John C. Medaille has a long quotation, I believe from Charles Clark.  
Charles is a colleague whom I respect immensely, and I suspect the  
quote is longer than what is in the post. In any case, the thought at  
the end comes from Schumpeter  
  
"This _vision_ is pre-analytical in the sense that it  
exists before theoretical activity takes place.[Charles M. A. Clark]"  
  
Schumpeter wrote: "Now it should be perfectly clear that there is a  
wide gate for ideology to enter into this process. In fact, it enters  
on the very ground floor, into the preanalytic cognitive act of which  
we have been speaking.  Analytic work begins with material provided  
by our vision of things, and this vision is ideological almost by  
definition.” (History of Economic Analysis, p.42)  
  
But I take issue with what follows as he shows his absolute faith  
in the scientific method: "But we also observe that the rules of  
procedure that we apply in our analytic work are almost as much  
exempt from ideological influences as vision is subject to  
it....[these rules] tend to crush out ideologically conditioned error  
from the visions from which we start." (ibid. p. 43)  
  
  
And the trouble with relying solely on pure logic (as was suggested   
in one of the posts),  is that it involves reasoning from  
assumptions, and they can be value-laden, even in statistical  
studies, as some feminist economic research has pointed out. In this  
respect, there is a very wide gap between the natural and social  
sciences.  
  
Sumitra Shah