In my first post in this series, I wrote about how desirism handles an issue that Peter Railton brought up in his article, Moral Realism, (The Philosophical Review, Vol. 95, No. 2 (Apr., 1986), pp 163-207.) This being the assertion that two people can be in agreement on all of the facts of the matter and still make different value judgments.
I handled this by saying that two people who agree on all of the facts of the matter will agree about the relationships that exists between objects of evaluation of desires, yet can still be expected to have different desires. This is true in the same way that two people who agree on all of the facts of the matter will know the locations of things, but will not necessarily share the same location.
Railton reports this fact by saying"
This claim is defended in part by appeal to the instrumental (hypothetical) character of reason, which prevents reason from dictating ultimate values.This should be taken as another way of saying that knowledge of facts does not change an agent's desires any more than it would change an agent's age or height.
Here, we do have to remember the distinction between desires-as-ends and desires-as-means. Because desires-as-means contain beliefs about how the means relate to ends, knowledge of facts are relevant to desires-as-means.
However, in this discussion, we are talking about desires-as-ends, or what Railton calls "ultimate desires". This term, "ultimate values" has an important ambiguity in that it does not actually distinguish between an agent's desires-as-ends as I use the term and "value judgments" (or beliefs about relationships between states of affairs and desires). This ambiguity creates a risk of confusion later on.
This is said to generate a problem in that a debate on matters of morality would seem to require starting with some agreement about ultimate values. Since ultimate values cannot be determined by reason, then agreement on matters of morality cannot be determined by reason.
Again, I want to stress the distinction between value-judgments (statements about relationships between objects of evaluation and desires) and an agent's desires-as-ends. The objection above requires equivocating between these two concepts.
Railton's first response to this, however, brings up the fact that even the most objective of empirical claims depends to a substantial degree on shared values.
The long-running debate over inductive logic well illustrates that rational choice among competing hypotheses requires much richer and more controversial criteria of theory choice than can be squeezed from instrumental reason alone. Unfortunately for the contrast Ayer wished to make, we find that argument is possible on scientific questions only if some system of values is presupposed.Indeed, scientists do share a common set of values. The value of science itself and its methods is one that cannot be defended by a laboratory experiment without, at best, begging the question regarding the value of conclusions drawn from laboratory experiments. Given the fact that even science is not as value-free as many believe, the claim that debates on morality also requires shared values seems less significant.
I think that it is still important to determine whether the shared values necessary for debates on morality to be the same shared values required by scientific debates, or if morality requires its own unique shared values. Reducing morality to a science would seem to require the former, and that has not, as of yet, been established.
As it turns out, agreement on matters of the form, "Agent has a desire that P" and "P is true in S" are, in fact, matters where agreement requires the same values shared by science. Where value-claims are taken to be claims about these types of relationships, then the only values that the agents need to agree upon to reach the same moral conclusions are the values that guide scientific research.