Arnold Zwicky zwicky at STANFORD.EDU
Tue Aug 11 14:52:33 UTC 2009

On Aug 11, 2009, at 6:06 AM, James A. Landau wrote:

> Vanity Fair July 2008 article "How the Web Was Won: An Oral History
> of the Internet" page 100 column 2:
> <quote>
> In 1969, ARPA [Advanced Research Projects Agency] gave the job
> building "interface message processors" (I.M.P.'s), otherwise known
> as "node" or "packet switches"---the cruical hardware for sending
> and receiving bursts of data---to Bolt, Baranek & Newman.  In a
> congratulatory telegram to the company, Senator Edward M. Kennedy
> referred to I.M.P.'s as "interfaith" message processors.
> </quote>

certainly some kind of malapropism (perhaps based on a mishearing in
which the unfamiliar (to EMK) "interface" was heard as the more
familiar (to EMK) "interfaith"), but "interfaith message processor"
doesn't make much sense; it just has familiar parts.

here's what i said on the matter on Language Log on 8/8/07:


Ever since I got into the eggcorn business, people have been
nominating errors as eggcorns, or asking if some error is an eggcorn.
The American Dialect Society mailing list has a thread headed
"Eggcorn?" every so often, and I get lots of mail with that header.
Some of these errors are already in the eggcorn database, some are
lovely new finds, but others don't seem to me to be eggcorns, for one
reason or another.  The latest chapter began on August 3, with an
"Eggcorn?" posting from Wilson Gray asking about the following, from
MacUpdates comments:

   >If the author cleans up that one glitch, then I'll make a b-line
for his app.<

Larry Horn and I saw no semantic motivation for b-line and suggested
that it was probably as opaque for the writer as bee-line would have
been.  And I launched into yet another discussion of things that
aren't eggcorns but resemble them in some respects.  Here's a somewhat
spiffed-up version of what I said.

1.  Pails.  B-line exemplifies a fairly common error type, involving a
part X of an expression that can be parsed out but can't be easily
assigned a meaning: in [bi]-line, line is a recognizable element, but
what is [bi]?  The name of the letter B?  The verb be?  The noun bee?
The proper name Bea?  Something unique to this expression (a "cran

... If you can think of an item pronounced like X, or something close
to it, that would seem to contribute some sense to the whole
expression, then interpreting the expression as containing that item
and spelling the expression accordingly produces an eggcorn (or, of
course, gets you the right analysis and spelling).

On the other hand, if you're stumped about the identity of X -- that
is, if the larger expression seems irretrievably idiomatic to you --
you can just pick some existing item Y pronounced like X, ideally one
of the right sort of category to fit where X occurs (so, for [bi]-
line, a noun); you'll probably be biased towards picking a frequent
word, or one with a short spelling, or maybe you'll pick one at
random.  The result is a type of error i'm now calling a PAIL, after
the (very common) spelling "beyond the pail", where the baffling noun
pronounced [pel] is taken to represent the everyday noun pail; yes, it
doesn't make sense, but then idioms are like that...

i take "interfaith message processor" to be an instance of the
"pail" ("b-line" etc.) type of malapropism.  interesting, but not an


The American Dialect Society -

More information about the Ads-l mailing list