1.In a recent paper Carl Sachs provides a helpful discussion of Richard Rorty's thinking concerning the question of naturalism (Sachs 2009). He distinguishes between two positions that he terms 'non-reductive physicalism' and 'pragmatic naturalism', claiming that Rorty has moved, under pressure from critical remarks by Bjorn Ramberg (2000), from an endorsement of the first to an endorsement of the second. Moreover, the latter is the more cogent position, according to Sachs. The central difference between non-reductive physicalism and pragmatic naturalism is that the latter but not the former sees what Sachs, following Ramberg, calls the vocabulary of agency as having a privileged, transcendental role in relation to other vocabularies, in particular, all descriptive vocabularies, including those of the sciences. In this paper, I will be concerned to raise some doubts about pragmatic naturalism and the reasons Sachs gives for preferring it to non-reductive physicalism; and, further, about whether Rorty or more generally a pragmatist who also wants to be a naturalist really should or needs to subscribe to pragmatic naturalism rather than non-reductive physicalism.Both non-reductive physicalism (NRP) and pragmatic naturalism (PN) reject a reductive naturalist position, according to which a certain scientific vocabulary - often taken to be that of completed physics - provides us with a full, basic account of what is true, or what the 'facts' are. All other vocabularies have to be 'made sense' of in terms of this basis, i.e. they have reduce to it (at least in principle), or else be seen as in some way or other making less than literally true claims. This is something that Rorty has steadfastly opposed through different avenues of thought. One central one (developed in Nonreductive Naturalism in Rorty 1991) has been the Davidsonean line that while there may be just one reality 'out there' consisting of physical particles in motion (or whatever the physicists tell us) there is an indefinite number of mutually incommensurable ways of describing that reality, none of which can non-question-beggingly lay claim to providing the one true description of it. The position is (semantically) non-reductive but also, apparently, naturalistic in that it allows all events can be described in microphysical terms. We might call it 'ontological physicalism'. Rorty's naturalism goes however further than this ontological thesis in that he also claims there are no radical discontinuities between non-human and human cognitive activity, and that Darwin's theory of evolution by natural selection has provided us with a framework for understanding our place in nature alongside that of other organisms (see Inquiry as Recontextualization: An Anti-dualist Account of Interpretation in Rorty 1991). Sachs calls this the continuity thesis. Both NRP and PN seek to uphold both these features of Rorty's naturalism.NRP can be said to stop there, but PN adds a further twist. Whilst both NPN and PN are steadfastly anti-reductive about different vocabularies, they diverge on the further question of whether there is anything especially special about that mode of understanding and explaining of ours which we apply to each others' activities, involving the attribution of beliefs, desires and verbal meanings to explain action: the vocabulary of agency (Ramberg 2000). The vocabulary of agency can be equated with a certain conception of what others have called folk or common sense psychology. For Davidson, common sense psychological explanations answer to constitutive ideals that are quite different from those of natural science, rendering any hope of reduction of the former to the latter forlorn: this is the thesis of the anomalism of the mental (see Davidson's 'Mental events', and other essays collected in his 1980). Initially this thesis operated with a notion of scientific explanation that involved strict laws, but in later work Davidson stressed that psychological explanations are distinctive compared to those of any natural science, 'exact' or otherwise, implying that there is a more significant fault-line between the vocabulary of agency and all scientific vocabularies, on the one hand, than between any two vocabularies within the latter class, on the other (Davidson 1987). …