▻ HARMAN'S BASIC POSITION
Let's take a moral judgement :
'S ought to do A'.
On Harman's account, if I make this judgement I assume that S has motivating reasons to do A. Not just any motivating reasons, of course, but sufficient motivating reasons. This means something like :
S's motivational attitudes are such that, in his or her circumstances, A is the action supported by the best reasons, X.
This assumes agreement between myself and S that A really is the action supported by the best reasons and that these reasons are X.
Harman assumes the possibility of a moral community united in a consensus that
A is the action supported by the best reasons, X. The consensus will extend to a range of other actions, B, C, D ... , all of which the community agrees are supported by the best reasons. There is also agreement on what these best reasons are in each case.
There is a problem here. Harman specifies only a vague level of consensus. Such a community is formed only when 'each of a number of people intends to adhere
to some schedule, plan, or set of principles, intending to do this on the
understanding that the others similarly intend' (G. Harman, 'Moral Relativism Defended', Philosophical Review, 84, 1975, 4).
It is hard to see how such a vague agreement can yield anything so precise as 'the best reasons' to support any particular action. Something far more circumstantially specific is required. That's my first problem.
▻ HARMAN'S RELATIVISM
But assume, what motivates the question, that the 'schedule, plan, or set of principles' is specific to the group; and that S is a member of the group by virtue of an implicit contract (though it could be explicit). The group's morality is relative to the group. They have adopted it; and no-one outside the group has, all else equal, sufficient motivation to act on that morality. This gives us a form of moral relativism.
▻ SKEPTICISM AND SUBJECTIVISM DO NOT ENTAIL HARMAN'S RELATIVISM
Grant practical moral skepticism : there is no rational reason to be moral.
Grant moral subjectivism : moral judgements express (nothing but) attitudes.
I can't see how entailment would work. This is my second problem :
1 There is no rational reason to be moral.
2 Moral judgements express (nothing but) attitudes.
3 Moral judgements, such as S ought to do A, apply only if the
agent who is the object of them has contracted into a group every
member of which intends to adhere to some schedule, plan, or set of
principles, intending to do this on the understanding that the others
1 and 2 do not entail 3.
▻ HARMAN IS NOT A PRACTICAL MORAL SKEPTIC
There is a third and final problem. Unless I am much mistaken, Harman believes that under the contractual conditions he describes, it really is the case that 'A ought to do S'. Whatever this is, it is not practical moral skepticism - or moral subjectivism.