In 'A Theory of Human Action' (1970) Alvin Goldman launched an attack on what has become known as the Anscombe-Davidson Identity Thesis. In brief, this is the thesis that our acts are our body movements, and that all the different effects of that movement do not entail that different acts have been performed, but only that an identical act has different descriptions. In her response to Goldman, Anscombe (1981) claims that Goldman is arguing at cross-purposes. I will argue that this (...) is partially true, but only because she accepts what I shall call the Symmetry Thesis. This thesis in turn is natural if you accept a third thesis called the Irreducibility Thesis. These three theses form a consistent set that the defenders of the Identity Thesis would accept and its detractors would deny. It follows that the best way to attack the Identity Thesis is to attack the Irreducibility Thesis on which it is based. I believe that this can be done and that Goldman is basically right, but I will not be concerned with a full defense of Goldman's method of individuating actions against its competitors. Rather, I want to show the grounds, some of them extralogical, upon which preference can be made. (shrink)