In the first part of this paper, I express doubts that Tarski and Carnap were guilty of some confusions about the relations between truth and meaning, attributed to them by Soames. In the second part, I consider Quine's Carrollian argument against conventionalism about logical truth, discussed only briefly and approvingly by Soames, and I explore the question whether some not obviously incorrect forms of conventionalism about logical truth, such as what I call "finitary conventionalism", are immune to Quine's argument.

One of the dominant traditions in normative ethics is characterised by the attempt to develop a comprehensive moral theory that can distinguish right from wrong in a range of cases by drawing on a philosophical account of the good. Familiar versions of consequentialism, deontology, and virtue ethics have emerged from this tradition. Yet such theories often seem to lack the resources needed to evaluate the broader contexts in which moral dilemmas arise, which may cause them to encourage moral complicity. Context-insensitive (...) complicity of this sort receives surprisingly little direct philosophical attention, despite its being a ubiquitous concern for ordinary moral agents and despite the threat it poses to this form of ethical theorising. The present paper sketches the problem more formally and canvasses some leading responses before locating its source in the implicit distinction between moral and non-moral domains at the root of much traditional normative theorising. (shrink)