Monday, July 23, 2007

Cain't Got No Time: Grammaticality Judgments at the Edge of Acceptability

After avoiding listening to her for months (I might read Rolling Stone when I have a free copy lying around, but I tend to be selective in acting on its recommendations), I was eventually put into a situation where I found myself listening to Amy Winehouse. Specifically, her song "Rehab." While the music appeased the musical ear, something flagged down the attention of my linguist's ear. See if you can hear it just after 0:36 in the YouTube link, and again at 1:23.

Unless I'm mistaken, Winehouse sings
I cain't got seventy days.
And later
I cain't got the time
All of the lyrics printed online that I can find transcribe both lines with ain't, but I hear a velar obstruent before the ain't. I wasn't aware that the "southernism" cain't was recognized as a "Americanism" across the pond. If any British English speakers have insight on this, please let me know.

Now I know those of us from Northern Indiana like to think we speak "plain" English, but the fact of the matter is my dialect has as much in common with southern US dialects as it does with more eastern dialects. So it's not the word cain't that caught my ear but its use. For those unfamiliar with the word, a quick Google search on "cain't" produced the following examples.
I'm just a girl who cain't say no.
Oh you cain't getta man with a booooook!
Poor Bill, He Cain't Help It.
Cain't you do nothing 'bout them weeds?
Make 'em an offer they cain't refuse.
I Cain't Get No Wireless.
Cain’t yew afford no gas?
It just ain't fair if you cain't cheat!
Ya cain't get thar from hee-yah!
Why cain't we get the FDA to label food made in China?

Prussian Blue: Them thare girls cain't sing.
The word means essentially the same thing as can't but perhaps more forcefully negative. Some of the results returned by the search clearly mock the word (and by extension, its users) and some results are even references to other song lyrics (Rodgers and Hammerstein's "I Cain't Say No"). Did Winehouse appropriate the word cain't to add authenticity to her fake soul vocals only to use it erroneously? I thought a few Google searches would prove that hunch.

The only search result I could find of "cain't got" was on a German LiveJournal page.
you cain't got no chance with cupid
In contrast, "cain't get" returns 715 results as of this posting, one of which was quoted above. Why the difference? A speaker of more prestigious dialects might point out that got is not an infinitive whereas get is, but the got here is being used as a verb denoting possession. As in "I got five weeks left." In at least my dialect, some negative auxiliary verbs can appear before this verb: don't got, ain't got, haven't got. Others cannot: *can't got, *won't got. Could there be a statistical effect intervening, flagging the construction as ungrammatical when what it means is "I haven't heard this before"?

I gathered some quick google stats on the likelihood of cain't appearing with verbs of possession in comparison to other negative auxiliaries.

don't can't ain't cain't didn't couldn't wouldn't Totals
got 428,000 13,000 2,240,000 1 96,700 687 630 2,779,018
% 15.40 0.47 80.60 0.00 3.48 0.02 0.02
get 84,100,000 46,700,000 28,800 715 17,500,000 6,420,000 2,250,000 156,999,515
% 53.57 29.75 0.02 0.00 11.15 4.09 1.43
have 340,000,000 2,930,000 49,600 45,900 44,300,000 3,070,000 17,100,000 407,495,500
% 83.44 0.72 0.01 0.01 10.87 0.75 4.20
own 2,130,000 142,000 523,000 5 444,000 25,300 47,000 3,311,305
% 64.33 4.29 15.79 0.00 13.41 0.76 1.42
hold 2,370,000 2,020,000 1,990,000 4,910 727,000 1,120,000 433,000 8,664,910
% 27.35 23.31 22.97 0.06 8.39 12.93 12.93
buy 7,060,000 2,160,000 2,870,000 3,930 2,070,000 443,000 1,050,000 15,656,930
% 45.09 13.80 18.33 0.03 13.22 2.83 6.71
find 2,170,000 51,000,000 6,340,000 13,400 2,580,000 4,060,000 547,000 66,710,400
% 3.25 76.45 9.50 0.02 3.87 6.09 0.82
Avg % 41.78 21.25 21.03 0.02 9.20 3.92 3.93

Based on these numbers, there appears to be a dis-preference for cain't got, cain't get, and cain't own. Whether that difference is significant remains to be determined with better data. Only the one of these strikes me as sounding ungrammatical. This raises interesting questions for the ways in which statistical feedback informs grammaticality judgments. I doubt the differences of the chart can be accounted for solely by semantics.

One interesting result is that ain't got appears much more frequently than would be predicted from the other patterns. This might be a matter of informal style favoring aint' over don't. The possessive meaning of got is less likely in more formal styles. Another factor could be that ain't got is preferred over ain't have. Perhaps for formality reasons.

1 comment:

toothychum said...

Thank you! Someone else noticed! I distinctly heard the same thing.