Particularism is a view concerning the status of moral principles. Put simply, the particularist wants to say that moral practice doesn't require them. As we shall see, in Dancy's hands this claim becomes a claim about the nature of reasons.
First of all here's how Dancy models Particularism:
(P) Moral thought and judgement is not dependent on the provision of a suitable
supply of moral principles
Opposed to (P) is Generalism:
(G) Moral thought and judgement is dependent on the provision of a suitable
supply of moral principles
What exactly is a moral principle? Let's take the example
(A) 'Breaking one's promises is wrong'.
(A) specifies a feature of an action and ascribes a moral predicate to it. So (A) is roughly equivalent to 'If Φing constitutes the breaking of a promise then one ought not Φ'. Now, to say that one ought not do something is just to say that there is reason not to do it. So we need to get clearer on the nature of reasons if we are going gain an understanding of (A), (G), and (P).
Reasons are things which favour or disfavour a course of action or the adoption of an attitude. So to say that breaking one's promises is wrong is to say that the breaking of a promise is a feature of an action which disfavours one performing it.
An important distinction we should note is that between a contributory reason and an overall reason. When we are deciding whether or not to Φ the first step we take is to determine the reasons for and the reasons against Φing. After we have made such a determination we then make a decision as to where the reasons ultimately fall: either in favour or against Φing. Sometimes we might decide that there is more reason not to Φ than to Φ, other times not. The reasons which we surmise in our first step are the contributory reasons. In the second step we determine the overall reason.
Contributory reasons are not sufficient reasons - the existence of a contributory reason in favour of Φing is not enough to determine that Φing is what one ought to do; they contribute to our decision but do not determine it. This is because there might be other contributory reasons which disfavour Φing which override or defeat the contributory reasons which favour Φing. So a contributory reason holds only 'as far as it goes' which is just to say that it is not a sufficient reason. The correct specification of a contributory reason, then, is 'In so far as Φing constitutes lying, one shouldn't Φ'. The qualification 'in so far as' signals the fact that lying engenders a contributory reason.
Once we have a list of the contributory reasons for and against a further step must be taken in order to determine what we ought to do. This further step consists of weighing up the reasons for or against in order to determine where they ultimately fall. If they ultimately favour Φing then we have overall reason to Φ. If they ultimately disfavour Φing then we have overall reason not to Φ. Note that on this construal overall reasons are not different kinds of reasons over and above contributory reasons - to have an overall reason to Φ is just to have a set of contributory reasons which defeat those reasons which disfavour Φing.
According to (A) actions with a certain feature - actions which constitute the breaking of a promise - are things which we have reason not to do. For reasons which we don't need to go into the reason generated by this feature and which (A) specifies must be taken to be a contributory reasons - that is, a reason which holds only 'so far as it goes'.
I have been talking of features of actions generating reasons. Some features generate (contributory) reasons which favour, other times reasons which disfavour, and other times fail to generate reasons at all. To introduce some important terminology we might say that a feature which generates a reason which favours has positive valency, one which generates a reason which disfavours a negative valency and one which fails to generate any reason at all has nil valency.
We are now in a position to better understand the nature of a moral principle. A moral principle identifies a feature of an action with positive or negative valency. Of course, principles are supposed to do other things as well - they are supposed to guide our moral decisions, for instance. But Dancy is interested in the role principles play in specifying reasons. We will see that he wants to claim that it is because principles are supposed to specify reasons of a certain sort that they are not fit for purpose. The terminology introduced above will come in handy in the following discussion.