Semantic Compositions

Brunch at the modal buffet

Mrs. SC's birthday comes up in a few days, and tomorrow, we're going for brunch with her parents. As of this writing, though, the question of where remains outstanding. So this post is prompted by a conversation we just had about a possible destination:

Mrs. SC: I don't want to go to the Flying Feast ([not its real name, but it's by an airport -- ed.]) because I don't want an all-you-can-eat buffet. I feel like it means "eat all you can".

SC: But if you ate less, it wouldn't be an all-you-can-eat buffet. It would be an "all-you-feel-like-eating" buffet.

Mrs. SC: No, it would be an "all-you-would-eat" buffet.

So which is it? SC readily buys into the idea that it can't be an all-you-can-eat buffet if you don't eat all you can, but it might not be the case that you eat all you would if you could specify the amount of food in advance. Maybe you're on a diet, and wish there would be some externally-imposed portion control. Or maybe you go out with other people (who goes to a buffet alone?), and feel like you eat less than you otherwise would because you don't want the other people thinking you're a pig. No single modal verb seems to cover all this, though, in any tense -- not "can", not "will", not "shall". Not "could", not "would", not "should".

But somebody would have a great tag line for a competitor to Souplantation if they used the slogan "The all you should eat buffet".

Inching forward

This morning, your host received an e-mail with a back-formation he hadn't heard before. Specifically, the message said:

"We'll be reviewing the inchstone transition plan."

Inchstone was new to SC, but it was immediately recognizable as a derivation from milestone. In the context of a series of small steps to be taken, it makes sense. Not having heard it before, your host wondered if there's evidence that it's catching on, or if it's just an obvious derivation that might have been independently coined by a variety of different people (including the individual quoted above). For the sake of completeness, we'll look at the whole spectrum of possible English-system length measurements:

inchstone: 965 hits
footstone: 108,000 hits -- but a preexisting word meaning "grave marker"
yardstone: 871 hits
furlongstone: 5 hits, with 3 unique ones
milestone: 29,000,000+ hits (some of which are just names)
leaguestone: 7 hits, 3 of which are genuine

At least anecdotally, it appears that the milestone derivatives are used in a very self-conscious fashion to indicate that the user is trying to calibrate the degree of what they're discussing. Some examples, with emphasis added by SC in all cases:

The recovery team will create detailed miniature milestones known as “inchstones.” (link)

"Resist overkill," Humphreys said. "If you can measure the status of a project at the milestone level, don't go to the 'yardstone' level. If you can do it at the 'yardstone' level, don't go to the 'footstone.'" (link)

And Easter Weekend siguals [sic] a milestone--leaguestone?--of sorts for Isenfir: The Shire will be old enough to drink. Not many of its constituent members, but the Shire. (link)

Small milestone? Not a very elegant or satisfying description. It seems like there should be another term - a furlongstone, or its equivalent - for the marking of this small unit of progress. (link)

Although the metaphor of calibration along a scale is the same, metric uses appear to be more uniformly ironic:

[Blog Post Title] Another Milestone Reached...well, this is more like a centimeterstone. (link)

[Blog Post Title: Meterstone] I'd call it a mile stone, but it is closer to 1000 which is more the metric thing. (link)

today is the thirteen month mark for my boyfriend and i...
its not usually considered a huge milestone (if i'm canadian shouldn't it be kilometerstone?) (link)

There are actually 479 hits for kilometerstone (versus just the one above for centimeterstone, and less than 100 for meterstone). Perhaps this is due to the fact that kilometers are the closest metric unit of measure to a mile, or perhaps there's merely some principle of self-restraint where one pun in a novel coinage is enough. Inches or centimeters, many of these groaners are just begging to be...stoned.

I'm bringin' syntax back

On the occasion of this evening's Grammy awards, SC is reminded of a topic he's been meaning to address for a while, specifically a song by Justin Timberlake. It should be noted that your host never listens to Mr. Timberlake's ouevre, and is familiar with his "music" entirely and exclusively through exposure to it at the gym. Aside from the boy-band taint, SC despises Mr. Timberlake for the associations he's created in the public mind with the name "Justin", formerly a sign of discriminating good taste among those parents sophisticated enough to use it.

But back to the subject at hand, a ditty of Mr. Timberlake's titled "SexyBack". This artless title manages to combine one of the most annoying typographical conceits of our time -- the gratuitous elimination of spaces between words -- with an ugly grammatical choice. One might think from the juxtaposition of these words that the song would be about someone having a particularly sexy back, and that the title would reflect a construction like NP-[Adj N]. And one would be wrong. The first line of the song goes, "I'm bringin' sexy back", a construction which requires the title to be extracted across phrase boundaries; i.e. VP-[V-bringin' NP-[N-sexy] PP-[P-back]]. If you're finding this page through Google, and don't understand extraction, think of it this way: You can say "What's J.T. bringin' back?" or "Where is J.T. bringing' sexy (to)?" (that latter one is awkward to your host's ear, though), but "What where is J.T. bringin'?" is just wrong.

If those representations are still a little too abstract in making the point, see if you can recognize the classic songs that would be mangled by the application of the operation we'll call Timberlake extraction:

• ItOnHomeToMe
• HerInSomeDay
• YourFriendsBehind (thanks to Neal Whitman for reminding SC of this one)

Get this man a copy of Barriers. Or maybe a copy of The Cambridge Grammar of the English Language.

A la modal

SC has been watching with amusement the outpouring of interest at Language Log over Dick Cavett's thoughts on language. Aside from Mr. Cavett's status as a usage crank, there is an important presupposition in need of explicit cancelling, but your host will get there in a roundabout way.

When SC first started working for his present employer, in 1999, he was a mere 23 years old. The unspeakable horror known as Episode I had just recently been released, and so Star Wars nostalgia was a popular topic of conversation. At the start of a meeting where SC was the youngest person in the room by a good 20 years, several people expressed their memories of standing in line to go to the original movie. Mistaking an excellent opportunity to keep his mouth shut for a chance to perform some good-natured needling, your host promptly spoke up with, "I wish I remembered Star Wars coming out, but I was pushing 1 at the time".

It is in that spirit that SC gets to his point: who is Dick Cavett, and why should we care about his particular brand of usage crankiness? A cursory review of his biography on Wikipedia suggests that he formerly made a living as a comedian, a fact not at all evinced by his not-at-all funny debut column (there were some recognizable attempts at humor, but SC declined to laugh in order that the bar should be set at least high enough to limbo under). Rather like Star Wars, you seem to have needed to live through the '70s for Dick Cavett to have any sort of visceral appeal. Your host will at least credit him with having been funny at some time in the past, as Wikipedia also quotes him having said:

I went to a Chinese-German restaurant. The food is great, but an hour later you're hungry for power.

That actually did get a chuckle from SC, but then, he also likes the section on Hitler jokes from the Big Book of Jewish Humor. Not so good for Mr. Cavett's contemporary relevance.

But now to Mr. Cavett's work as a usage crank. Prof. Liberman remarks of him:

He does threaten that "if I ever find myself once again with the senator who spoke of his 'incredulous' experiences, I shall pop him one" -- but has anyone who said "I shall pop him one" ever actually done so?

The answer to the rhetorical question is obviously no, but not necessarily for the obvious reason. SC was actually...popped...more by the modal verb preceding that phrase than by the blustery use of the phrase itself. A Google search for "pop him one" turns up 507 hits; adding that "shall" in front drops it to 0, as does "must". "Might", "could", "would" and "should" all fared similarly, varying between 4 and 40 hits. Lesson: if you're trying to affect rough speech, using constructs that mark you as educated is a Bad Idea.

But then it occurred to SC that maybe Cavett's problem wasn't the "shall", but the "pop him one". Maybe he's stuck in a 1970s time warp where he's considered funny, and hasn't caught up with the more recent usage "tear X a new one" (where "one" is synonymous with a bit of posterior anatomy). With 718 hits, the usage rate on the web is pretty much comparable to "pop him one", so that's out (and oddly, the polite form of the new phrase gets 3x as many hits as the unexpurgated version). One thing that doesn't change, though, is the near-total absence of hits following modals -- 0 for "shall" before both forms, and just one for "must" before the aforementioned vulgarism. "Would" is the only modal to break 40 hits before "tear him a new one". So we can only conclude that despite Dick Cavett's declinist views about usage -- "I see it as a glass half empty — and evaporating quickly" -- some things haven't changed.

[A methodological note: Obviously, most (if not all) of the usages of both "pop him one" and "tear him a new one" that Google has catalogued couldn't predate the Web. A cursory inspection of the results from both searches suggests that the users are by and large different groups of people, but it's practically impossible to tell if the users of "pop" are uniformly significantly older than the users of "tear" from the web alone. I can't find a dictionary citation of "tear him a new one", and the earliest OED citation of "pop him one", with the dummy ditransitive object, is 1958. They're both contemporary enough that Google's corpus is probably not very different from what might be gotten from browsing popular literature of the '60s-'80s as far as "pop" goes, although strictly speaking, it's possible that people's knowledge of the newer form might influence how they use modals with the older one.]

McDonald's, "burgerhouse"

SC is on the mailing lists of too many restaurants to count, but he always looks forward to the uniquely self-promotional gems* produced by Noel Turner, the genius behind Turner New Zealand. Readers who don't subscribe to Gourmet or Bon Appetit may not recognize the name from anywhere, since his only restaurant is in Costa Mesa, CA, but Mr. Turner has long been one of the premier suppliers of high-quality beef and fish to the restaurant industry. Since he does such a great job, SC is all too happy to plug for him. But on to the linguistic topic at hand:

For the upcoming Valentine's Day (Men: You have 28 days left to get your act together. SC warned you.), Mr. Turner sent out a missive describing his planned five-course menu, including the following line:

This time of year always brings a special feeling of love and happiness to the hearts of many, and I wanted to let you know that 'Love Is In The Air' at our Turner New Zealand Steakhouse & Seafoodhouse on Wednesday evening February 14th!

It's only mildly corny, but your host's attention was drawn to the word "seafoodhouse", which was a novel coinage to him. It makes obvious sense as a back-formation from steakhouse, but does anyone use it? Off to Google, which gives 706 hits for a search on "seafoodhouse" alone. However, this is a grossly misleading count, as the original pages end after only 94 hits, a good number of the pages in question are spam, and more than a few simply contain the URL of a defunct website for a possibly-closed restaurant called the Chatham Fish House in Massachussetts. A couple of hits also just contain the URL of Landry's Seafood House, a still-going restaurant in Texas. A few more hits are from archived versions of prior Turner e-mails (and somehow, SC missed the usages).

So the usage of "seafoodhouse" as a distinct lexical item appears to be pretty much unique to Noel Turner. But this piqued SC's curiosity further. How about other (food)-house items?

It turns out that this is a pretty sticky research problem using Google alone. "Chickenhouse" returns 33,000 results (the usual caveat about Google result sets of this size obviously applies), but most of the hits in the first 5 pages (after which your host quit looking) are actually for places named "Chicken House", and simply making use of a space-free URL, or they're spam, or Google thinks it knows what SC wants better than he does, because the page doesn't have "chickenhouse" anywhere in the source HTML, not even in a filename (see here for an example). "Pastahouse" yields just under 14,000 hits, but suffers from all the same problems as chickenhouse. It does yield an authentic example from a restaurant in Denmark, but this is vitiated somewhat by the likelihood that the restaurant owners don't speak English. "Burgerhouse" yielded the same size result set as pastahouse, and offered a genuine result from a place called Gordy's Burgerhouse in Texas. While a search for burgerhouse together with McDonald's didn't yield any actual examples of someone describing the place that way, it did turn up a second legitimate English usage -- of a Hawaiian McDonald's nestled in the former location of a restaurant called Kenny's Burgerhouse.

Having just about run out of primitive food types, your host also tried "italianhouse", "mexicanhouse", and "chinesehouse", all with apparently entirely irrelevant results. SC didn't go into this expecting a productive usage, but his working hypothesis was that there would be at least a couple of people with Noel Turner's sense of humor and playfulness with language, and that they'd run restaurants. There aren't. The man really is an original.

*Turner's signature linguistic cue is the making of every ingredient into a proper noun, specifically a Turner Noun. Here's his description of the upcoming Valentine's Day meal:

We will be celebrating Valentine's with a Five-Course Extravaganza serving the finest quality organic and all-natural food on the planet. You will begin the evening with your choice of Turner King Salmon Sushi or Turner Clams and Turner Calamari; all from the cool, clean oceans of New Zealand. Next, enjoy our 'raved-about' Turner Lobster bisque; the real thing, made 100% from Turner Lobsters. Move onto the third course, a choice of organic Caesar or Herb salad. Your fourth course, the main event of the evening, our acclaimed free-range Turner Lamb Rack or our world-famous Turner Beef Filet Mignon, 100% grass-fed and all-natural; no antibiotics and no hormones. Or if you are in the mood for the finest seafood available, then delight with our ocean-fresh Turner Orange Roughy or Turner King Salmon. Relax over your fifth course, while you enjoy our Mama Turner's Pavlova, Brandy Cherry Creme Brulee or freshly chocolate-dipped organic strawberries. At your pleasure, we have paired each course with an exquisite New Zealand wine.

He has explained this in the past by claiming that if he didn't specify that everything was "Turner (Food)", you might think he was cheating and slipping something that didn't meet his standards into his menu. Verbose? Yup. Pompous? Gloriously so. It all really is that good, though.

Syntax up with which I will not put

Today is the start of the annual Home Entertainment Show, taking place just a few short miles from Victory Mansions. While your host might write about what he sees there over the weekend, today he'll simply use it as a peg to bring up a blog post by Fred Manteghian, a contributor to the surprisingly long-lived Ultimate AV.

Mr. Manteghian's post dealt with the 5 top-selling front projectors in the U.S. today, which is quite tangential to our topic. In the course of introducing his subject, he writes:

Yup, the U.S. of A. Love it or leave it, we're the country everyone wants to break into, not out of. Although if I ever end another sentence in a preposition, I'll agree to be expelled.

Ah, the old canard about the ungrammaticality of ending sentences in prepositions. This occasioned a commenter to respond with a delightful rejoinder on the subject, and while SC can't vouch for the authenticity of what follows as an Anthony Burgess quote (the aforementioned blog post is the only such citation listed in Google), it just has to be passed on:

A preposition is a good word to end a sentence with - it's a rule of Latin, anyway, not English. As Anthony Burgess once wrote: "What do you keep bringing that book you know I don't like being read to from out of up for?."

Got argument structure?

Neal Whitman has provided an interesting piece of data that clearly demonstrates get with multiple different argument structures at once. Your host can't think of any syntactic work that accounts for a single verb being in multiple grammatical categories at once. It's sort of a reverse serial verb construction; instead of multiple verbs operating on the same object, with possibly clashing mophology (but the same intended tense/aspect/valence/etc.), it's just one verb with multiple lexical representations (or functional projections, take your pick) operating simultaneously on multiple objects. It's quite possible that your host is simply ignorant of an important piece of work on this point, but he's fairly sure he's never read any sort of syntactic analysis that doesn't assume just one allowable argument structure at a time for a given instance of a verb, however convoluted the operations that produce it might be.

Right now, the only idea SC has for a way out of the conundrum is to suppose that something is actually going on with the verbal complements in question (make and brush), and that the apparent objects of get really aren't. But Prof. Whitman's "unpacking" of the uses of get strikes him as correct, and this is causing enormous headaches for your host.

All Bally's ads changed!

Your host is presently in New Orleans (more on that shortly), but that doesn't actually explain his absence for the last week; he's simply been running low on ideas again. However, travel usually helps refill the tank, and this time is no exception. So here goes:

Gambling is legal in New Orleans, albeit apparently only in certain areas zoned for it, and so there are billboards all over the city advertising the local outposts of the Bally's and Harrah's chains. The ads for Bally's in particular have been causing SC's brain no end of syntactic distress, as they are all based around a slogan of highly questionable grammaticality (always in all-caps): "ALL SLOTS CHANGED!".

To get the semantics out of the way, further explanatory information on a number of the billboards indicates that the intended meaning is "all of our slot machines have been altered to produce payout odds more favorable to the player". (For example, some billboards include the added slogan "All slots, all loose, all the time".) But your host can't get that meaning out of the phrase as it stands. It would be better with an auxiliary, as in "All slots are changed!", or even better with a perfective, as in "All slots have been changed!". As the phrase stands, the only interpretation that SC can recover from the raw "ALL SLOTS CHANGED!" is along the lines of "all of our slot machines have been physically removed and replaced with different units".

Mrs. SC argues that it's not as categorically bad as SC claims. Syntactically, she sees no difference between it and, say, "My mind changed", for which we can construct some valid contexts (A: "What's different?" B: "My mind changed."). Perhaps so, but even in those contexts, your host would expect to hear something more like "I've changed my mind".

On taxi cabs, the phrase gets embedded in a larger noun phrase which makes it marginally more acceptable to SC: "Tell the driver, 'Take me to the place with ALL SLOTS CHANGED!". It would still sound better to your host as "Take to the place where all the slots have been changed!", but it's an improvement.

Self-eating food

SC isn't sure if this is an entirely new use of a morpheme, but if once is a novelty and twice is a trend, then our food is becoming more active. At least that's the message that some marketers seem to have latched onto these days.

The first indication of this came to SC a few months ago, when he first heard commercials for a new 7-Eleven product known as the -- go ahead and laugh -- "Southwest Jalapeno and Cheese flavored Big Eats Griller Sausage" (judging by the placement of the trademark symbol, the fact that it comes in a hot-dog bun doesn't seem to be part of the name). Even if your host was inclined to eat pork or jalapeños -- and this doesn't look likely to change him -- there's no way he'd ask for anything at a 7-Eleven by using a nine-word name. But more interesting is the "-er" morpheme; it's not a "grilled sausage", it's a "griller" sausage. What, it cooks itself?

Then, just today, SC came across "Goldfish Sandwich Snackers" (scroll down to see them), a snack consisting of two Goldfish crackers with a filling between them. Your host isn't quite sure what a "snacker" is supposed to be, aside from someone who eats a snack. So are these crackers self-eating?

Mrs. SC adds as a possible additional example of this morpheme the frozen food sold under the name "Poppers". SC is a bit unsure of whether or not it's the same usage, since he thought it was meant to be a description of the action performed in eating them (as in, you're supposed to pop them into your mouth). If that's the intent behind "snackers" -- you're supposed to snack on them -- that would make a certain amount of sense as well. But unless 7-Eleven wants customers to do the cooking, they can't mean that you're supposed to come into the store and grill their sausages yourself. (Although maybe they do; one of Southwest Airlines' rules is "Think how we can transfer the workload to the customers".)

So what's going on with "-er"? Among its various meanings are "one who does X" (i.e. "killer" is "one who kills"), or intensifying whatever it's attached to ("cold" -> "colder"). But neither of these really makes sense in the examples given above. It's actually quite hard to see what "er" adds in any of the above examples aside from some additional phonological material; would 7-Eleven's glorified hot dog make any less sense if it was called the "Big Eats Sausage" or the "Big Eats Grilled Sausage"? Would Pepperidge Farm's latest come off any worse as "Goldfish Sandwich Snacks" -- or if they have to have that "-er", as "Goldfish Sandwich Crackers"? Perhaps it's wrong to be looking for a grammatical function here, and it's really another phonestheme.

Welcome to Boston. Now go home.

While SC is home and working, Mrs. SC is on the road interviewing for possible residencies. After interviewing at a prestigious hospital affiliated with America's oldest university, she checked in this afternoon to inform your host about a very interesting billboard seen in the downtown Boston area.

It seems that a local anti-gun campaign has produced an ad that reads:

Welcome to Boston. You're more likely to live here.

Mrs. SC's first read, and that of your host's, was that it meant "you're more likely to be a resident of this city than a tourist if you're reading this sign". But of course, if it's associated with your expected lifespan, the proper reading is "if you're here, you are less likely to die over a given time period than in some other place". The difference comes down to a delightfully ambiguous attachment for "here".

Bracketing for the sentence runs roughly like so:

Tourism: [you're more likely [to live here]]
Lifespan: [you're more likely to live [here]]

Not having seen the sign, it's not clear to SC if the ambiguity is readily resolved by some other visual element. It's an odd phrasing, though, and one wonders exactly what effect they were going for that wouldn't have been captured by something like "Welcome to Boston. Our traffic will kill you before a handgun does", or maybe "Welcome to Boston. Now that the Red Sox aren't losers, our residents are less likely to blow their brains out". At least neither of those slogans are open to multiple interpretations.

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.

All hail King Taco

A friend asked an interesting question this evening. I have no idea if there's anything to it, so I'll just throw out the data, and let the tortilla chips fall where they may.

This friend observes that restaurants which either serve primarily American food -- or at least Americanized -- seem to order certain words in their names differently than restaurants which are primarily ethnic.

American: Planet Hollywood, Taco Bell, Burger King.
Ethnic: King Taco, Noodle Planet (if you're from the L.A. area, you'd recognize these as Mexican and Chinese fast food, and considerably more ethnic in their appeal than Taco Bell).

They're all made by stringing nouns together, but to the extent that we can discern a pattern (once is an exception, twice is a rule, right?), the words which are common to both groups show up in opposite positions. And there's something mildly -- I cannot deemphasize this enough -- plausible about the idea that "King Burger" is less natural to the native English speaker than "Burger King". Your host is also willing to play with the idea that "Hollywood Planet" doesn't sound right, but on the other hand, "Planet Noodle" doesn't sound very good to SC, either. "Bell" is the founder's last name; SC submits that "Bell Taco" sounds ungrammatical, in part because it's almost begging for a possessive (i.e., "Bell's Tacos").

As I said before, I'm not sure there's anything going on here, but the data is sparse. And there are all sorts of confounding factors -- measuring the relative English fluency of the people naming restaurants, testing for selectional restrictions (maybe "planet" is best preceeded by food items, and best followed by place names), testing for phonological factors...lots of things could be involved. Of course, if your host felt strongly enough about wanting to know, he'd assemble a corpus of restaurants whose names include "king" and "planet", and try to gather the necessary supplementary data to figure out if these words really do have different orderings depending on the fluency of the people using them. In the meantime, however, SC suspects his friend is...out to lunch.

It's not all about parts of speech

This morning, Bill Poser wrote on Language Log about a horrifically bad special from the BBC on historical linguistics. In the process, he mentioned a documentary from PBS' NOVA series (transcript here), which was made in response to it. Judging by the transcript, it was a well-informed production, bringing together a group of world-class linguists to talk about the historical method. Your host wishes to focus attention on one of them in particular, the late Joe Greenberg.

Greenberg did cross-linguistic studies on a scale that most of us can only dream of, starting with a 1963 paper modestly titled "Some universals of grammar with particular reference to the order of meaningful elements". In that paper, he put forth a model for making generalizations about language which contained two categories of interest: exceptionless (or "absolute") universals, and statistical universals, described as those occurring with “overwhelming greater than
chance frequency".

Unlike the typical parsing-obsessed syntactic theoreticians of today, Greenberg spent a lot of time describing things at the abstract level of S(ubject), V(erb) and O(bject). That's not to say he wasn't concerned with getting word order right, but given just one lifetime to do research, he preferred to worry about discovering general facts like "Languages with dominant order VSO are always prepositional" (his third absolute universal from the 1963 paper). At some future point, all the niceties about phrase-internal structure could be described in terms that were known to be rock-solid, because the examples and counterexamples would have already been worked out. Greenberg's project wouldn't let him -- or even a few generations of successors -- live long enough to see this concluded.

The mindset of Greenberg's research is thus very top-down: given a large collection of languages, find some general truths, and then start applying them specifically. This contrasts radically with the bottom-up approach in most of theoretical syntax today: work on one language, and infer some general principles. Yes, argumentation in syntax often relies on showing that a principle proposed for one language has neat applications in another language, but as a general rule, it's hardly ever the case that principles are explicitly tested over anything like the thousands of languages that Greenberg worked on. Which isn't to say that anything like that is practically possible, unless there was a universal agreement to freeze syntactic theories for a few years and go code millions of examples into a database using one, consistent framework. Which would then be obsoleted as soon as it produced results inducing changes in the theory that had been used to produce it. So it's not necessarily a bad thing that theoretical syntacticians don't sit around waiting for validation on the grand scale.

Having said that, SC has long lamented that he only became aware of much of this too near the end of his grad student life. Not that corporate America, where SC chose to work from the beginning, funds this sort of research, but it's as true for other people who stay on the academic side as it was for him. A greater appreciation of large-scale cross-linguistic facts would do a lot to fix the theoretical squabbles that plague the field today (recognizing, of course, that it is precisely such squabbling that makes linguistics fun).

Andrew Sullivan does infixation

SC readers know that Andrew Sullivan is often a source of material (here, here, and most recently, here). Your host tries to avoid discussing Mr. Sullivan's (he's entitled to Dr., but SC hasn't seen him insist on it) views in this forum; our purposes here are largely linguistic. This one's a beaut.

In his latest article for the Sunday Times of London, AS attributes a coinage to Mickey Kaus:

Here's a word that deserves to be entered into the political lexicon. The blogger Mickey Kaus coined it. It's "pandescender."

As Mr. Sullivan explains it, the word is a combination of "pander" and "condescend", and is a verb meaning to simultaneously pander and condescend. In the usages of both Sullivan and Kaus, it accepts the suffix "-er" to derive the noun "pandescenderer", meaning "someone who both panders and condescends". Since the plain verb form contains the string "-er" itself, we must assume that it derives from one of the words that was used to make the compound, and thus that Mickey Kaus' grammar (and Andrew Sullivan's, since this doesn't bother him) features a productive infixation rule, like so:

pander + condescend -> pan- + descend + -der -> pandescendder

Since English does not generally permit geminate (for nonlinguists, "long") consonants, we may also assume that some phonological process reduces the output of the rule to "pandescender".

Now, it is true that most English speakers actually allow at least some infixation, even though it's not taught in schools. SC still remembers his very first linguistics class, where the professor demonstrated this fact by uttering "in-f***ing-credible". Let's see if we can take a stab at characterizing the morphological process that allows this (note: the following analysis is less serious than SC's usual standards of rigor, such as they are).

Note that in both cases, the infixed word is inserted after the first syllable of the infixed word. So there's a positional effect. It's hard to identify a clear phonological rule; while place of articulation remains stable at the infixation point for "pandescend", there is no common phonological feature at the junction of "in" and "f***". So maybe it's just a positional issue. Let's try a few new coinages, trying to keep things in the same domain of lexical semantics to eliminate effects there:

Based on the obliviousness of the average cell-phone user to the outside world, how about "chatter" and "ignore"? "chatignorter". Maybe just "chatignorer". Or maybe "igchatternore". The first two sound too much like someone in the process of ignoring a chat, which is the one thing cell-phone users don't do; meanwhile "igchatternore" has a certain onomatopoeic quality to it that your host rather likes. It sounds like something that would come out of the mouth of someone not paying attention.

Or perhaps from the combination of "mobilize" (which Merriam-Webster gives as a synonym for "drive") and "imprecate" (a nice way of saying "curse"), we can fashion a useful description of what people do while stuck on freeways during rush hour. "moimprecatebilize" probably violates most people's intuitions about the sounds of English, never mind the grammar, but "immobilizeprecate" felicitously adds the sense of being immobilized.

It seems abundantly clear to your host that infixation is actually capable o