July 25, 2004

The thin line between error and mere variation, part 3: They started to saying

Something I'm always on the lookout for is a kind of speech error, the inadvertent syntactic blend. There are easy cases, and then there are hard ones, where I start by thinking that what I've just read or heard is a slip, but when I Google around, I discover a pile of examples suggesting that some people have grammaticalized the construction; what almost surely began life as a slip has now been installed in the grammars of some speakers as one of the available options.

Case in point: someone who reports on the radio that "economists started to saying that..." I took this to be a simple slip, blending "started to say" and "started saying". And maybe it was, in this case. But it looks like, for some people, "to V-ing" is now a third scheme of complementation for "start" (and perhaps other verbs).

As background, a few cases I think really are inadvertent slips:

(1) It really becomes down to a question of... (commentator on local elections, NPR's Morning Edition, 2/26/04) ["becomes a question of" x "comes down to a question of"]

(2) He'll have a lot of more competition. (posted to ADS-L, 7/15/04, from NPR's All Things Considered, 7/14/04: cycling expert talking about Lance Armstrong in the Tour de France) [presumably: "a lot more competition" x "a lot of competition"]

(3) ...look at the power of art has to... (posted to ADS-L, 6/25/04, from NPR's Fresh Air, 6/25/04), Alain de Botton in an interview) ["look at the power of art to..." x "look at the power art has to..."]

(By the way, these slips could use some serious study. I wouldn't want to give the impression that the blending just involves strings of words. But that's a topic for another day.)

And then, the precipitating example:

(4) ... economists started to saying that... (interviewee on PRI's Motley Fool Radio Show, 7/25/04)

I caught this because it sounded odd to me. On the other hand, it didn't sound like total Martian. Maybe the speaker slipped. But maybe he was using a construction that was in his grammar but not mine. Maybe, even, he was using the V-ing form to indicate a semantic subtlety -- perhaps the continuative sense associated with progressive V-ing forms -- that I didn't get.

So I Googled. There were about 35 relevant Google web hits for "started to saying", 5 google groups hits; about 20 web hits for "start to saying", 4 groups hits; and 5 web hits for "starts to saying", no groups hits. Not too shabby. I tried some other complement verbs: "started to going" got about 300 relevant web hits, about 130 groups hits; and "started to thinking" was similar (about 300 web, about 150 groups). Not shabby at all. It began to look like I was once again contemplating the thin line between error and mere variation (see part 1 and part 2). Look at some of the examples:

(5) Q: I broke up with my ex-boyfriend for various reasons, one of which was his drug habits as he had started to going to clubs and doing pills (he's 22). (www.channel4.com/health/microsites/ H/health/magazine/drugs/youask_ex.html)

(6) However, that wore thin quite quickly - I started to thinking of ways of integrating much deeper layers of meaning. (www.shaav.com/art/art-intro.html)

These are easily seen as continuative (and I find them relatively easy to understand). Examples involving repetition, but over short periods of time, are easy to find; these I find a bit harder to understand:

(7) Then some big dude came over and backed me into the corner. Where I immediately started to saying "I thought she was legal! I thought she was legal! (www.ubersite.com/m/28836)

(8) The walls started receding back and forth. And then the ceiling started to going down and going up. And that hard, cold cement floor started to shaking. (seattlemedium.com/News/article/ article.asp?NewsID=39654&sID=36)

And then there are pure inception examples, which strike me as very weird indeed:

(9) ...Cook campsite, after that my sleeping bag started to getting wet and i decided to move over to the YHA hostel after the 48 hours of rain. (www.bikeforums.net/archive/index.php/t-22925)

So maybe the semantics thing is just (mostly) a red herring. Maybe some people simply have three constructions -- V to V:Base, V V-ing, V to V-ing -- where most of us have only the first two. Maybe some people have a semantic distinction and others don't. Lots of things are possible. Here's a paper topic for linguistics students!

In any case, there's a lot more variation going on with the government of forms of complement verbs than most scholars of English think. For some entertaining moments, check out the verb "try" with various base-form complements: "try start", "try go", "try say", and the like. Googling will pull up a lot of junk, but also a respectable number of examples from what appear to be competent writers who are native speakers of English. Further study is called for -- like, examining the productions of these speakers over significant amounts of text, interviewing them, etc. -- but I don't think such examples can be dismissed as mere errors just because the people who produce them are in a fairly small minority.

zwicky at-sign csli period stanford period edu

Posted by Arnold Zwicky at July 25, 2004 06:34 PM