In my post on the semantics of ‘ought’, I intimated that my end-relational analysis of ‘ought’ could also handle categorical uses, but tried to bracket that claim and those uses out of the conversation. Kris has invited me to explain, however, and so here I am pasting the crucial parts of my paper. The strategy is fundamentally the same as the one I have laid out in my JoE and PI papers.
In what follows, by a ‘categorical use’ of ‘ought’ I mean where it is used in giving advice to an agent but not in the form of how that agent might attain some end he or she desires or intends. An ‘instrumental use’ is one where the advice does have this form.
Suppose that my principle OUGHT(e) gives the meaning
of all normative ‘oughts’ (i.e. they all have an ‘in order that…’ operator). How then would we go about identifying which
proposition is expressed by the utterance of an unqualified ought-sentence?
First we look for a relevant end in the context. Does the conversation
preceding the utterance address the realization of some end that is a plausible candidate? By hypothesis, in
the categorical case it does not. Second, does the nature of the subject matter
suggest a plausible candidate? With artifacts and other objects possessing an
obvious function, such as cars, clocks, referees, bombs, and seeing eye dogs,
we can talk about how they ought to behave without explicitly qualifying the
‘ought’. I don’t think this is the case with categorical uses of ‘ought’.
Third, we can observe that categorical uses of ‘ought’ are
characteristically addressed to persons in the mode of practical advice; i.e. directed to others for the conversational purpose
of influencing those others’ conduct or directing them as to what to do. Does the utterance then presuppose the agent’s ends? Of course not, since by
definition categorical uses of ‘ought’ are not instrumental uses of ‘ought’.
All these possible cues fail us in
the categorical case. Presented with a categorical use of ‘ought’, therefore,
we must move to a higher level of exegesis. Since we cannot identify what the
implicit ‘in order that…’ operator is, can we identify what the speaker is
trying to communicate by uttering an unqualified ought-sentence without an
appropriate ‘in order that…’ operating in the context? The rules of
conversational etiquette dictate that the operator can be left unexplicit just
in case it need not be made explicit: i.e. just in case it is an assumed part
of the context shared by speaker and audience. Our question then becomes: what
are we as audience to make of the fact that an utterance treats an end-operator
as an assumed part of the shared context when it is not?
Speaking as if something were uncontroversially
the case when it is transparently not often has a rhetorical function of
expressing the demand that it be the case. For example, asserting ‘You will
come here!’ where the intended audience has other plans functions as a
demand. Plausibly, therefore, the categorical use of ‘ought’ functions
rhetorically as a demand that the relevant end be an assumed part of the
shared context. Furthermore, since categorical uses of ‘ought’ are made in the
mode of practical advice, directed at determining what to do, they express the
demand that the relevant ends be assumed as ends which the agent him or herself
desires or intends. Categorical use of unqualified ought-sentences would then
function in conversation rhetorically to demand that agents be motivated
towards certain ends. I claim therefore that the categoricity of the assertion of ought-sentences is a
phenomenon of the rhetorical use of the end-relational ‘ought’.
However how can a speech act demand certain
ends if it does not make clear which ends those are? First, categorical uses of
ought-sentences can have rhetorical value even if they fail to communicate to
which end they are relativized. It can be enough that a speaker communicates
this much: there is some end, such that I demand that you desire or intend that
end, and for the sake of which you ought to φ. In effect,
this communicates the demand that an agent be motivated to φ. Second, the existence of this kind of rhetorical use of ‘ought’ makes
possible a social institution whereby certain ends are socially expected of
agents: a ‘morality’. Where such an established institution exists, an audience
is able to glean from content and categorical use of an ought-sentence that it
assumes relativization to these moral ends. This, I contend, is what happens
when we make moral assertions.
On the hypothesis that the meaning of the categorical ‘ought’ is given by OUGHT(e) it therefore turns out to be plausible that categorical uses of ‘ought’
function to express demands and attitudes, just as expressivists claim.