Professional Documents
Culture Documents
EGU
Reykjavík
24 August 2021
Downloaded from https://academic.oup.com/book/43848/chapter/370810763 by University College Dublin user on 15 September 2022
. . . and the boon twins Art and Con, aged thirty-seven years, who
measured in height when in their stockinged feet three feet and
four inches and who weighed in weight when stripped to the buff
seventy-one pounds all bone and sinew and between whom the
resemblance was so marked in every way that even those (and
they were many) who knew and loved them most would call Art
Con when they meant Art, and Con Art when they meant Con,
at least as often as, if not more often than, they called Art Art
when they meant Art, and Con Con when they meant Con.
SAMUEL BECKETT
Downloaded from https://academic.oup.com/book/43848/chapter/370810763 by University College Dublin user on 15 September 2022
Introduction
intentions are only one part of the speaker’s communicati e intention. The act
of referring is merely one part of a speech act intended to produce an attitude
in an addressee, whose content involves the referent. The mature capacity
necessary to represent the belief that the one thing is identical to the other
thing. Impossible, because if they have two distinct concepts they are not
really confused.
profoundly implicit in the sense that the thinker lacks the very capacity to
perform some types of self-representational acts. This is paradoxical because
we still, I argue, want to ascribe attitudes involving such concepts to the
Edenic Intentionalism
At this stage, the ground has been cleared for an account of the relationship
between confusion and reference, resulting in the so-called edenic theory.
8 ntro ct on
Talking About: An Intentionalist Theory of Reference. Elmar Unnsteinsson, Oxford University Press.
© Elmar Unnsteinsson 2022. DOI: 10.1093/oso/9780192865137.003.0002
. ap ea o t ent t 11
A Pu le a out Identit
What exactly does Frege’s first puzzle have to do with identity or identity
statements? Interestingly, according to what Joan Weiner (1997: 269) calls
‘the standard interpretation,’ there is not supposed to be any connection at
all between the two. Michael Dummett held, in his ‘What is a Theory of
Meaning?’, first published in 1975, that the puzzle can be formulated in terms
of any ‘atomic statement’ (Dummett 1993: 24, 86). Richard Mendelsohn
(1982: 281–282, 2005: 30) presents a more detailed argument for this claim,
which has been repeated or endorsed by many philosophers since. In his
book on the puzzle, Nathan Salmon (1986: 12) argues that the puzzle has
virtually nothing to do with identity. His thought is, roughly, that since the
puzzle can easily be stated without any appeal to identity statements it follows
that it is really about something else, namely the cognitive significance of
referring expressions as such. Recently, Peter Hanks (2015: 126–128) has
restated the argument.
The standard interpretation of the puzzle, however, is wrong. I will argue
that any statement of the puzzle makes essential reference to statements or
beliefs involving the identity relation. It makes essential reference to what
I will call the Be ef Mo e of the mental state of identity confusion in
Chapter 2. The claim is not that Frege had a theory of identity confusion.
Rather, the model identifies the nature of the mental state speakers will
be in when Frege’s puzzle—and other similar puzzles—will apply to them.
Frege himself took the puzzle to be about why true, coreferential identity
Others have argued for a similar conclusion on exegetical grounds, see Weiner (1997) and
Beaney (1996: ch. 6).
12 att t e ascr pt ons an fre e’s c rse
Obviously, I believe that there are good reasons to develop such a theory, the only point
here is that Frege’s puzzle doesn’t provide one in the way most have assumed.
My formulation here is intended to be neutral on the question of the proper bearers of
semantic difference. The reader is free to think of (1) and (2) as sentence types, dated utterances,
propositions, statements, or something else.
. ap ea o t ent t 13
It seems, however, that (1) and (2) attribute the same property to the same
individual. Both are true if and only if that individual x—the one named both
‘Tully’ and ‘Cicero’—has the property of being an x such that x is an orator if x
Thus, P seems unassailable. And it is only natural, on this basis, to hold that
the puzzle arises simply by considering the different semantic properties of
coreferring singular terms like ‘Tully’ and ‘Cicero’. And, so, extant solutions
to the puzzle either deny the intuition that there is such a difference, or
they involve the claim that the semantic content of a name is not simply its
referent; perhaps it’s a Fregean sense of some sort.
Yet, P is demonstrably false. To start, note that many theorists would agree
that the puzzle as stated so far is not well motivated. We need to know, more
precisely, in what sense there appeared to be a semantic difference in the
first place. Usually, the missing link is provided by claiming that (1) and (2)—
and (3) and (4)—convey different information to someone who understands
both sentences (e.g., Fine 2007: 34). I assume here that something along
these lines is necessary for a full statement of the puzzle. Furthermore, most
theorists take the notions of ‘conveying different information’ or having
‘different cognitive effects’ to be epistemic, calling for relativization to an
epistemic agent.
So, we need to specify the minimal conditions for being an epistemic
agent S such that (1) and (2) have different cognitive effects on S even when
S understands both. A natural suggestion is that the condition will have
something to do with S’s epistemic access to the identity or distinctness
Tully. Proving the hypothesis would show that the puzzle essentially involves
identity. There are, then, two relevant epistemic states for S.
14 att t e ascr pt ons an fre e’s c rse
people are aware that John F. Kennedy = JFK. Therefore, Frege’s puzzle does
And the same would apply in the monadic case. This is because, generally,
Those who are inclined to think that the acronym and the full name are, in fact, one and
the same expression, only written differently, should be reminded that ‘scuba’ is—or was?—
also an acronym. The two expressions are certainly susceptible to Frege’s puzzle as traditionally
conceived.
. ap ea o t ent t 15
say that the antecedent and consequent must have the same truth condition
and not merely the same truth value.
that Cicero = Tully in virtue of ha ing the false belief that Cicero ≠ Tully.
Finally, S’s ignorance may consist in the fact that S lacks the true belief
standard picture, and this is what I called Frege’s Curse before, both puzzles
give rise to one question: What do singular terms contribute to semantic
content, in any context, which is different from a mere referent? And this is
cognitive significance from ‘a = a’ when ‘a’ and ‘b’ corefer? Positing senses
as the cognitive contributions of any term like ‘a’ and ‘b,’ in any context, does
indeed count as a possible solution to this problem. Even more, it counts as
a possible solution to the second puzzle as well, as Frege himself argued. But
it is important to realize, in light of the argument so far, that the first puzzle
allows for a much more limited approach, which focuses on the peculiarities
of identity as such. This is where edenic intentionalism comes in.
will presuppose either that John F. Kennedy is identical to JFK or that they
are distinct. If the speaker is confused, she presupposes the latter, which is
false. If correct, this is rather interesting, because it opens the door to an
lots of small ants but only one big ant. Frida never realizes, however, and
starts calling ‘both’ of her ants Joe, finding it only slightly odd that she can
only ever see one of them at any given time. But she believes they don’t like
It’s clear, then, that this particular utterance of (8) is informative while, in a
different context, (8) could have been uttered by Frida herself without being
anything more than an expression of the trivial truth that a particular ant
is self-identical. So, we have yet another formulation of the Fregean puzzle.
Can senses or modes of presentation save us? No, they can’t. The mode of
presentation associated with the first occurrence of ‘Joe’ in (8) is the same as
the one associated with its second occurrence. And this is something Frida
has always known. Notice that her dad’s gestures are not essential to the
example, so the introduction of ‘demonstrative’ senses, perhaps presupposed
by the speaker, would not help (e.g., García-Carpintero 2000, Perry 2012; see
Unnsteinsson 2019a and Chapter 2 of this book for more discussion). In this
case, the name ‘Joe’ may be used to designate either one of two ants. The two
ants just happened to be one. Thus, it seems, (8) can be informative even
when both the Fregean and the Russellian propositions expressed are trivial.
This would appear to be a counterexample to what Aidan Gray (2016: 346)
has called ‘minimal descriptivism.’
The edenic theory, I argue, generally does a better job when it comes to
explaining triviality. Some relations are necessarily reflexive and others are
necessarily irreflexive. The clearest examples are identity and distinctness,
respectively. Identity is a relation that necessarily holds between an object
and itself, and nothing else. Distinctness (or non-identity) is a relation that
holds necessarily between an object and every other object, and not itself.
A speaker who utters a sentence containing two names, ‘a’ and ‘b,’ and is a
22 att t e ascr pt ons an fre e’s c rse
competent user of those names, will either believe that they corefer or that
they disrefer. If the belief is false, the speaker qualifies as confused on the
Be ef Mo e . But if the belief is true, they will say something trivial if they
Attitude Ascriptions
and Kroll 2020). But I am convinced that a proper treatment calls for a
more significant division of labor than theorists seem inclined to accept.
At a minimum, singular thought either indicates a representational state
A P
The theory of representational states like belief and desire is in principle
separate from the theory of the representational properties of various rep-
resentational acts, including mental acts like judging and speech acts like
expressing.
They might add that I just cannot admit the truth to them or myself. Now,
taken in the act-sense (11) would express something false. If my assumptions
about self-deception are accepted, I do indeed believe that my daughter
is doing drugs and there is then a sense in which (11) is true, but I am
not disposed to perform any act in which I represent myself to myself as
so believing. Ordinary language sometimes seems to break down when we
try to describe this type of situation. For example, sometimes when people
26 att t e ascr pt ons an fre e’s c rse
discuss self-deception they say things like (11) and add ‘but he doesn’t know
that he knows,’ and so on.
But if we take (11) in the state-sense, it is true, just as (10) is true if it is only
uestioning Attitudes
Bill and Biff are a pair of identical teddy bears to be introduced in the next
chapter. Strictly speaking, both (12) and (13) are only true in the state-
sense, because neither Lois nor my daughter are supposed to have access
Suppose, now, that she says to someone ‘I wish Clark were here,’ and the
response is, ‘Don’t you also wish Superman were here?’ She would have to be
lying if she said ‘No.’
Conclusion
This chapter proceeds as follows. First, I will describe two plausible models
of the mental state of confusion, the Be ef Mo e and the Concept
Mo e . Next I show that the most influential argument against the Be ef
o Models of Confusion
T e Be ef Mo e