Re Michael Gibbons' point about normative influences, from John Davis 
 
There surely must be different levels of normative influence.  On the 
one hand, one has reasonably well defined moral theory, such as  
utilitarianism, rights theory, Rawls, etc.  On the other hand, one has 
concepts which are more loosely value-laden, such as freedom, progress, 
class, etc.  I think it is the latter people are thinking of in the 
current context, but what may be more interesting is the degree to which 
these value-laden notions (sometimes better thought as ideological notions) 
support or bias a purportedly positive analysis of facts and causes as 
implying a particular moral theory.  That is, if we use the language of 
freedom all the time (say, to the exclusion of unmet needs), to what 
extent does that mean that our science or positive theory so-called  
only 'fits' a rights-type moral theory?  Or, if we use the language of 
preference satisfaction (the value-laden instrument), to what extent  
does it impair speaking about social welfare as including justice 
considerations?  Given that social science discourse is value-laden, 
then, how does one keep it open to multiple moral theory perspectives, 
all of which philosophers can defend more or less successfully in principle.