Continuing the discussion of English orthographic gemination, Bill Poser observes that he sometimes finds himself writing "Jeniffer". This is not an experience that rings a bell for me, but Bill is clearly in tune with the zeitgeist, or anyhow the Jennifergeist:
f |
ff |
|
n |
869,000 |
481,000 |
nn |
15,400,000 |
120,000 |
Keith Ivey emailed to point out that "[t]wo accepted variant spellings of words borrowed from Spanish provide examples of an added geminate and a lost one: afficionado [and] guerilla." And notice that the result in each case is consistent with the orthographic pattern seen in the contingency tables for Attila, Karttunen and Jennifer: a preference for a single consonant paired in an adjacent syllable with a double one, in either order.
Qov emailed to say that "The ones I have to watch for are parallel, accelerate
and tomorrow. I don't
quite understand how the numbers in the tables prove your thesis, but Google
finds many more tommorows than tomorows and many more paralells than paralels."
Indeed, and also consider the relative paucity of "tommorrows". Here is the contingency table for tomorrow, which shows basically the same pattern that we've seen before:
| r | rr | |
| m | 67,700 |
14,300,000 |
| mm | 228,000 |
189,000 |
The case of variants for "parallel" is somewhat different, because there are apparently three different consonants involved to some extent in the confusions, and two of them are L's:
FORM |
ghits |
| paralel | 162,000 |
| paralell | 65,700 |
| parallel | 13,800,000 |
| parallell | 94,500 |
| parralel | 8,700 |
| parralell | 2,200 |
| parrallel | 31,200 |
| parrallell | 475 |
The analysis here is a bit more complicated -- maybe later, I have a grant proposal to write. I'll also see if I can find another, more accessible way to come at the explanation of the statistical analysis of contingency tables, to supplement the one I provided here
Posted by Mark Liberman at March 30, 2004 08:19 AM