Try and/to

[Arnold Zwicky]

so much for the morphosyntax of TRY AND.  on to the semantics/pragmatics.

but, first, a word on historical (and synchronic) derivation.  suppose
we're faced with two variants, A and B, with very similar meaning,
and one or more of the following is true:
  (a) A is standard, B nonstandard;
  (b) A is formal, B informal/colloquial/conversational;
  (c) A is significantly more frequent in texts than B;
  (d) A has more parts (segments, syllables, morphemes, words) than B.
it is incredibly tempting to jump to the hypothesis that A is
historically prior to B, that B is in fact historically derived from
A.  (and, often, to make the further jump that B is *synchronically*
derived from A as well.)

now, we all know, at some level, that this jump in reasoning has
little to back it up.  we teach our classes, correctly, that
appalachian english is not standard american english altered in the
mouths of lazy louts, in fact not a *version* of some other current
variety of english at all - historically, or synchronically.  and so
on.

yet when faced with the quasi-serial GO V construction, even seasoned
linguists tend to start out by assuming that GO V must have its
historical source (and perhaps also its synchronic base) in the
longer, less dialectally restricted, more formal GO AND V or GO TO V
construction.  and to make similar assumptions about TRY AND - that
is, to refer TRY AND historically (and perhaps also synchronically)
to the more formal, and possibly more text-frequent, TRY TO.

as far as i can tell, using (a)-(d) in rules of thumb for
hypothesizing historical antecedents or synchronic "basic forms" is at
least as likely to lead to error as to truth; variants are quite
likely to have had separate histories.

even in case (d).  yes, abbreviation, omission of predictable material
in context, etc. are quite clearly the motivations in some historical
changes (in abbreviated informal english questions of the sort "Come
here often?", for instance), but this is not at all obviously what
went on in many other cases ("Give me that penguin!" vs. "You give me
that penguin!"; "I know pigs can't fly" vs. "I know that pigs can't
fly"; "all of the birds" vs. "all the birds").  great caution is to be
recommended.

and even when abbreviation/omission was a factor historically, that
doesn't mean that later generations of speakers continue to treat the
B variants as synchronically derived from the corresponding A
variants.  the A and B variants will share elements of structure, no
question.  but the default analytic assumption ought to be that they
are simply different constructions.

----

end of rant, and back to TRY AND/TO.  an earlier message from PAT
tries to tease out a semantic/pragmatic distinction between the
constructions.  i think this is basically correct.  in fact, the
semantic/pragmatic distinction follows the difference in syntactic
structure between the two constructions, and this fact suggests a much
more plausible historical scenario than the one in which TO is somehow
replaced by AND.

at the most superficial level, TRY TO is a verb + [marked infinitive]
complement construction, and TRY AND is a coordinate verb
construction.

we would expect then that the semantics of TRY TO would have a single
predication, of a subject's engaging in an Attempt, with the Attempt
itself involving a goal Action; Attempt is expressed by TRY, the
Action by the VP complement of TRY.  pragmatically, the Attempt is
foregrounded.

the semantics of TRY AND, however, would have a double predication, of
a subject's engaging in an Attempt, plus this same subject's engaging
in the goal Action; Attempt and Action are expressed syntactically by
a coordinate V (TRY) and VP, respectively.

this is, in fact, a brief account of what happens in ordinary
coordinate VPs: each VP is predicated separately of its subject.
"They left the room and cursed us" entails that they left the room and
that they cursed us (and suggests that the latter event followed the
former).  the first VP can certainly be the single V TRY in such
examples: "We told them that leaving the room would be incredibly
difficult.  They tried(,) and left the room."

but, let's face it, trying on its own is not something people are
inclined to engage in.  trying comes with its goal. ordinarily, TRY
AND is cataphoric; "try and X..." conveys 'try to do X and [hope to]
succeed in doing X'.  pragmatically, the focus is on the Action and
not the Attempt, so that in many circumstances TRY AND will be
understood as 'do X, and put some effort into it'.

the result of all this is that "try and X...", which can have its
straightforward reduced-coordination meaning, is also open to
reinterpretation as a construction in which the the second verb is
the contextually important one, while the first is pragmatic spice.
that's the TRY AND construction we've been discussing.

it could have arisen at any time from ordinary coordination.  even
if english hadn't had the infinitival TRY TO construction.

so, historically, there's no reason to think the TRY AND and TRY TO
constructions have much to do with one another.  or synchronically.

arnold (zwicky at csli.stanford.edu), apologizing for the extremely
  off-the-cuff semantics/pragmatics