The following is the first installment of an article on the philosophy of Gottlob Frege. Students of philosophy may find here something of interest.
Among the comic strip faithful it is well known that
(1) Lois believes that Superman is a hero
(2) Superman = Clark Kent.
However, it is just as well known that
(3) Lois believes that Clark Kent is a hero
is false. Indeed, she feels him to be somewhat of a coward. But rather than dismissing the Daily Planet’s best report as an eccentric, readers of the comic find that her beliefs about Superman (i.e. Clark Kent) make the couple’s relationship intriguing.
But how can this be? According to Frege, the reason Lois can have one belief and fail to have the other (and remain sane) is that her two beliefs are about different things. Of course saying this is merely to say what every speaker already knows. But to explain the semantics of it is to do something philosophically important. And the means by which Frege explains the difference is by his celebrated distinction of sense and reference.
The mere distinction of sense and reference is not enough to account for Lois’ doxastic consistency however. For although ‘Superman’ and ‘Clark Kent’ may have different senses, both still refer to Superman. And this being the case, one should be replaceable by the other salva veritate. Frege, of course, understood this problem well and offered a further distinction. Names that occur after a believes-that clause or any other propositional attitude clause do not have their normal sense of reference. Their reference becomes what was the original sense (sense of the name or definite description outside a propositional attitude clauses) and the sense becomes what Frege calls the indirect sense. Thus ‘Superman’ in (1) does not refer to Superman, but the the sense of ‘Superman’.
There are difficulties with Frege’s view however. If we take (1) together with
(4) Superman is a hero,
we can conclude that
(5) Lois believes at least one thing that is true.
Given Frege’s account of indirect sense, though, this inference is cut off. This is due to the fact that ‘Superman’ in the substantival clause of (1) and ‘Superman’ in (4) refer to different things â€“ or, better put, different types of things. The first refers to the normal sense of ‘Superman’ (to use Kaplan’s convention, mSupermanm), the second refers to Superman. (Compare the inference of (5) from (3) and (4) where the same type of non sequitur is committed.) One might propose that we substitute
(6) It is true that Superman is a hero
for (4) and thereby assure a valid inference to (5). Both tokens of ‘Superman’ now refer to the same thing. (Compare the invalid inference of (5) from (3) and (6); the proper names in the substantival clauses now refer to the right type of thing, but not the right thing since mSupermanm is not identical to mClark Kentm. This shows that the inference of (5) from (3) and (4) is doubly in error.) The problem is that Lois’ true belief turns out to be about mSupermanm rather than Superman.
A further problem to the Fregean picture is what to make of propositional attitude clauses embedded within another propositional attitude clause. Take, for example,
(7) John believes that Lois believes that Superman is a hero
or even more gruesome,
(8) Paul believes that John believes that Lois believes that Superman is a hero.
Assuming we can make sense of the indirect reference (and sense) of ‘John’ (i.e. specify conditions of synonymy), what do ‘Lois’ and ‘Superman’ refer to? To carry on Frege’s analysis we would have to say the indirect indirect reference and indirect indirect indirect reference of ‘Lois’ and ‘Superman’ respectively. And concomitant with each stratum of reference is a sense. Indirect senses clutter the world up enough. Now it seems as if we are forced to countenance an indefinite number of types of senses. Surely something has gone awry.
Indirect reference is tolerably clear enough; it is the standard or usual sense of the term. And the indirect sense is what gets at the indirect reference. But this is to put the cart before the horse. Frege’s view is that sense is the conduit to reference. One object can be referred to by multiple senses, but one sense can only refer to one object. As Russell says, “There is no backward road from denotations to meanings.”
These are the standard problems with Frege’s distinction between sense and reference when applied to propositional attitude clauses. In the next section I shall cover the deeper criticism of his view as given by Quine.