"confuses X for Y"

[Arnold M. Zwicky]

On Sep 2, 2007, at 7:04 PM, Mark Mandel wrote:

> What's wrong with the analysis of "confuse X for Y" as a blend of
> "confuse X with Y" and "mistake X for Y"? I must be missing
> something here, probably the specifics of your taxonomy.

i begin to despair that i can ever be clear about this.  i'm not
denying that "confuse X for Y" is in some sense a combination of the
patterns of "confuse X with Y" and "mistake X for Y"; i'm saying that
most occurrences of "confuse X for Y" are not inadvertent slips in
which two different formulations of a thought compete with one
another in production, with the result that parts of each are
realized.  the many (a few hundred thousand) google webhits for
{"confuse * for"} don't look at all like slips; they look like things
the writers intended.

tellingly, among these hits are some misquotations of "We must not
confuse dissent with disloyalty" (Edward R. Murrow, _See It Now_
broadcast, report on Senator Joseph R. McCarthy, 3/17/54) as "We must
not confuse dissent for disloyalty" (plus a few misquotations as "We
must not confuse dissent and disloyalty").  the versions with
"confuse for" are telling because people usually misquote things so
as to fit their own grammars/lexicons.

there clearly are a lot of people out there who think that "confuse X
for Y" is just another construction of english -- quite possibly,
people  who prefer it to "confuse X with Y" and "confuse X and Y".

how could  this happen?  let's start with some basic stuff.  for the
moment, i'll put aside the coordinated variant.

we're talking about "misidentification" verbs that take two non-
subject arguments, which i'll refer to as R and W.  when an
expression x serves as R, it's conveyed that x expresses the "real"
or "right" identification of some denotatum d; when x serves as W,
it's conveyed that d is wrongly identified as x.  (hence the R and W.)

there are two relevant syntactic variables: which of R and W
functions as direct object (the other then functioning as an oblique
object), and what preposition marks the oblique object.  for
misidentication verbs in general, R usually serves as the direct
object, W as the oblique.   for one set of misidentification verbs
("confuse" and "mix up"), W is standardly marked by "with"; for
another ("take" and "mistake"), W is marked by "for".  typical (R, W)
pairs for "confuse with":

   (dissent, disloyalty), (accidentals, essentials), (correlation,
cause and effect), (jingoism, patriotism), (activity, productivity),
(assertiveness, aggression)

and for "mistake for":

   (landmine, frisbee), (ribaldry, humor), (the acolyte, the high
priest), (him, Rafael Nidal), (hemlock, parsley), (notoriety, fame),
(kindness, weakness)

now some possibly relevant facts: "confuse" and "mix up" allow
coordinate objects with a misidentification interpretation ("don't
confuse dissent and disloyalty"), while "take" and "mistake" do not;
and coordinate objects with a misidentification interpretation, like
coordinate objects in general, can occur in either order ("don't
confuse disloyalty and dissent").   i speculate that the symmetry of
coordinate objects promotes the possibility of a symmetric
interpretation for the objects of "confuse" and "mix up" in general.
and, in fact, these verbs sometimes occur with R and W reversed:
"confuse cause and effect with correlation" alongside "confuse
correlation with cause and effect", and even edward r. murrow quoted
as having cautioned against confusing "disloyalty with dissent".  as
a result, marking W with "for" is clearer than marking it with "with".

so i'm suggesting that "for" is imported from the "mistake" pattern
to mark the status of the two objects of "confuse" more clearly.

arnold

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