Try and/to

[Arnold Zwicky]

this is about the details of these constructions (their form and
meaning), not about whether one has been increasing at the expense of
the other, or whether they've been cycling up and down, or whatever.

first, form.  john patrick villanueva (20 july) suggests that "_try to
is a construction that acts as a fully productive verb, whereas _try
and_ is a chunk used ... only for non-negative imperatives."  no
problem with the first of these claims, but i doubt that the second
one is quite accurate, even for villanueva.  note that on 23 july,
rudy troike cites uses of TRY AND in the complement of modal verbs
(with an example from our own don lance).

the generalization here, for most american speakers at any rate, is, i
believe, the very same one at work in the "quasi-serial verb" (Go/Come
see who's at the door!) construction of modern english examined by
geoff pullum in his paper in the Ohio State University Working Papers
in Linguistics 39 (1990) (with discussion of various similar
constructions, including TRY AND and infinitival-complement
constructions, and careful bibliography for the generative
literature).  the condition is what pullum calls the "inflection
condition": (simplifying somewhat, and disregarding some fascinating
variation) the two verbs must each be in a form identical to their
base form.  [assuming that both verbs must agree in person and number
with their shared subject.]

this, of course, allows all base forms, including the imperative:
  Go see who's at the door!     Just try and stop me!
and complements of modals:
  I'll go see who's at the door.   I'll try and stop them now.
and all marked infinitives:
  I asked you to see who's at the door / to try and stop them.
as well as other infinitival complements:
  I watched the kids go see who was at the door  / try and stop her.

it also allows "present subjunctives"
  I insist that you go see who's at the door / try and stop this.
and present tense forms that are identical to the base:
  Whenever I'm asked,
    I go see what's happening outside /  I try and stop the fighting.

but it disallows past tense forms generally:
  *I went saw/see who was at the door.  *I tried and stop(ped) them.
and other present tense forms:
  Whenever she's asked.
     *she goes see(s) what's happening / *she tries and stop(s) them.
and nonfinite forms other than the base:
  *I've often gone seen/see who was at the door.
     *She's often tried and stop(ped) them.
  *Ok, I'm going see(ing) who's at the door.
     *Ok, I'm trying and stop(ping) it.

[all the TRY AND examples must be understood with the 'do action,
and put some effort into it' semantics of the special construction.
most, if not all, of the asterisked examples are perfectly fine
with simple reduced-coordination readings.]

this is not the end of the story.  in fact, the inflection condition
above actually holds only for the *first* verb in the TRY AND
construction.  [for some speakers, this is also true for the
quasi-serial construction, as pullum discusses in considerable
detail.]  the condition on the second verb is slightly different: the
second verb *is* in the base form (even if that would not agree with
the subject).  this subtlety is visible only for the verb BE, which
(unlike all other english verbs) has present-tense non-third-singular
forms distinct from the base.

agreement would predict things like:
  *I try and am quiet.  *We try and are quiet.
it would seem that these are out because they violate the inflection
condition on the second verb.  but examples like the following are
fine (in fact, parallel examples have already appeared in earlier
postings on TRY AND):
  I try and be quiet when I'm asked.
  We try and be nice to Cousin Lee, but it's a chore.
this is evidence that the second verb simply *is* in the base form,
and that second-verb agreement is suspended.

(to put this in a deliberately startling formulation: the AND of the
TRY AND construction, like infinitival TO, governs the base form of
its complement VP.  so the TRY AND construction shares this condition
with the TRY TO construction, but shares the first-verb inflection
condition with the quasi-serial verb construction.)

[pullum and i noticed these facts many years ago, and i've lectured
about them in various places but never got around to writing them
up.  i'm not aware of any place where this description is given
explicitly in print, but by now someone else might have published
it.  i'd be grateful for references.]

i have a lot to say about the semantics/pragmatics of TRY AND and
TRY TO, but i'll save that for a later message.

arnold (zwicky at csli.stanford.edu)