Brad's response makes me feel somewhat like a British Idealist in the  
18th century might have felt after perusing Boswell's account of Dr.  
Johnson's refutation of idealism--kicking a stone furiously, "I refute  
it thus", said Johnson. Let me say a little in response: 
 
On Sat, 2 Dec 1995, Brad De Long wrote: 
 
> Hmmm... 
>  
> Whenever I read something like Kevin Quinn's: 
>  
> >the British empiricists thought that all our knowledge 
> >received its warrant in virtue of its reference to self-interpreting 
> >sense-data. Modern neoclassical economics continues this tradition with 
> >its naivete about the "data" that sort out true from false theories and 
> >its pre-Kuhnian belief that the data are "theory-independent". 
>  
> I find myself confused. It does not seem to me to make... sense... :-) 
>  
I'm sorry to be confusing. Brad had asked for the sense in which  
Enlightenment thinkers were foundationalists. I picked the British  
empiricists as examples, since they are the thinkers most relied-upon in  
standard accounts of the methodology of economics. The "dogmatism" of  
empiricism is Quine's territory, of course. So let me use some of his  
insights. Most famously, Quine emphasized the way theory never confronts  
the data in atomic sized bits. Any particular theory can be "saved" by  
suitable reinterpretation of other bits. The hard-core of neoclassical  
economics is rational choice theory. This is itself never "tested" :  
instead, in order to be a candidate for explanation at all, by the rules  
of the neoclassical game, a proposed account must specify the  
maximization problem to be solved by the "agents". Otherwise, we are guilty 
of "ad-hoccery" and "implicit theorizing". Awareness of the dogmatic role 
of  
rational choice theory in the neoclassical project (post Quine and post  
McCloskey) might lead to humility and the willingness to appreciate and  
encourage a plurality of paradigms which do not put rational choice  
theory in their hard core. Neither logic nor "the data" compel such a 
choice. 
 
> What could a phrase like "self-interpreting sense-data" possibly mean? 
 
>  
That theory is underdetermined by the data (Quine again.) 
 
> Suppose that this year's budget reconciliation bill cuts $90 billion over 
> the next seven years from income support for poor single-family 
household; 
> that the elimination of the federal requirement that states "match" 
federal 
> funds leads states to cut back an extra $40 billion, that the typical 
> households whose income has been cut doesn't manage to replace much of 
that 
> extra income by increasing hours spent in the paid labor force; and that 
> the roughly 25% reduction over the next seven years in the income and 
> spending of the poorest twentieth of Americans has dire social 
> consequences--infant mortality up, hundreds of thousands of homeless 
> children in the streets of America's major cities, cholera and typhoid 
> making a comeback, and so forth. 
 
All of this seems to me horribly likely. 
>  
> A neoclassical economist, looking at this hypothetical scenario, would 
read 
> it as evidence that the neoconservative critique of welfare made by 
people 
> like Charles Murray was largely incorrect--that the post-WWII AFDC 
program 
> did _not_ lead the typical recipient to choose _welfare_ rather than 
> _work_, and that the post-WWII AFDC program was a transfer program with a 
> relatively small excess burden. 
 
Look at the way you've restricted yourself already. If Murray can show  
you in the data that the welfare system did substantially reduce labor  
supply, and undoing it increases it, you would applaud the "reform". A  
feminist economist would remind you that the net effect here would almost  
certainly be a reduction in labor, where the latter is understood to  
include non-wage, reproductive,"caring" labor--probably the most important  
labor any sustainable economy has to do, and a type that is in obvious  
ways inappropriately modelled with rational choice techniques. (See Sue  
Himmelweit, "The discovery of unpaid work", *Feminist Economics* 1, 
#2:1-19). 
 
> How would this hypothetical neoclassical economists' analysis be flawed 
by 
> its "naivete about the 'data'," by its "pre-Kuhnian belief that the data 
> are 'theory-independent'," or by its acceptance of "the British 
empiricists 
> thought that all our knowledge received its warrant in virtue of its 
> refrence to self-interpreting sense-data"? 
 
Showing that AFDC has a small "excess burden" is fine. Don't imagine  
though that this "empirical question" is what's being debated in the  
welfare reform debate. The Murrays and Gilders are making moral arguments  
that need to be answered if there is to be a chance of defeating  
repugnant policies. They rely on stilted conceptions of freedom and 
autonomy  
that are part of the deep appeal of neoclassical economics generally,  
conceptions it behooves us to pay attention to, contest and challenge. They 
have  
managed to convince people that the very real destruction of community  
and family they are seeing stems from the behemoth state, and not from  
the "marketization of everything" that policy has, however haltingly,  
served to palliate to an extent. Above all will this sort of work require  
that we leave our treasured fact/value distinction behind--another,  
perhaps the most important, empiricist dogma to add to Quine's list. 
 
Kevin