I would appreciate a quick explanation:  
What is the "theory of choice" that some on this list want economics to be?  
  
To provide some (presumptuous) context for my question:  
  
I presume that an economic "theory" must have empirical content, and that those proposing
that economics is a "theory of choice" are indeed proposing that as a "theory of choice"
it has empirical content.  That is, they are not simply asserting that we will not
understand past human behavior unless we tell stories rooted in human agency.
  
So naturally I am interested in knowing a prediction that emerges from an economic model
that is due to this 'theory of choice' independently of the additional structure (e.g.,
constraints) within the specific model chosen to represent that theory.  I presume that
anyone who wishes to define economics as a "theory of choice" stands ready to offer many
such examples.
  
- I presume preference stability is an important structure and that anyone who want to
root behavior in choice and choice in preference will want to discuss both the stability
and origin of preferences
- I presume Gary Becker's 1962 JPE article is relevant, despite its inverted conclusions.
(I.e., he really shows that the study of constraints adds more to predictive content than
any story about choices, at least in his settings.)
- I presume Alan Kirman's 1989 EJ article is relevant   
I notice Michael slyly changed 'theory' to 'study' and added constraints, to get a more
reasonable definition of much microeconomics.  Perhaps he still gives away too much by
conceding without discussion that we can study choice without any direct mention of
behavior, but exploring that seems to go a bit astray at the moment.
  
My guess is that those proposing that economics is a "theory of choice" are really
proposing something simple and not particular to economics: that if we combine a folk
understanding of human psychology (including very metaphysical beliefs about agency and
deep presumptions about preferences) with enough concrete particulars about the situations
in which humans are embedded then we get better at making guesses about behavioral
outcomes.  So for example if we learn of an institution where some nonproductive activity
is wealth generating, we expect to find individuals engaged in this activity and to find
that these individuals did not have obvious ways to produce more wealth with equal effort.
Of course knowledge of the legality of the activity, the likelihood of being observed, the
status of the activity, the local ethical analysis of the activity, the depth of ethical
commitment in the population, and many other factors will also help us to guess the
outcomes.
  
But that is just my guess based on what I have seen so far in this thread.  
  
Thanks for any help,  
Alan Isaac  
  
PS I hope my references to the literature will move this discussion a tiny bit in the
direction suggested by Tony.