I read Mark's post on "harp back to" vs. "hark back to" this morning and then, not 15 minutes later, coincidentally read an article in which both "harks back to" and "harkens back to" are used (the former in the article's abstract, the latter in the body).
Subsequent Google searching revealed the following:
whG |
whG |
||
| harpen back to | 0 |
harken back to | 15,100 |
| harpens back to* | 9 |
harkens back to | 38,400 |
| harpened back to | 0 |
harkened back to** | 5,480 |
| harpening back to | 0 |
harkening back to** | 13,100 |
| TOTAL | 9 |
72,080 |
* All 9 hits for "harpens back to" are separate copies of the same music review: "The later songs are more bare in a way; the band sheds some of the piano and effects and harpens back to their old sound a little."
** I also found a small handful of hits each for "harkenned back to" and "harkenning back to".
This is not inconsistent with Mark's "harp on" hypothesis, but I think it offers more support for the hypothesis that the "harp back to" variants are cases of (lexicalized) place assimilation: since none of the forms with -en pit the /k/ against the /b/, alternate (assimilated) forms with /p/ are not expected and thus not found.
Incidentally, the OED lists "harken" as a variant of "hearken", of which most of the senses are more or less the same as the first three of "hark" ("give ear, listen"). None of the senses refer to the hunting-dog-call business, and none are declared to be used with "back".
[ Comments? ]
Posted by Eric Bakovic at June 28, 2004 05:18 PM