.

While at the theater, your host saw a poster advertising an upcoming movie entitled Anacondas, a sequel to a film that SC never would have guessed merited one, Anaconda. Now, the title of the new film is not meant to be taken as a sign that the movie features two snakes, whereas the first one featured only one; presumably, there will be a multitude of snakes in the movie. The point of the plural morpheme is only to indicate that the film is a sequel, an interpretation which requires the reader to be aware that there was an Anaconda before there was (were?) Anacondas.

A similar trick was pulled with Alien, which begat Aliens. Had English retained the historical three-way contrast between single, dual, and three-or-more plural, Aliens could have been the title of the third movie, and the lack of fresh titles might have prevented the abomination known as "Alien: Resurrection" from ever making it onto a screen.

Use of the plural morpheme isn't the only way that sequels are derived from titles. Adjectives can be altered with intensifiers: witness "Grumpy Old Men" and "Grumpier Old Men". Of course, "Grumpiest Old Men" wouldn't have worked, which perhaps might be the real reason why rather than another sequel, the next Lemmon/Matthau pairing got a new title altogether.

Of course, even when the conventional "2", "3", or, in painful failure to recognize when the joke's gone bad, "5" or even "10" gets tagged on the end of a title, there are still ways of being unconventional about it. SC's favorite is the pattern used for the Naked Gun films, featuring the sequel numbers 2 1/2 and 33 1/3. Had O.J. Simpson not been himself, we might well have seen 444 1/4, a number never to be broken even by the James Bond films (all of which SC has seen at least twice, even the lamentable Living Daylights). Or the number might be facetious rendered in a foreign language, as was done for Hot Shots: Part Deux. Or the number might simply be rendered through a bit of clever typography; the gimmick of following MIB (Men In Black) with MIIB (Men In Black II) probably will only work once, but it worked well.

Although your host had only set out to catalog grammatical encodings for sequels at the start of this post, it ends up being a quite amusing demonstration of the arbitrariness of the relationship between grammatical forms and functions. All of the above forms serve the same function, namely the identification of sequels, but they don't really serve the same function anywhere else. Adjectival intensifiers normally don't tell you anything about the count of the nouns that they describe, and English plurals normally don't pick out "2" as opposed to any other number, even if we could imagine concatenating them to do so (presumably, nobody is ever going to try to indicate "Anaconda 3" with "Anacondass" -- but maybe in that unique case, they should). While these forms certainly don't act as substitutes for each other in other contexts, we have no difficulty picking up their meanings in this one. It's just one more reminder that our capacity to reanalyze and reuse grammatical forms will always exceed even the most diligent efforts to fit them neatly into little boxes and declare an end to the project.

El Caliente

SC and his wife were at El Pollo Loco this evening for dinner, where we were treated to a new, not necessarily welcome, sight.

A life-size cardboard cutout of a Hispanic male model perched over a grill was located next to the cash register, bearing the legend: "El Caliente, Master of the El Pollo Grill". Your host can prove he's not making this up; look here.

As yet more proof that advertisers would benefit from not smoking whatever it is that they presently make use of, the chief creative officer of the agency responsible is quoted in the press release as saying "El Caliente is nothing less than the spiritual leader of El Pollo Loco and its millions of male and female devotees." SC can reliably report that he did not in fact feel any urge to genuflect before the picture.

At the bottom of the cutout is a message promising that even if "El Caliente" isn't in the kitchen, he promises that his "passion, devotion, and unusually expressive eyebrow are". At this last one, your host smirked: "Ah-ha! Caught in a syntax error, are we?". He then called over Mrs. SC to gloat about his find. After reading it, she pronounced the sign grammatically correct, and expressed bewilderment at SC's behavior. How could we have reached different conclusions?

Although there aren't any pictures of "El Caliente" available on the El Pollo Loco site (but you can see the model here), Mrs. SC correctly noted that, like Alan Colmes, El Caliente has one eyebrow prominently raised above the other. The difference in our interpretations of the sign thus comes from integrating the visual cues of the picture with the text of the sign. Or, in your host's case, failing to do so.

Essentially, there are two potential parses of the sentence:

(1a) [passion, devotion, and unusually expressive eyebrow]-NP [are]-VP
(1b) *[passion]-NP, [devotion]-NP, [and]-CC [unusually expressive eyebrow]-NP [are]-VP

(Key for those inexperienced with the Penn Treebank: NP="noun phrase", VP="verb phrase", CC="conjunction")

The former case groups multiple singular nouns together in a collection which can be treated as a plural because it has more than one member; the latter case treats each noun as a singular item, all of which are separately coordinated with the verb, which must therefore be singular.

Your host isn't sure how likely 1b really is, as can be demonstrated by simply pulling out a single element and trying to make use of the singular verb form:

(2a) "Even if El Caliente isn't in the kitchen, his passion and devotion is."

That doesn't sound very good to SC's ear; reader intuitions may vary. This doesn't actually work better for me, even though I feel like it should (since a singular count noun is right next to the verb, the fact which triggered my intuition the first time):

(2b) "Even if El Caliente isn't in the kitchen, his passion and unusually expressive eyebrow is."

Since "passion" and "devotion" aren't ordinarily count nouns (yes, I know, "among X's many passions/devotions..." is a perfectly valid construction, but that's not the usage here), it seems to me that "is" and "are" ought to be interchangeable in 2a, and that "is" ought to be obligatory in 2b. More than that, I find myself thinking that some kind of parallelism ought to hold here, with an "is" in each clause. But I find myself thinking that "are" actually works better in both cases, an intuition that annoys me greatly, because it suggests that it was quite odd of me to parse the original a la 1b instead of 1a.

Nitpicking political speech

Neal Whitman, guest-blogging at The Volokh Conspiracy, graciously gave SC a mention on account of a recent post on recursive acronyms. Since, as Ryan Gabbard recently noted, Eugene Volokh prefers links back to other forms of recognition, as does SC -- there will never be a tip jar here -- it seemed appropriate to come up with a post on some recent Volokh topic. It didn't take long.

Despite the fact that SC didn't make it to a second post without referring to Slate, and despite the fact that your host cited the publication in his recent comments on Samuel Huntington and scale-based humor, the truth is, SC doesn't read Slate very often. So until he saw how annoyed Eugene Volokh is with "Kerryisms", he had no idea that Slate was even running such a feature.

The ostensible point of Slate's Kerryisms feature, as set out in the opening column, is as follows:

Since 2000, Slate has poked fun at George W. Bush for his torture—some say it's merely abuse—of the English language. Our "Bushisms" collection captures (as Editor Jacob Weisberg explains in his latest volume) the president's ignorance, incuriosity, laziness, and thoughtlessness expressed in frequent gaffes. Now that Democrats have settled on a presumptive presidential nominee, it's time to cast an equally cold eye on the pomposity and evasiveness of John Kerry.

Now, the controlling opinion on this genre of journalism is indisputably Mark Liberman's line:

You can make any public figure sound like a boob, if you record everything he says and set hundreds of hostile observers to combing the transcripts for disfluencies, malapropisms, word formation errors and examples of non-standard pronunciation or usage. It's even easier if the critics use anecdotes based on the perceptions and verbal memories of equally hostile listeners.

SC considers this view of "(politician)-isms" dead on, and has previously applauded Prof. Liberman for lighting out after some fairly ludicrous attempts to make "Bushisms" out of perfectly reasonable pronunciations. Since nobody at Language Log has yet taken a similar hack at the Kerryisms feature, and since Semantic Compositions attempts to be a rigorously nonpartisan blog, it seems appropriate to apply the Liberman philosophy to Slate's efforts to be bipartisanly asinine.

The red flags went up for SC as soon as he read Will Saletan's description of the methodology employed in parsing Kerryisms:

Here's how to read a Kerryism. The text below is Kerry's quote translated into plain English. Kerry's actual quote, however, is full of caveats and pointless embellishments. To read these, click the numbers above the text, which will take you to the caveats and embellishments, presented as footnotes. (Words in brackets before a number are what a normal person would have substituted for the ornate phrase Kerry delivered. To see the ornate phrase, click the number and read the footnote.) To return to the main text, click the number at the beginning of any footnote. To see the whole quote as Kerry delivered it, with all the caveats and embellishments, click here. To get back to this Kerryism, click "Return to English version."

A mercifully short example of the genre goes like so:

Question: But no regrets about those votes [for NAFTA and the China trade agreement]?

Kerry: [1]Sure.

—Democratic presidential primary debate, Milwaukee, Feb. 15, 2004

[1] I regret the way that they haven't been enforced,

Verbatim:

I regret the way that they haven't been enforced, sure.

Even the most uncharitable observer should have no problem determining what Kerry's opinion is in this case. Inferring from other examples of Saletan's attempts at deconstructing Kerry's speech, "Plain English" is just English without adjectives, adverbs or subordinate clauses:

Plain English: Let me just say that the abuse of Iraqi prisoners is unacceptable.

Verbatim: Let me just say very quickly that the horrifying abuse of Iraqi prisoners, which the world has now seen, is absolutely unacceptable and inexcusable.

So intensifying a statement is what passes for evasiveness? Apparently, SC has spent the last quarter-century under the erroneous assumption that emphatic rhetoric indicates particularly clear agreement with the point being made. Guess not, at least in Saletan, a dialect of English which SC will claim dibs on having identified and added to the generally agreed-on body of languages.

It's not terribly hard to figure out how this feature was cooked up. As political speech goes, the canonical example -- "I actually did vote for the $87 billion—before I voted against it." -- is what Talleyrand might have had in mind when he reputedly said "This is worse than a crime, it's a blunder". Like any politician not intent on going down in flaming defeat, Senator Kerry tried to frame his position as being something that both pro- and anti-war voters could support.

The problem with a statement like this, especially since there have been other examples, is that it's erroneous to characterize voting for drafts of bills as constituting a meaningful voting record, when only the vote on the final bill really matters. Neither the ACLU nor the American Conservative Union, both of which have long published annual scorecards evaluating legislators' fidelity to liberal or conservative principles respectively, counts procedural votes in their analyses, and SC would be surprised to learn that any other organization with serious lobbying clout behaved differently. But all that a few examples like this proves is that John Kerry is a politician. Is it really the case that no other prominent politician has ever made a statement trying to claim credit for being on both sides of an issue? SC rather doubts it.

This sort of thing wouldn't happen if media hacks didn't seek, as Prof. Volokh puts it, to fit stories into "stock plots". As soon as one incident occurs that has even the faintest whiff of prospective "meme-hood" about it, politicians are tagged for life. Dan Quayle couldn't spell to save his life after the infamous "potatoe" incident, a point which must surely rub him raw every time he sees a newspaper run a spelling correction. Gerald Ford, an all-American football player at the University of Michigan, had a reputation for clumsiness derived largely from a famous picture of him slipping on a staircase (in the rain, no less!) as he exited Air Force One. Michael Dukakis couldn't shake an image as a wimp after another famous photo of him looking rather uncomfortable while taking a ride in a tank.

Perhaps some sort of guilty bipartisanship motivates Slate's writers to feel that they have to push a storyline about Kerry's language problems in order to even the score with the overdone story about Bush. It would be far better for political journalism if they came to a different conclusion about maintaining balance, and dropped their quixotic pursuit of both "isms" altogether.

Political kabuki

Languagehat mentions a post at a blog called Far Outliers on the subject of "political kabuki". He mentions not having heard the phrase before; while SC has heard "kabuki" used in exactly the way that the original poster writes -- "a way to denote an empty political performance" -- he wasn't sure that the collocation was all that common. So in order to inspect the use of the term, off to Google. The numbers are sparse, and presented in descending order:

"political kabuki": 91 hits
"kabuki" used alone at National Review: 32 hits
"kabuki" used alone at Reason: 12 hits
"kabuki" used alone at The New Republic: 9 hits
"kabuki" used alone at The American Spectator: 8 hits
"kabuki" used alone at Mother Jones: 2 hits

There's just one hit from Z Magazine, but not in a political use. So much for it as a meme among the political scandal rags (as SC affectionately refers to the partisan commentariat).

Your host then formed a hypothesis that such a usage would be more out of bounds among mainstream newspapers than political magazines. With such low frequencies to begin with, this is all but untestable, but some amusing usages turned up nevertheless:

At the LA Times, 6 hits, all nonpolitical, including one line with the context: punctuates statements that he believes should be apparent to anyone harboring a shred of sense with hand and facial gestures suggesting Borscht Belt Kabuki". (The article's not available; this is just from Google's snippet of text.)

From the New York Times, 53 hits, all cultural, at least after a sort. What to make of a movie review that starts off " Viva Las Vegas, one of Elvis Presley's most popular vehicles, adheres as rigidly to formula as a Kabuki dance"?

From the Washington Post, 21 hits, including a Bill Clinton speech transcript, where the former President comments: "You know, this book, "Spin Cycle," it implies that this kabuki dance between the White House and the press is some kind of a recent phenomenon." It occurs in a few online discussion contexts, hosted by the Post, as well. A chat participant compliments Post humor writer Gene Weingarten: "Your line about the Raelian spokeswoman that she "looks like a kabuki Tweety Bird" was dead on." Meanwhile, Post movie critic Rita Kempley replies to a comment with "I liked Jeremy Irons as the Kabuki Morloch".

The Wall Street Journal isn't searchable online without a subscription (which SC has, but links wouldn't work). Opinion Journal is, though, and it appears that columnist and former Reagan speechwriter Peggy Noonan has a bit of a taste -- but just 2 hits worth -- for the kabuki metaphor herself.

Taken as absolute numbers, this isn't exactly suggestive of "kabuki" as a hot metaphor for the political process, although it's clearly not an unkown one. It would be instructive to compare these frequencies with other political metaphors for catchiness, but a few cursory attempts by your host ("lock box" and "straight talk") proved to be not particularly interesting either (the latter complicated by its existence as something other than a John McCain 2000 primary slogan). Kabuki's an interesting metaphor for politics, but it seems destined to remain the preserve of terminal punditry addicts rather than a mainstream phrase.

A barbeque for vegetarians?

One of SC's souvenirs of the good old days the 1999 stock market bubble is a subscription to an e-mail list run by a clown and stock tout named Steve Harmon. He hasn't paid attention to Mr. Harmon's advice in years, but hey, it's free, so your host is too lazy to unsubscribe. Aside from that, occasionally Mr. Harmon will say something so bizarre that it's good for a laugh or two.

Anyways, this evening, SC got a mailing touting the Semantic Web that started off promisingly with "Did you know your dog is a lot smarter than the world's fastest super computer?". But the real gem comes later, when he's trying to explain how cool the SW really is, and writes:

Computers have about as much contextual knowledge, which I would term "awareness," as a vegetable at a BBQ for vegetarians. BBQ'ed carrot anyone? Computers and context are like Weird Al Yankovic winning a Grammy for best male vocalist.

As Dave Barry likes to say, I am not making this up. Your host also really loves the scare quotes around "awareness". Also for the record, SC considers Weird Al a much more enjoyable singer than the actual winner of last year's Grammy for "best male pop vocal performance" (Mr. Harmon really ought to have specified the category he had in mind).

Actually, the metaphor is worth paying attention to for linguistic reasons, aside from being a guide to how not to write. Mr. Harmon is trying to mix two well-known metaphors: 1) the association of mentally deficient (or brain-dead) people with vegetables, and 2) the association of things people that are in trouble with being cooked (i.e. "his goose is cooked"). The problem is that vegetables are rather semantically anomalous as the objects of a barbeque. And who ever heard of a barbeque for vegetarians, anyway?

This sort of anomaly used to be analyzed as an example of the relationship between syntax and semantics, and was called a "selectional restriction", the idea being that certain words couldn't be followed by certain semantic classes without creating a sentence which was at least odd, if not ungrammatical. For example, "I drank the steak" is anomalous because "drink" needs to be followed by a type of liquid (spare me any contexts involving putting a steak in a blender; SC doesn't want to imagine it). There's nothing ungrammatical about the sentence, in the sense that "drink" is followed by a noun phrase as it's supposed to be. Examples like this make it clear that not all constraints on sentence structure are purely syntactic. Apparently, though, some writers are more aware of these issues than others.