has made an important contribution with an argument that I would loosely paraphrase
Although moral philosophy
is highly diverse in its methods and conclusions, it almost always involves an
effort to identify concepts or words that have consistent moral significance.
For instance, when we examine a complex case of adultery, we may detect numerous
features that are morally relevant: promise-breaking, deceit, self-indulgence,
lust, pleasure, happiness, love, freedom, and self-fulfillment. We may not know
how to judge the case, since its features push us in various directions. But we
do know—or we think we know—the valence of each concept. Regardless
of our overall judgment of an adultery story, the fact that it involves a broken
promise makes it worse than it would otherwise be. The fact that it expresses
freedom or increases happiness makes it better. And so on.
This kind of
analysis has the advantage of allowing what Dancy calls "switching arguments."
We form a strong opinion about the moral polarity of a concept that arises in
well-understood cases, and then we apply (or "switch") it to new situations.
So, for example, if we admire conventional marriage because it reflects long-term
mutual commitment, then we ought to admire the same feature in gay relationships.
But what if moral concepts do not have the same valence or polarity in
each case? What if they are not always good or bad (even "all else being
equal"), but instead change their polarity depending on the context? Clearly,
this is true of some concepts. Pleasure, for example, is often a good thing, but
not if it comes from observing someone else’s pain—then the presence of pleasure
is actually bad, even if it has no impact on the sufferer. In my view, it is a
mistake to isolate "pleasure" as a general moral concept, because one
cannot tell whether it makes things better or worse, except by examining how it
works in each context.
Philosophers have always been eager to reject some
potential moral concepts as ambiguous and unreliable; but they have wanted to
retain at least a few terms as guides to judgment. Thus, for instance, Kant drops
"pleasure" and "happiness" from the moral lexicon, but "duty"
remains. It would be revolutionary to assert, as Dancy does, that "every
consideration is capable of having its practical polarity reversed by changes
in context." Dancy believes that no concepts, reasons, or values have the
same moral polarity in all circumstances. Whether a feature of an act or situation
is good or bad always depends on the context, on the way that the feature interacts
with other factors that are also present in the concrete situation. To shake our
confidence that some important moral concepts have consistent polarities, Dancy
provides many examples in which the expected moral significance of a concept is
reversed by the context. For example, truth-telling is generally good. But willingly
telling the truth to a Gestapo agent, even about some trivial matter such as the
time of day, would be regrettable. Returning a borrowed item is usually good-but
not if you learn that it was stolen, in which case it is wrong to give it back
to the thief.