[Fwd: can/can't]

[Arnold Zwicky]

[this discussion is a spin-off from the TRY AND/TO thread.  it started
on 21 july when don lance objected to the negative in "I'll try and
see if I can't do that" - in which the "try and" is irrelevant to the
negation issue.  a further spin-off is now going on about the
phonetics of CAN and CAN'T; i have nothing particularly revelatory to
say about that.  but i'd like to pursue the syntactic/semantic issues
a little further.  larry horn (23 july) briefly discusses what i
take to be a conversational-implicature treatment of the negative in
"I'll see if I can't do that", referring to negatives in protases and
in modal contexts, with a reference back to jespersen.]

i'm dubious that this is merely conversational implicature, though.  i
think what we have here is an idiom (perhaps a previously undiscussed
one) of modern english. compare (1) and (2):

(1) See if you can get across the finish line.
(2) See if you can't get across the finish line.

(1) is neutral; it's meaning is, roughly: 'Determine if the following
assertion is true or not: You can get across the finish line.'
(2) has a neutral reading ('Determine if the following assertion
is true: You can't get across the finish line.'), but you have
to torture the context to bring it out.  the natural and easily
available interpretation is as a biased version of (1), with the
meaning of (1) plus the assumption that the assertion is true;
it conveys a prediction as to what's going to happen.

this 'predictive SEE' interpretation requires not only negation in the
embedded clause, but also the specific verb SEE in the clause above.
SEE in (1) can be replaced by FIND OUT, DISCOVER, or DETERMINE, but
the biased interpretation of (2) disappears under such a substitution.
either IF or WHETHER can be the clause-introducing element in (2) (as
well as (1)), but only (1) involves a true embedded alternative
question:

(1') See if you can get across the finish line, or if you can't.
(2') ??See if you can't get across the finish line, or if you can.

moreover, (2) is truly predictive, future-oriented, to the extent that
though the content of (1) can be reported in a past-tense description,
that's not possible with (2):

(1'') So you saw if you could get across the finish line.
(2'') ??So you saw if you couldn't get across the finish line.

(the force of the prediction can be relatively modest, as in (2),
or extremely strong, as in:

(3) Just see if I don't outrun you!)

finally, the embedded clause in examples like (1) allows polarity
items that appear in negative clauses, in conditionals (whether
positive or negative), and in yes-no questions (whether positive or
negative):

(4) See if you can ever get across the finish line.
    See if you can move at all.

(cf.: You can't ever get across the finish line.
       You can't move at all.
      If you can/can't ever get across the finish line, I'll cheer.
       If you can/can't move at all, I'll be astonished.
      Can/Can't you ever get across the finish line?
       Can/Can't you move at all?)

but these polarity items are not possible in examples like (2),
despite the fact that (2) has both a conditional/interrogative marker
(IF or WHETHER) and explicit negation in the embedded clause.  with
respect to polarity items, the predictive SEE construction acts like a
positive assertion:

(5) ??See if you can't ever get across the finish line.
    ??See if you can't move at all.

(cf.: *You can ever get across the finish line.
       *You can move at all.)

[the question marks, rather than asterisks, are there on some examples
only because of the possibility of literal readings.]

undoubtedly, there's a lot more to find out about predictive SEE.
these are just some quick thoughts (well, the product of some
sleepless time in the middle of last night - tortured by ADS-L!).

arnold (zwicky at csli.stanford.edu)