Re John Davis' query:  I vote for opportunistic application of a 
shell discourse (neat phrasing there, John) over natural product of 
formalism. 
     It's more problematic for me, as a historian, because more than 
a few economists have discovered they can get away with a LOT in 
terms of bluffing what economic theory is, since most historians have 
very little training, and less understanding, about mainstream  
economics today.   
     So none of you have heard mainstream economists use this, eh? 
That's interesting.  (I have, but only around economic historians or 
historians.) 
     How about using pareto optimality, or pareto superiority, as 
the yardstick for social welfare?  Seems to me to conveniently 
exclude initial endowments of wealth and/or power as issues in a 
situation where they are very much issues.  And it might give a 
local optimum, but can't possibly answer questions about POTENTIAL 
welfare.   
     Perhaps some of the language I'm hearing in economic history is 
an extension of this line of reasoning.  (If no trade will make  
anyone better off, social welfare is maximized.) 
     -- Mary Schweitzer