F1 The characteristics of the sort of moral system which I want to discuss are as follows. First of all it is a value system, a way of evaluating as good or bad in varying degrees. Secondly it has at least some values which are 'objective'. If something is alleged to be objectively good (or bad), then it is supposed to be good (or bad) for everyone. When there are objective values we can reasonably engage in argument, and we can do so with the presumption that one (or more) of the parties that differ is in some absolute sense wrong. If we are dealing with a subjective value system then it is not necessarily the case that of two people differing in their evaluation of something at least one is wrong. The third characteristic which interests me is what I shall call the coercive element of the system. It is rather difficult to settle exactly what does or does not constitute coercion, but this does not matter. I don't use that word because it precisely marks what I am looking for, it is the closest word I could think of.
F2 To further emphasise here my desire to get away from studying the meanings of words I point out here that it is of no interest to me whether or not any of these characteristics are necessary characteristics of moral systems, for that would be an enquiry about the way the work 'moral' is used. In asking the question 'is morality necessary?' I had in mind an area of study, to which I wanted, as best I could, to point myself. I had in mind also, roughly, a result, which I wanted to give some grounds for: that morality is not necessary. In attempting to avoid the linguistic stumbling blocks inherent in terms of the above characteristics, leaving open the possibility that I may later decide to investigate further as yet unnoted characteristics of what might be called moral systems.
F3 So the initial reformulation runs like this, as three questions. Is a value system necessary? Need we have an objective value system? Is it necessary for our value systems to endorse any sort of coercion, and if so, what sort of coercion, how much, and when? This reformulation eliminates one problem, the word 'morality', and introduces a fine frop of new problems, 'value system', 'objective', and 'coercion'. I shall leave these new problems for the while so that I can devote a few words to my rather suspect use of 'necessary'.
F4 In using 'necessary' it should be clear that I do not mean 'logically necessary'. It's more the sort of necessity which is involved when I ask myself whether I really need a new car, or, even vaguer, when I ask my errant neighbour whether its really necessary for him to make such a noise. Now I want to avoid having to pin down this word just as I have sought to avoid leaning heavily on the word 'morality'. This is possible because I am not looking for a yes/no answer. Questions of necessity often do not receive yes/no answers in ordinary discourse. A perfectly respectable response to a question about necessity of a certain course of action is a list of dire consequences likely to follow failure to adopt that course. Where the necessity is not as clear cut the response might be a list of both pros and cons, though in such a case the response is likely to be : 'No, its not necessary but ......, while on the other hand .....'. In this latter case we do get the 'no', the assymetry suggests plausibly that it is easier to refute a necessity that to establish it.
F5 This preliminary skirmish with 'necessity' does not, unfortunately, entirely settle the problem. I have suggested that the sort of necessity involved is more of a prudential necessity than a logical one, and we should deal with such prudential problems by looking at the 'pros and cons'. But how are we to identify these 'pros and cons'. How do we know when something is or is not 'likely to follow' from something else, and should we, as philosophers be engaged in investigating things which look rather like contingent probabilities? Isn't this the province of science?