The foregoing had the function of clearing the ground for the discussion of reference to
God and the status of predicates (concepts) applied to God that will constitutes the bulk
of this chapter.
First, the question of reference to God. How are we to pick out God as what
end p.227
we are thinking or talking about? By virtue of what is the statement (thought) directed to
God rather than to something (someone) else or to nothing? I address these questions
against the background of the main alternatives for understanding singular reference in
general.
Perhaps the most natural answer to the question “By virtue of what do we refer to a
particular individual?” is the descriptive one. One refers to Hillary Clinton by having in
mind a uniquely exemplified description, for example, “the junior senator from New
York” or “Bill Clinton's wife.” Note that both of these descriptions themselves contain
attempted singular references: New York, Bill Clinton, and the present time. (Prior to
Hillary's election, the junior senator from New York was Charles Schumer.) And these
descriptions are typical in that respect. It is rare to find purely qualitative properties that
are uniquely exemplified, like “the first human being to run a four-minute mile.” The
dependence of most such descriptions on other singular references has the consequence
that although descriptivist reference is not uncommon, it can hardly be supposed to
constitute a way in which reference could be instituted from scratch. This is hardly a
problem for reference to God, however, for this is one case in which we can find a
proliferation of descriptions that do not contain other singular references and that
uniquely apply to God if to anything: “the omniscient knower,” “the omnipotent agent,”
“the source of all being for everything other than itself,” “the necessarily existent being,”
and so on.
The idea that reference always, or even usually, depends on such descriptions, has been
effectively criticized by Kripke, Donnellan, and others. Kripke (1972) points out that
there are cases of successful reference to X in which the subject, S, does not have, and
does not suppose herself to have, any description that uniquely applies to X. Thus, he
suggests that many people use “Aristotle” to refer to the famous philosopher with that
name without being able to specify any identifying description other than “a famous
philosopher” or “an ancient Greek philosopher.” He also argues that even where S has a
description that he takes to fix the reference to X and succeeds in referring to X, it isn't
always by virtue of that description. These cases are divided into (1) those in which
nothing uniquely satisfies the description and (2) those in which it is something other
than X that uniquely satisfies it. Kripke illustrates (1) with Jonah, on the assumption that
none of the putatively uniquely identifying descriptions from the story succeeds in
identifying the prophet about whom the legend grew up or anyone else. He illustrates (2)
with a story about someone who succeeds in referring to the mathematical logician
Gödel, where all he knows about Gödel is that he proved the incompleteness of
arithmetic. But suppose that it was someone other than Gödel who did that. Kripke
maintains that the speaker can still be referring to Gödel even though the only uniquely
satisfied description he has available is satisfied by someone else.
Kripke's suggestion for an effective nondescriptivist way of securing reference