OK, everybody, listen up. Sit back, hang on to your chair, and brace yourself. I'm going to violate the ultimate taboo.
No, I'm not going to defend the Sapir-Whorf hypothesis. I've already done that. And I'm not going to argue that chimps, dogs and parrots are starting down the road to language. I'm still thinking about that one.
I'm going to venture into territory where no linguist has ever dared to set foot -- until now. I'm going to praise William Safire.
The poor guy writes a few sensible paragraphs on gone missing in the NYT magazine, and he gets artfully slammed for it by our own David Beaver. Now, sometimes Safire deserves a rebuke. Many of his 50 "fumblerules" are recycled nonsense from centuries of self-appointed experts who invented linguistic principles out of thin air and tried to impose them by force of cultural authority. However, his discussion of gone missing is reasonable, and even uses some grammatical terminology correctly.
David expresses surprise that Safire writes as if "the opinions of people who don't have Language Maven stamped on their business cards actually count". This is unfair to a man who once wrote a book entitled "In Love with Norma Loquendi", referring explicitly to Horace's dictum that language will change "si volet usus / quem penes arbitrium est et ius et norma loquendi" ("if it be the will of custom, in the power of whose judgment is the law and the standard of language").
Safire starts with a question from a reader, Daniel Baldwin of New York: ''My intuition tells me that the term goes missing is grammatically incorrect,'' he writes. ''Here is a possible explanation: It is proper to link goes with a gerund (e.g., goes fishing) but not with a participle (e.g., goes missing). Am I on the right track?''"
Baldwin is all wet, actually. Go fishing involves a different sense of go: transitive sense 3 "to engage in" in the American Heritage Dictionary's entry, rather than intransitive sense 10.b. "to come to be in a certain condition". And fishing in go fishing is probably not a traditional gerund at all, since it can't be replaced by a noun (in contrast to "I regret destroying it" vs. "I regret its destruction").
Safire responds "A technically correct track -- I salute all gerundologists -- but headed in the wrong direction." I take this as a brief, polite way to pat Baldwin on the back for using words like gerund and participle (which Safire is probably as uncertain about as almost everyone else is), though in the service of an analysis that's wrong. Safire goes on to say "This is a tale told by an idiom that leaves many of its users vaguely uncomfortable", and I think this is correct, if a bit compressed.
Here at Language Log, we're not constrained by the petty word count restrictions of a New York Times column, and so I can go into more detail in support of Safire's point of view. As he explains, "one sense of to go is ''to pass from one state or place to another'", and this is the sense that is involved in go missing as well as many other expressions. His analysis is not new, but it's correct. The OED puts go missing under sense 44 of go:
44. To pass into a certain condition. Chiefly implying deterioration. a. With adj. complement: To become, get to be (in some condition). (Cf. COME 25a.) to go less: to be abated or diminished. Also with n. complement: to become, use, or adopt the characteristics of (something specified); to go all ____: see ALL C. 2c; to go bush: see BUSH n.1 9e; to go missing: to get lost; to go native: to turn to or relapse into savagery or heathenism; also transf. (cf. FANTI b); to go ____ on (someone): to adopt a particular mode of behaviour towards or affecting (that person); to go public: to become a public company.
Safire notes that go missing is "British English", and supplies an earlier citation than the OED does. He quotes "a naval correspondent for The Times of London on Aug. 10, 1877, in a dispatch about the Turkish armies in the Balkans'' as writing "I was obliged to return to Adrianople to get some supplies, as a box which should have reached me at Tirnova had gone missing.'' The OED's earliest citation is from 1944:
1944 E. BENNETT-BREMNER Front-Line Airline (1945) viii. 50 Qantas Empire Airways have been called upon to conduct searches for missing aircraft, and it was only natural, therefore, that being ‘Johnny-on-the-spot’ they should be asked to join in when aircraft went missing.
As the OED suggests, "go PREDICATIVE" often suggests that the predicative is a kind of deterioration: go bad, go ballistic, go bananas, go bankrupt, go blank, go cold, go crazy, go dead, go gray, go Hollywood, go lame, go mad, go native, go numb, go nuclear, go nuts, go sour, go vacant, go wrong.
Sometimes the corresponding positively-evaluated condition doesn't work: go bad but ?go good (in the sense of "become good", not as an informal version of "go well"), go crazy but ?go sane. However, there are some positively evaluated conditions in common inchoative collocations with go: go live, go platinum, go blonde.
On the other hand, there are also common adjectives expressing deterioration that don't seem to work well with go, preferring get instead: ?go sick, ?go fat, ?go dizzy, ?go sleepy, ?go antsy.
Note also that the meaning of the predicate is often restricted: thus went dead is fine if you're talking about a phone line or a radio, but doesn't work to mean that an animate being died.
All in all, the collocational propensities of go, in this construction, seem much more like derivational morphology than like normal compositional syntax. A good shorthand term for this situation would be idiom, and that's what Safire calls it. Good for him.
I need to temper my praise with a bit of criticism. Safire says that gone missing "may well stretch our hard-wired sense of syntax." This is completely nuts, at best. Is Safire saying that we are genetically disposed against making an inchoative out of a motion verb and a predicative? This would be weird biology and worse typology. Is he saying that our genes have been programmed by evolution to resist inchoatives that involve becoming lost? or include two-syllable predicative words starting with /m/? This is beyond even the bounds of journalism.
No doubt Safire means something much less grand. I bet his train of thought went something like this: "gone missing is on the edge, stylistically; I'll call this 'stretching our sense of syntax' in order to get the alliteration I'm famous for (remember the 'nattering nabobs of negativism'? those were the days!); and I'll add "hard-wired" as a nod to my fellow language maven Steve Pinker..."
Prof. Beaver expresses uncertainty about whether Safire has read The Language Instinct or not:
NEWSFLASH: Safire reads Language Instinct
Of course, I could be wrong about this. All I have to go on is the following paragraph in Safire's latest On Language piece...
I also can't be certain that Safire has read The Language Instinct, but Google tells me that he reviewed it for the New York Times, calling it "[a] deliciously erudite, if somewhat grainy, glass of Metamucil for the legion of English speakers troubled by irregular verbs."
Lila Gleitman has also testified that Safire has read some of her work, as well as Pinker's, though her experience was apparently not a positive one (perhaps because of the ritual uncleanliness alluded to earlier in this post?):
I believe the worst nightmare for any linguist would come in these three parts:
(1) being cited by William Safire in the NY Times
(2) being cited approvingly by William Safire in the NY Times
(3) being cited approvingly as claiming the OPPOSITE OF WHAT ONE HAS CLAIMED by William Safire in the NY Times
These three nightmarish events happened to me (and my collaborators, Henry Gleitman and Elissa Newport) this week. Probably this was God evening things up for granting me a fond linguists' dream last year, i.e., being elected President of LSA.
Continuing briefly on the subject of go missing as opposed to William Safire... There are some more elaborated structures that seem to be more permissive than simple inchoative "go MODIFIER", like "go (all) MODIFIER on PRONOUN":
(link) Alexa just went all Googley on us.
(link) I asked Bernard for a few moments of his time, and he went all poetic on me.
(link) moveabletype went all 730 on me and purged everyone that had signed up for notification after may 31.
(link) they looked at their desks and scuffed their feet and went all shy on me.
Even just "all" seems to help:
(link) The Day the Universe Went All Funny.
as do some other modifiers:
(link) One of the projects, restoring a historic pier in Swansea, went somewhat pear-shaped when it was discovered the relevant planning permission hadn't been approved.
There also seem to be special cases involving colors (go white, go red, go yellow, go brown, go green) and some other semantic classes.
In fairness to Daniel Baldwin, there really is an issue about the syntactic category of the complement of verbs like seem -- "this seems disturbing" vs. "*he seems sleeping" -- and missing may fall on the wrong side of that line -- "*it seems missing". However, gerund vs. participle is not the issue, which in any case will have to wait for another day.
Update: Eric Bakovic points out to me that the Safire quote on Pinker -- "metamucil" etc. -- is actually from a review of Words and Rules. Oops. I still bet that Safire read The Language Instinct long ago -- or at least had one of his researchers summarize it for him.
Posted by Mark Liberman at June 30, 2004 06:20 AM