Merkins--(origin in metanalysis)

[Laurence Horn]

At 9:57 AM -0500 10/22/06, Cohen, Gerald Leonard wrote:
>     Maybe this was already mentioned in the thread (if so, my
>apologies), but the origin of "Merkins" would be a sort of
>metanalysis of  the singular "American" (An example of metanalysis
>is "an ekename"--literally "an also name" being incorrectly analyzed
>in popular speech to "a nekename," which gives modern "nickname").
>
>      So, someone would say "I'm American" (adjective), which would
>be reanalyzed to "I'm a Merikan" (noun).  Then, once someone can be
>"a Merikan," more than one would be "Merikans." Then by slurring
>(I'm sure there's a more precise phonological description/rule for
>this): "Merkins."
>
except that in this case, the relationship with "America" would have
likely prevented the opacity that's normally the precondition for
this kind of reanalysis (cf. the loss of initial n- in apron, adder,
orange, umpire and the acquisition of the initial n- in newt,
nickname, n and, for a while, nuncle).  I doubt there was metanalysis
in this case, although it's hard to prove either way.

(The closest equivalent would be an + other > a + nother, as
discussed here earlier, but in that case "other" doesn't really
continue to function as an adjective, so you do get *some* loss of
transparency, certainly more than in "American" given "America".)

LH

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org