Arnold Zwicky's Blog

Ira Glass on public radio’s This American Life #454, Mr. Daisey and the Apple factory (first aired 1/06/12):

Mike Daisey. His one-man show about Apple is going back on stage this month in New York at the Public Theater. The full show has this entire other story line about Steve Jobs that you will have to buy a theater ticket if you want to hear.

See anything notable about that last sentence? Many people don’t, though there’s some tradition in the syntactic literature for treating it as problematic.

Focus on the relative clause in that sentence:

head: this entire other story line about Steve Jobs

relative: that [ you will have to buy a theater ticket
[ if you want to hear ___ ] ]

The relative has the relativizer that plus a clause (delimited by the outside brackets) with an adverbial subordinate clause, with subordinator if,  inside it (delimited by the inside brackets). And that adverbial subordinate clause has a direct object “gap” in it (indicated by the empty underlining) — a gap that’s “filled”, in semantic interpretation, by the head. In metaphorical language that’s been standard in generative grammar since Haj Ross’s 1967 Ph.D. dissertation, the filler is “extracted from” the position of the gap.

Ross explored a variety of configurations in which extraction from particular positions seemed to be barred (compact summary by John Lawler here). Extraction from an adverbial subordinate clause is one such case; it’s a violation of Ross’s Complex NP Constraint (CNPC). So the Daisey relative clause is a violation.

But there’s a huge literature on variability in how acceptable violations of the Ross constraints are, and on the possibility that the Ross constraints are primarily (or exclusively) “soft” conditions on ease of processing, rather than “hard” conditions on grammaticality. In fact, the Daisey relative clause is of a pretty frequent type of apparent violations that seem to present very little difficulty in processing: extraction of an object, in a adverbial clause with the subordinator if or whether, with a head that’s semantically indefinite. Examples are so common, and so generally unremarkable, that I haven’t bothered to collect more than a handful.

Here are a few more, with the head boldfaced and the position of the gap indicated by empty underlining:

He [Bill Clinton] ended up saying some things that who knows if he would have said ___ if he thought that he was on the record, so to speak. (Jacob Soboroff, interviewed about citizen journalism and the 2008 race, NPR’s Morning Edition Sunday 6/15/08)

Nice treatment, and one that I’d forgotten, if I ever read ___. (Larry Horn on ADS-L 11/18/10)

There are books up there that I can’t remember whether I’ve read ___ or not. (Joshua Foer, “Secrets of a Mind-Gamer”, NYT Magazine 2/20/11, p. 32)

In this configuration, a resumptive pronoun can fill the gap (“rescuing” the CNPC violation, but yielding a type of gapless relative that isn’t standard in English) —

this entire other story line about Steve Jobs that you will have to buy a theater ticket if you want to hear it

(and similarly for the other examples). But in this case the gapped relative strikes most people as about as acceptable as the gapless alternative; certainly, neither is particularly hard to process. I don’t know if an eagle-eyed editor would insist on rewording such sentences to avoid both alternatives (because they both “sound wrong”), say by breaking them in two:

The full show has this entire other story line about Steve Jobs; you will have to buy a theater ticket if you want to hear it.

(and similarly for the other examples).

This entry was posted on January 26, 2012 at 9:55 am and is filed under Relativization, Syntax. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.