Charlotte asked:

What is Frege's puzzle? Why did he reject the metalinguistic solution and change to reference and
sense? What is his second solution and does it work any better than the first?

and Alex asked:

I'm writing an undergraduate essay about Frege, which is, "Is sense a semantic property of singular
terms?" I would greatly appreciate any help on this subject as it is very difficult and I don't understand
it!! Thank you.

============

In his essay, 'On Sense and Reference' ('Uber Sinn und Bedeutung') Frege presents a puzzle about
the notion of identity. Identity statements, of the form A=B, can convey factual information about the
object designated. To take Frege's example, we now know that the sun we see in the sky is the same
objectwhenever it appears. Once, people did not know this. Or, to quote an example popular with
academic philosophers, we all know that Superman is Clark Kent, but Louis Lane does not. If she
were to discover Clark Kent's identity, this would be knowledge. But what exactly is this knowledge
and how is it represented in the statement, Superman=Clark Kent?

There seem to be just two alternatives:

*A statement of identity describes a relation which every object holds to itself, and does not hold to
any other object.

*A statement of identity describes a relation which holds between two names which refer to one and
the same object.

On alternative (1), when we say that Superman=Clark Kent, what we actually stateis that a certain
individual is identical with himself. But this is hardly news! On this reading, there is no difference in
informational value between the statement, Superman=Clark Kent and the statement,
Superman=Superman.

On alternative (2), when we say that Superman=Clark Kent, what we actually state is that the name
'Superman' designates one and the same individual as the name 'Clark Kent'. This 'meta-linguistic'
solution which Frege originally adopted, looks more promising. To know that an object is designated
by a particular name is a piece of factual information. For example, I ask what my neighbour's new
dog is called and she tells me his name is 'Bruce'. Now I know something I didn't know before.

But Frege rejects this second alternative. Why? This is what he says:

[T]his relation would hold between the names or signs only in so far as they named or designated
something. It would be mediated by the connexion of each of the two signs with the same designated
thing. But this is arbitrary. Nobody can be forbidden to use any arbitrarily producible event or object
as a sign for something. In that case the sentence a=b would no longer refer to the subject matter, but
only to its mode of designation; we would express no proper knowledge by its means.

Philosophical Writings of Gottlob Frege P. Geach and M. Black eds. Blackwell 1970.

As an undergraduate, I remember puzzling over this. What is Frege saying? Of course, I am free to
invent my own arbitrary name for Bruce. Suppose my name for Bruce is 'Bonzo'. For me, it is not
news that 'Bruce' designates the same entity as 'Bonzo'. For you it might be. Not having a great eye
for dogs, you are unaware that the naughty 'Bonzo' I point out tearing up my vegetable patch is the
very same well behaved 'Bruce' you were introduced to by my neighbour last week when you were
invited over for tea. So what's wrong with this account?

There's nothing wrong with it. It is perfectly all right. The statement that 'Bruce' designates the same
entity as 'Bonzo' will convey information to you just in the case when the statement that Bruce=Bonzo
conveys information. They are not the same statement, of course. The first statement refersto names
while the second statement usesnames, but does not refer to them. (You refer to a name when you
put it in quotation marks.) However, the two statements do the same job.

But that is precisely the reason why the meta-linguistic analysis doesn't make any contribution to
solving the original puzzle. It looks as though the meta-linguistic analysis gets you somewhere,
whereas in fact it doesn't. The question, which still hasn't been answered is, What is it that is
characteristic of all and only those cases of statements of the form A=Bor of the form The object
designated by 'A'=the object designated by 'B'which succeed in conveying factual information?

Frege's solution is to propose a distinction between the senseof a name and its reference. The
reference of a name is the entity which it designates. The sense of the name, for a given individual or
group of individuals who use that name, depends upon — now, here comes the difficult bit —
something(Frege calls it the 'mode of presentation' of the object) which is not the same as the entity
itself. In the statements, 'Superman=Clark Kent' or 'Bruce=Bonzo', the names on each side of the '='
sign have the same reference but a different sense.

Objects, as I would put it, have sides. Every object that we are acquainted with, potentially has sides
from which that object would be unrecognizable to us. Occasionally, we succeed in connecting two
disconnected sides together and recognize that fact by asserting an identity statement.

It is therefore is absurd to claim that the semantic valueof names like 'Bruce' or 'Geoffrey' is the
object which they refer to. If that were the case, one would have to conclude that we can never know
the meaning of any name. If one did know the meaning of a name, then one would have to know the
object from every possible side, knowledge to which, as Frege laconically remarks, "we never attain".

But donames have a semantic value, in Frege's sense? Is there any useful point in looking for the
mode of presentation, or senseof a name like 'Bruce', or 'Geoffrey'? Does my neighbour's dog have a
'Bruce' side and a 'Bonzo' side? Only in certain artificially restricted examples. In real life, modes of
presentation overlap in exceedingly complex ways. One would have to conclude that Frege's
argument for a sense/reference distinction for proper names as a solution to the puzzle about identity
statements is totally unconvincing. Names have a fluid and variable currency,not a fixed 'semantic
value'.

Geoffrey Klempner

Dear Alex:

It can be difficult to understand sense and reference but I think this is just because reference always
seemed to be called "meaning" as if sense was irrelevant to semantic theory. However, there are two
different aspects to meaning.

Firstly, there is thought and speech. What we refer to is the object or referent. When we "mean"
something, we refer to it and, in doing so, we use a particular sense, or a description. You may know
John as "the bloke in the pub" and your friend might know him a "the man who works in the
bookshop". You can talk about the same person without knowing it, but you are both talking of the
same person (you have the same reference), and you may come to both realise it when you come to
agree on senses — or descriptions — under which you know John. Sense is essential to this aspect
of meaning. You can't talk about John without a sense, i.e. a description under which you know him,
or you wouldn't understand what you were talking about. For Frege, this type of meaning was both
sense and reference. It is an account of an individual's understanding of a sentence. Frege called this
propositional as opposed to sentential.

The other aspect of "meaning" is the relation between a sentence and the fact in the world. When we
use language we use propositions which express our relationship to, or our understanding of, the way
the world is. There is what we know of John, but there is also John himself who embodies all facts
about himself. The sentence which contains as the referent the word John, picks out John as the
individual of whom there are facts which are true or false of him. The sentence "John works in the
bookshop" directly refers to a particular person as an objective item and doesn't need to carry sense
to have meaning. An ordinary sentence — as opposed to a propositional thought — is given meaning
in terms of its truth conditions, so is directly determined by extra-linguistic facts such as those in the
world.

Dear Charlotte:

According to Gareth Evans in The Varieties of Reference, Frege did not abandon one semantic
theory for another, but recognised that more than an extensional analysis of sentences was needed if
a theory of meaning was to encompass an account of what it is to understand a sentence.

His initial theory of meaning was in terms of truth value. The meaning of a sentence is determined by
whether or not a description is true or not of the referent. What Frege realised, which is the reason for
his sense/reference distinction, was that in many cases we might know an object or person under one
description yet not under another. What an individual knows about a referent is its sense, or
intension. The sentence relation between sentences and extra-linguistic entities is extensional and
doesn't account for what people understand by their language and a theory of meaning should be
able to provide some explanation of differences in understanding as well as being able to underpin a
theory of communication.

Evans identifies Frege's initial recognition of the problem of understanding in regard to a theory of
meaning in an unpublished letter in which Frege gave an example of a mountain discovered from
different directions by two explorers. One explorer calls it 'Afla' and discovers it's height to be 5000
metres, the other calls it 'Ateb' but knows nothing of it's height. The second explorer can successfully
refer to the mountain as Ateb and might come to discover that Ateb is 5000 metres high. He would
thus believe "Ateb is 5000 metres high" but not "Afla is 5000 metres high" because he doesn't know
that Ateb is Afla. Therefore, these cannot be the same thought: There are two senses, one referent.
It is possible not to know that the mountain is not snow-capped in the summer, so Frege's theory of
sense shows how it is possible to informatively communicate with others about objects. Frege's
example of Hesperus and Phosphorus (the evening star and the morning star) in 'On Sense and
Reference' illustrates the same point. On the extensional analysis, which Frege came to recognise as
inadequate, if you believe that "Ateb is 5000 metres high" then you would believe that "Afla is 5000
metres high".

As a theory of meaning the sense/reference account works to an extent, but lies slightly
problematically alongside Frege's account of extensional meaning of sentences. Some sentences,
about fiction for instance, include names which do not refer to objects. On a logical extensional
analysis, this type of sentence would be false because it fails to refer. If a theory of meaning was
simply extensional, a person using a sentence about a fictional character would fail to say anything.
There would be sense without reference for a class of propositions and sentences. Michael Dummett,
the main interpreter of Frege, has held that for these sentences a third, indeterminate truth value
should be introduced. Evans holds that Frege took a more Russellian view of fictional sentences.

One specific problem arising from Frege's account is what Susan Haack in Philosophy of Logics
refers to as "The morning star paradox". On the extensional analysis, the morning star is the evening
star — both terms refer to the planet Venus — which is an identity relation, and so it is necessary. If
the morning star isthe evening star then that fact could not be otherwise. However, because there
are two senses involved here, it follows that it is contingent. Not everyone knows that the morning
star is the evening star. We can easily conceive that they might have been two different stars, given
the different senses.

Rachel Browne