Fictional conundra

Bill says,

As I see it, the central problem in the philosophy of fiction is to find a solution to the following aporetic dyad:

1. There are no purely fictional items.
2. There are some purely fictional items.

The problem is that while the limbs of the dyad cannot both be true, there is reason to think that each is true.  (1) looks to be an analytic truth: by definition, what is purely fictional is not, i.e., does not exist.  George Harvey Bone, the main character in Patrick Hamilton's 1941 booze novel Hangover Square, does not now and never did exist.  He is not a real alcoholic like his creator, Patrick Hamilton, who was a real alcoholic.  What is true is that

3. Bone is a purely fictional alcoholic.

That (3) is true is clear from the fact that if a student wrote on a test that Bone was a teetotaler, his answer  would be marked wrong.  But if (3) is true, then, given that nothing can satisfy a predicate unless it exists, it follows that

4. Bone exists

and, given the validity of Existential Generalization, it follows that

5. There is a purely fictional alcoholic.

But  if (5) is true, then so is (2).

Bill suggests a number of responses to this aporia.  My choice is

E.  Dissolutionism.  Somehow argue that the problem as posed above is a pseudoproblem that doesn't need solving but dissolving.  One might perhaps argue that one or the other of the dyad's limbs has not even a prima facie claim on our acceptance.

The puzzle hinges on what we mean by ‘a fictional X’, for some concept term X.  Is this itself a concept term? It’s not clear.  I think we can be mislead into thinking that it is a concept term, and this gives rise to Bill’s paradox.

The usual way of understanding 'a Y X' involves a conjunction of the concepts X and Y.  The extension of Y X is thus a subset of the extension of X.  Example: X=Roman, Y=female.  But this goes wrong for Y=fictional.  When we list the alcoholics we don’t count the fictional ones, whatever they may be.  If we must interpret ‘fictional alcoholic’ in this way we must conclude that its extension is the empty set.  Similarly 'fictional X' for any concept term X.  Hence we are lead to Bill’s (1):  There are no [purely] fictional items.

Why then is Bill’s argument for (2) rather compelling?  I suggest that we are inclined to read

n is a [merely] fictional X

as a surface transformation of

[merely] fictionally, n is an X,

And this we take as

[merely] in some piece of fiction, n is an X.

Thus 'Bone is a fictional alcoholic' becomes 'fictionally, Bone is an alcoholic', which is understood as 'In some piece of fiction it says that Bone is an alcoholic', and this we take to be true by virtue of Hamilton's works.  This kind of transformation can be seen elsewhere. Compare

n is a [merely] fictional X   <--->  [merely] fictionally, n is an X

with

n is a possible X   <--->  possibly, n is an X
n is a past X   <--->  pastly, n is an X
n is a real X   <--->  really, n is an X
n is a false X   <--->  falsely, n is an X

On the right, sentential operators; on the left, pseudo concept terms.

Thus the inference from 'n is a fictional X' to 'there is a fictional X' is not valid unless the latter is read as 'fictionally, there is an X'.  That is, 'Some piece of fiction says there is an X'.

The interpretation of 'fictional' as a disguised sentential operator fits nicely with a phenomenon that Bill notes parenthetically: 

As Kripke and others have noted, there are fictional fictions, fictional plays for example, such as a fictional play referenced within a play.

A female female Roman is just a female Roman.  Doubling up the qualifier 'female' tells us nothing more.  But 'n is a fictional fictional character' can be seen as 'fictionally, fictionally, n is a character'.  That is, in some story there is a story in which n is a character.  This alone tells us that there is something odd about 'fictional' as a concept term.
 
None of this says anything as to the existence or otherwise of van Inwagen 'creatures of fiction'.

In an earlier  post  Bill claims to have refuted this kind of approach to the puzzle.  He says that in the 'story operator' solution,

4. Sherlock Holmes is a detective
5. Sherlock Holmes is fictional

are elliptical for, respectively,

6. In the Conan Doyle stories, Sherlock Holmes is a detective
7. In the Conan Doyle stories, Sherlock Holmes is fictional.

but (7) is clearly false.  I reply, of course, that 'fictional' is a pseudo concept term and that (5) makes an assertion 'outside' the Conan Doyle stories.  The move from (5) to (7) is invalid.  Bill  says that the argument can be made with other 'extranuclear' terms such as 'merely possible' and 'mythological.'  Indeed it can, and these turn out to be pseudo concepts also.

See also Actualism and Presentism.

I should add that there is, of course, a perfectly valid use of 'fictional' as a concept term, and that is in sentences such as 'Jane Austen's Emma is fictional', where we are referring to the novel and not the young woman it is about.  To apply 'fictional' to anything other than a representation is to make a category mistake.  More on how representations  come into this in Ficta and impossibilia