[Ads-l] Pronunciation of "(anti)semitic"

[Charles C Rice]

To fill out what Lawson wrote, I checked the Spoken section of COCA. It is collected using closed captions on news network discussion shows, so written in standard orthography (mostly). Token frequency pushes phonetic cohort effects rather than type frequency, and I figured this would be the closest accessible spoken corpus that would give a reasonable count.
*etic	8381 tokens or 66.44 per million
*itic	3183 tokens or 25.23 per million
So the former is 2.63 times more frequent than the latter. Just to be careful I added in searches using the <d>, but the ratio was very close:
*edic	375 (almost all medic, and after subtracting comedic and orthopedic)
*idic	123 (almost all represented by acidic and Hasidic)
Again it is about 3 to 1, so no effective change. That doesn't look promising for the necessary effect. 

One outside possibility is to add the <m>, so effectively capturing the last two syllables. The only result in *mitic/*midic is the target word semitic and its relatives. In *metic/*medic there is medic and cosmetic (with their relatives) and the adjective forms of arithmetic, all told over 1000 tokens. And the two pronunciations of arithmetic are also suggestive. There are no words I could see in the results that varied only by the penultimate vowel quality of [E] and [I]. This word arithmetic is the closest, even though in my speech the noun has a schwa and the adjective has the E. Maybe the E in "anti-semitic" arises as a form of the N/Adj ablaut? That's pretty far-fetched.

James Eric Lawson wrote:

No. The influence of the pin/pen (or rather mitt/met, however you want to canonize it, supposing that merger exists at all) merger is stronger and more likely than the influence of adjectives in -etic (other than a hopeful pathetic). The -ular adjective influence on nucular I'm more ready to accept. Roughly, there are more adjectives in -etic than -itic (400/300), which of course says nothing about the relative frequency in contemporary speech. The -etic of genetic, for example, might influence antisemetic use, but making a case for a strong influence seems difficult.

The rough count of -ular vs. -lear is 400/58, and the relative frequency of what I hope are representative examples seems to heavily favor -ular over -lear, completely unlike -etic/-itic.

--
James Eric Lawson

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

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