Eaker-wcpa-2006 - U of L Personal Web Sites

advertisement
Comments on Erin Eaker’s
“Are belief reports ambiguous between de re and de dicto readings?”
Peter Alward
WCPA Annual Meetings
October 14, 2006
Eaker argues that there is no genuine ambiguity to be found between de re and de dicto
readings or interpretations of belief sentences. She considers two ways characterizing the
distinction:
1. Psychological characterization
(a) De re belief sentences attribute de re belief to subjects
(b) De dicto belief sentences attribute de dicto belief to subjects
2. Truth-conditional characterization
(a) The preservation of subjects’ “ways of thinking” of objects is not
required for the truth of de re belief sentences
(b) The preservation of subjects’ “ways of thinking” of objects is required
for the truth of de dicto belief sentences
And she suggests either way, the distinction is usually taken to be encoded in linguistic
theory by means of the notion of scope:
3. Scope encoding
(a) In de re belief sentences, expressions referring to objects of belief have
wide scope
(b) In de dicto belief sentences, expressions referring to objects of belief
have narrow scope
Eaker criticizes both characterizations of the ambiguity, as well as the claim that it can be
understood as a scope ambiguity. First, Eaker’s presents the following argument against
the psychological characterization:
(i) even if the distinction between de re and de dicto belief can be drawn, this
distinction does not map on to the putative distinction between de re and de dicto
belief sentences: de re belief sentences can be used to report de dicto beliefs, e.g.
Second, she criticizes the truth-conditional characterization by undercutting its generality:
(ii) certain belief sentences – e.g., those using devices of direct reference to pick
out objects of belief – lack de dicto readings
And third, she presents a number of considerations which impede the identification of the
putative de re/ de dicto ambiguity with any kind of scope ambiguity:
(iii) the putatively de re readings of certain belief sentences – those containing
proper names and in which co-extensive predicates are substitutable, e.g., –
cannot be represented by taking the relevant expressions to have wide scope
(iv) there are scope distinctions that cannot be adequately characterized by the
putative de re/ de dicto distinction.
Although I endorse a version of the de re/de dicto ambiguity, there is much of Eaker’s
argument with which I do not take issue. First, I am willing to concede the argument
against the psychological characterization of the ambiguity. Even if there is an interesting
notion of de re belief – and I am neutral on this question – it is completely orthogonal to
the semantic ambiguity at issue. And second, although I am not entirely persuaded by
Eaker’s arguments against treating the de re/ de dicto ambiguity as a scope ambiguity, I
am not committed to treating them this way in the first place. Instead I am happy to take
‘believes’ to be a two-place predicate satisfied by subject-proposition pairs in both de re
and de dicto belief sentences. The semantic difference between them, on the picture I
have in mind, is that in de re sentences the propositions in question are Rusellian singular
propositions and in de dicto sentences they are general or Fregean propositions.1, 2
I do, however, take issue with Eaker’s objections to the truth-conditional characterization
of the ambiguity. Eaker seems to concede that that belief sentences which use proper
names or indexicals to pick out objects of belief have de re readings (I take it she would
say they have only de re readings). What I want to argue is that they have de dicto
readings as well. And my argument stems from the role they play in explanations of
behaviour. Suppose, for example, that as the Watergate scandal was unfolding, Nixon
sent some thugs to break into the offices of Woodward and Bernstein. Nixon’s
(hypothetical) behaviour might be explained as follows:
(a) Nixon desired to protect his presidency
(b1) Nixon believed that Deep Throat was a danger to his presidency
(c) Nixon believed that in order to protect his presidency from people dangerous
to it, he needed to discern their identities.
(d) Nixon believed that the identity of Deep Throat could be discerned by
breaking into the offices of Woodward and Bernstein
What is important to note is that if (b1) is given a de re reading here, we no longer have
an explanation of Nixon’s behaviour. Nixon stood in two distinct complex cognitive
relations to the single person, Mark Felt. Let’s call them the “Felt-relation” and the “Deep
Throat-relation.” Nixon engaged in the pattern of behaviour that he did only because he
1
Note: this does not presuppose there are two types of belief; rather there are two ways of describing single
types of complex doxastic state.
2
In order to provide a uniform treatment of de re belief sentences involving both definite descriptions and
devices of direct reference, one would have to suppose that in the former case descriptions have narrow
scope but can nevertheless replaced by co-referential (or co-denotative) expressions salva veritate. This
involves, in effect, taking there to be a semantically significant referential/ attributive distinction within the
scope of epistemic sentence operators, if not more generally.
believed Felt to be a danger while thinking about him by means of the Deep Throat
relation. If Nixon had instead believed Felt to be a danger while thinking about him by
means of the Felt-relation, he would have ousted him from his position as Associate FBI
Director, or had him assassinated, rather than breaking into the offices of a couple of
Washington Post reporters.3 If (b1) is a de re belief sentence, it makes no reference to
either cognitive relation and so could not serve to explain either pattern of behaviour. But
since it does explain Nixon’s behaviour, it must be given a de dicto interpretation.
Moreover, the same case can be suppose vis-à-vis belief sentences containing indexicals.
Suppose Woodward and Bernstein asked Felt why Nixon’s cronies broke into their
offices while talking to him from outside the bathroom stall in which he was hidden.
(Assume for illustrative purposes that they did not realize that Deep Throat was Mark
Felt). Felt could adequately answer this query by giving the same explanation with (b1)
replaced by
(b2) Nixon believed that I was a danger to his presidency.
But again, if (b2) were given a de re interpretation, it would make no reference to the
Deep Throat-relation and so could not explain Nixon’s behaviour. But it does. Exactly
how names and indexicals are able to latch on to cognitive relations between believing
subjects and the objects of their beliefs – or how such relations come to be truthmakers
for belief sentences containing devices of direct reference – is a complicated question.
Nevertheless, they must do so in order for speakers to be able to use them as they do.
Note: if you are worried that ‘Deep Throat’ is not really a proper name, you can further fictionalize the
case by supposing that Felt called himself ‘Fred’ rather than ‘Deep Throat’ when acting as Woodward and
Bernstein’s source.
3
On the received picture, we can distinguish between three levels of meaning or content:
character, semantic content, and pragmatic content. Character or linguistic meaning is a
property of sentence types and can be viewed as a function from contexts of utterance to
semantic content. Semantic content is a property of sentence tokens or utterances and
corresponds to what is literally expressed by a sentence on an occasion of use. And
pragmatic content corresponds to what a speaker pragmatically imparts or conveys by
means of what she says. If there is a genuine de re/ de dicto ambiguity, then the character
of a belief sentence yields a de re semantic content in some contexts of utterance and a de
dicto content in other contexts. And if there is a no ambiguity then the character yields
the same content in every context, except insofar as the sentence contains indexicals or
other context-sensitive elements.
Now it remains open to the univocalist to explain the de dicto readings of belief sentences
I identified above by appeal to de dicto pragmatic contents those sentences have in
various contexts of use, rather than de dicto semantic contents. There is, however, reason
to balk here. It entails that when we engage in psychological explanation – or at least
when we do so using devices of direct reference – we operate at the level of metaphor.
When one explain someone’s behaviour one does not tell one’s audience why the subject
did what she did. Instead one (cagily) makes implications and let the audience figure it
out.
The univocalist might, of course, move to Cappelen and Lepore’s more fine-grained
taxonomy and distinguish between semantic content – what is literally expressed by a
sentence – and illocutionary content – what is said, or asked, or requested, etc., by means
of it. And I would be happy to concede a univocalist account of semantic content as long
as my opponent grants me a de re/ de dicto ambiguity at the level of illocutionary
content.
Download