November 17, 2007

Definite descriptions

Geoff Pullum has posted about the considerable difference in the acceptability of "singular they" in the two sentences

(1)  Do not speak to the driver or distract their attention without good cause.
(2)  Do not speak to the king or distract their attention without good cause.

(which differ only in (2) having king where (1) has driver, yet (2) is much less acceptable than (1)).  Geoff's explanation turns on a difference in the way "definite descriptions" (roughly, singular count NPs that have the determiner the and denote contextually unique individuals) pick out referents -- a difference made famous by the philosopher Keith Donnellan in a 1966 Philosophical Review paper "Reference and Definite Descriptions", under the labels ATTRIBUTIVE and REFERENTIAL: the driver in (1) is used attributively, the king in (2) referentially.  (In fact, Geoff uses the term referential in distinguishing (2) from (1).)

What's cool here is that Donnellan's distinction shows up in a fact about how English they is used.

[Digression: definite descriptions, as understood in the philosophical literature, have both the properties D (roughly, uniqueness) and ArtDef (having the determiner the) that I talked about in an earlier posting, on (an)arthrousness, so they're "definite" in two ways at the same time.  The standard examples of definite descriptions are singular count NPs, though there are other NPs with both the properties D and ArtDef, like the recipients of this year's Nobel Prize in chemistry.]

[Another digression: Donnellan was responding to Bertrand Russell's analysis of the semantics of definite descriptions and Strawson's challenge to it.  Donnellan's analysis, in turn, has been disputed and defended over the years, in a rich and complex literature.  (For a summary of this history, look here.  And see below.)]

Taken out of context, the driver can be used either attributively (picking out whoever is uniquely the driver in the context) or referentially (picking out some specific person and saying that, in the context, this person is the driver).  The driver of this bus in

(3) The driver of this bus is insane.

will probably be interpreted referentially: person x, who is driving this bus, is insane.  So we'd usually use a singular pronoun, he or she, for anaphora to the driver of this bus, since the speaker of (3) will know the sex of x.  They (or he or she) would be much less felicitous.

But the driver of this bus in (3) can have an attributive interpretation: whoever is driving this bus is insane.  Perhaps the speaker of (3) judges that only an insane person would drive the bus the way this person does.  It's even possible that the bus company hires only insane people to drive this particular bus.   In such contexts, the sex of the driver is not particularly important, and might well be unknown to the speaker -- so anaphoric they (or he or she) is entirely natural.

So much for the main theme.  Now, a coda.

While searching on "definite descriptions" for links to add to this posting, I came across a book by that name, edited by Gary Ostertag and published in 1998 by the MIT Press.  I somehow managed not to buy it when it came out -- I buy an awful lot of books -- but it looks like something that would interest me: it's a compendium of the classic philosophical literature on definite descriptions.

You're thinking that it's been less than ten years since this book was published, and it's obviously a valuable resource, so it should be possible to find copies for sale.  Well, on the MIT Press site, Definite Descriptions is OUT OF STOCK.  Otherwise, things are dire; the book's a rare and expensive item.  (Maybe everybody who bought a copy has hung onto it.)  For the PAPERBACK (!) edition, we find:

Alibris lists one used copy, at $298.66.  Barnes & Noble lists one used copy, at $285.58.  These appear to be the same copy, from Actinia Bookstores in Baltimore.

Amazon has one used copy, at $211.01, from Specialty-Book in Ohio (which I've been unable to find anything about).

Books-A-Million's hard-to-find inventory has one used copy, at $437.10 (or a mere $393.39, if you belong to their "Millionaire's Club").

Powell's, AbeBooks, and Biblio list no copies at all.

Somehow I don't think I'll be filling this gap in my library any time soon.

Posted by Arnold Zwicky at November 17, 2007 12:49 PM