Caught in a radio news report this morning, this quote from Barack Obama, with the crucial bit boldfaced:
Obama said of a push for less financial regulation and lower taxes. “And why we would want to adopt something that we just tried and did not work, doesn’t make sense.”
This has a relative clause (that we just tried and did not work) in which a clause with an direct object gap (we just tried ___) is coordinated with a clause with a subject gap (___ did not work) [DO + SU]. As I noted in a Language Log posting on “Amazing conjunctions” back in 2005,
coordination of a clause with an object gap … and a clause with a subject gap … is usually judged ungrammatical, though there’s some question about what condition bars it.
In fact, a 1981 paper of Gerald Gazdar’s (“Unbounded dependencies and coordinate structure”, Linguistic Inquiry 12.155-84) treats such examples as ungrammatical and attempts to give an analysis that predicts that. But examples aren’t hard to find, in writing as well as speech; I myself seem to be given to writing relative clauses with this non-parallel structure.
Some examples, beyond those in my 2005 posting, from my files:
[speech] … something that President Bush supports but has been rejected by the House. [DO + SU] (Carl Kassell on NPR’s Morning Edition 3/28/06)
[writing] but maybe this is a construction that writing teachers have noted and corrected in student writing for decades, but never got turned into an explicit rule in the advice literature. [DO + SU] (AMZ in a posting to ADS-L 4/4/2006)
[writing] Something that both Foerster & Steadman and Kierzek get more or less right, but tends to be downplayed in later advice about the passive, is that how “important” the referent of Y is plays a role in choosing the voice for a clause. [DO + SU] (AMZ in a Language Log posting 7/22/06)
[writing] Coincidentally, I was just remarking to Chuck today that the use of the noun post to denote anything posted to a website is an example of a neologism that I’ve noticed but doesn’t bother me at all. [DO + SU] (Paul Kay in e-mail to me 3/28/07)
[writing] I seem to have mis-read Ron Butters’s posting on the subject, which I read as a j’accuse but was intended as a reductio. [DO + SU] (AMZ to ADS-L 2/7/08)
[writing] A handbook like this is going to be a useful resource… one that’s sensitive but eager readers will refer to again and again for tips, instruction and advice.[SU + PO] (comic strip Preteena 4/25/08, reported to me by Richard Sabey)
[speech; topicalization rather than relative clause] One of them I saw but got away. [DO + SU] (Doug Whitman 7/9/08, as reported by his father Neal on his blog)
[speech] … even the guys that didn’t like me and I didn’t like. [SU + DO] (Sgt. 1st class interviewed on NPR 11/8/09)
[writing] I was wearing something she’d never seen before and thought was really good-looking. [DO + SU] (AMZ in e-mail to friends 6/5/10)
Some speakers judge examples with non-parallel gaps to be unacceptable, others judge them to be of borderline acceptability, and still others find them acceptable, period. In the examples from me above, I produced the sentence first, then realized what its structure was, and on reflection decided that I was content with it as it stood.
January 28, 2012 at 5:56 am |
I think they work fine, and I would guess it’s because the notions of ‘subject’ and ‘object’ are purely determined by word order in English. As long as we can fill the gap, the lack of parallelism seems beside the point – a stylistic rather than grammatical constraint.
January 28, 2012 at 9:55 am |
“Purely determined by word order in English”? Word order is a signal (among several) of grammatical relations in English, but there are all sorts of alternatives to the canonical SU V (DO) order: fronted DO in Sushi he won’t eat, postposed SU in Into the room came the little pigs, and so on.
January 28, 2012 at 9:16 am |
A clear majority of these (7/9) are (DO,SU), a numerically equal majority have the conjunction but, and six of the nine have both properties. One senses a trend. Could this be something on its way to becoming a construction of English? I have no idea what the relevant historical facts are, nor am I adept at mining corpora for that kind of info. Have you had any thoughts on this?
January 28, 2012 at 9:47 am |
I’d noticed both properties, but was unsure of what to say about such observations on such a small (and unsystematically collected) corpus. But then I have no idea about how to do a systematic search through any corpus (much less the historical record) for this configuration.
January 29, 2012 at 10:57 am |
What I meant was “not marked by grammatical endings”.