November 08, 2007

The Google


In my seminar yesterday, I spent a little time reviewing the facts about (an)arthrousness in English proper names (recently discussed on Language Log here and here, with more to come), and the students brought up George W. Bush's reference to "the Google" last year, which was widely mocked, but not commented on here.


The event was a CNBC interview by Maria Bartiromo on October 23.  This site has a video clip, with the transcription:

HOST: I'm curious, have you ever googled anybody? Do you use Google?

BUSH: Occasionally. One of the things I've used on the Google is to pull up maps. It's very interesting to see -- I've forgot the name of the program -- but you get the satellite, and you can -- like, I kinda like to look at the ranch. It remind me of where I wanna be sometimes.

The WSJ version quotes "CNBC's unofficial transcript", which differs from the ThinkProgress transcription above in several respects: it gives Bush's "kinda" and "wanna" the standard spellings "kind of" and "want to"; it has "I forgot" instead of the non-standard "I've forgot" (I can't be entirely sure after many listenings to the passage, but I'm inclined to go with the ThinkProgress version); and it corrects "it remind me" (which seems entirely clear to me in the video) to "it reminds me".  But everybody has "the Google" (as well as the odd syntax surrounding it: "One of the things I've used on the Google is to pull up maps").

The class consensus was that Bush was analogizing to "the Internet" and "the (World Wide) Web", without realizing that Google doesn't take a determiner.

[Added later in the day: I've removed a brief attempt at an explanation for the arthrousness in "the Internet" and "the Web", since all that's important here is that they ARE arthrous.  Eric Christopherson asks about the history of arthrousness in "the Internet"; he notes that other computer networks and online services that were around in the early days of the Internet had/have anarthrous names: EFnet, DALnet, Fidonet, Usenet, AOL, Compuserve.  So he wonders if Internet once was anarthrous too.  An interesting question, which I hope someone will investigate.  And Alexis Grant reports another twist: 'my mom used to say 'on the email', as in, 'I sent you something on the email'" (so treating email like phone or fax).]

[Added 11/13/07: Mail on "the Internet" is pouring in; I will eventually post a summary.  Meanwhile, David O'Callaghan has pointed me to an Onion article "Google Launches 'The Google' For Older Adults".]

The main point I'd been making in class was that we need to distinguish several senses of DEFINITE, that is, several distinct properties that expressions can have.  In particular, we need to distinguish (as I did in the first of my recent postings) between NPs that are semantically (or pragmatically) definite (conveying uniqueness or givenness or both) from those that have the definite article in them.  There's clearly a connection -- it's not an accident that there's a custom of using the label definite in both cases -- but they are not coextensive.

First, there are NPs in English that are semanically definite, but don't have the as their determiner.  Several kinds of them: personal names, like Arnold Zwicky; other proper names, like, yes, Google, and Lake Worth, MIT, NATO, and several other types; NPs with possessive determiners, like Mary's father, and with demonstrative determiners, like this dog; and a number of others.

Second, there are English NPs with the as their determiner that aren't semantically definite.  Again, there are several types; here are two in which the NP is understood as referring to a type rather than an individual:

(1) the hospital in "She's in the hospital".  This is the American version; the British version uses a bare (anarthrous) NP -- "She's in hospital" -- to convey something like 'She's been hospitalized'.  The facts about these type-denoting location nouns, of both the arthrous and the anarthrous varieties, are quite complex, but the point here is only that there are some arthrous examples.

(2) the bus in "I came on the bus" 'I came by bus'.  Note the anarthrous variant with by rather than on.  Again, the facts about these conveyance-denoting nouns, of both the arthrous and anarthrous varities, are quite complex, but the point here is only that there are some arthrous examples.

Further wrinkle: there is some question about whether there is a specifically syntactic (rather than semantic) property of "definiteness" for NPs in English (and indeed in many other languages).  At issue is whether there is a class of NPs in English that plays some role in generalizations about the syntax of the language and is not simply identical to the class of semantically definite NPs.  But in any case, we do need to distinguish NPs with a semantic property I'll label D and NPs with the syntactic property I'll label ArtDef; ArtDef NPs are those with the as their determiner.  Most, but not all, ArtDef NPs are D NPs, and ArtDef NPs are in some sense the canonical examples of D NPs.

I've chosen somewhat artibtrary labels for these classes of expressions to discourage people from reasoning about them on the basis of the meanings and uses of the English words definite and definiteness; labels are not definitions.  It's sometimes been suggested to me that anarthrousness in proper names can be predicted, at least in part, from the extent to which the referent is a definite, or defined, entity, in the sense that it has clear boundaries.  I'll take up this idea in a later posting, but for the moment I'll just note that connecting ArtDef to this sense of definite looks like reasoning from the customary label.

Posted by Arnold Zwicky at November 8, 2007 02:31 PM