The funny thing about Hume is that he did believe that there was 
an algorithm from a pure core of bare particulars, but he could find 
no justification to warrent this algorithm -- but felt that one 
was demanded.  A look at the work of Edelman, Kuhn, Popper, (and sorry 
to use the word), gives us some reason to think that Hume was mistake about 
how to think of learning, and wrong about a pure coure of bare particulars 
in experience.  Wrong also to think that learning was an algorithm from 
this pure core, one which explains habit formation. 
 
And it is also correct that at least Hume found grounds for his account 
of learning and the existence of a neutral foundation of bare experience 
in the remarkable fact of communication -- the fact that people even of 
very different background can communicate, and do so in terms of the same 
particulars, the same components for conceiving things.  Nearly identical 
arguments can still be found in Quine. 
 
It is here were the work of Wittgenstein (and again Hayek and Edelman) 
does the most to undermine the justificationalist project in language and 
metaphysics (and Kuhn here too).  What these folks have done is to provide 
what has been called 'a reversal of metaphysics', turning upside-down 
what needs to be explained, and finding the ground for 'informational 
particulars' in a shared structure of goings on together, and several 
levels 
of selective processes at both the neurological and the social level. 
 
On Wittgenstein and the 'reversal of metaphysics' see Erich Rech, "Frege's 
Influence on Wittgenstein:  REversing Metaphysics via the Context 
Principle" 
(forthcoming in a festschrift for Leonard Linsky).  See also Richard 
McDonough, "A Culturalist Account of Folk Psychology" in John Greenword, 
ed. 
_The Future of Folk Psychology_. 
 
A summary of Edelman's work can be found in _Bright Air, Brilliant Fire_. 
 
For Hayek, see esp. his essay, "Rules, Perception, and Intelligibility", in 
_New STudies_, or his _The Sensory Order_ (book with that o D.O Hebb with 
marks the modern era in cognitive/neuro-science, introducing the Hayek/ 
Hebb synapse and articulating the many-many problem of classification -- 
see Kuhn on the importance of this for overturning the Cartisian project.) 
 
Greg Ransom 
Dept. of Philosophy 
UC-Riverside