Here are some issues where our position on fictional objects may make a difference. I quote Amie L. Thomasson, "Fictional Entities," in
A Companion to Metaphysics, Second edition. Ed. Jaegwon Kim, Ernest Sosa and
Gary Rosenkrantz, Blackwell, 2009: 10-18, p. 18.
I know not everyone will be interested in this. That's OK! Summary: she says it matters for
1. theory of language
2. whether one accepts other objects that don't have spatio-temporal existence (symphonies, laws, etc.)
3. whether existence can be distinguished from quantification
"First, as we have seen in section 1, it has relevance for our theory of language. If we deny that there are fictional entities (and so deny that we ever refer to them), we must explain how we can have true statements involving nonreferring terms. If we accept that there are fictional entities, we must explain how we can refer to non-existent objects (if we take a Meinongian view), merely possible objects, or abstracta (whether Platonist or artifactual)—a task that is especially difficult for causal theories of reference, since none of these entities are obviously a part of the actual causal order.
Issues regarding fictional entities also have broader relevance for work in
metaphysics. If artifactualists like Thomasson are correct, then whether or not one
accepts that there are fictional characters is closely connected to the issue of whether
one accepts other mind-dependent social and cultural objects such as laws and
nations, stories and symphonies. Moreover, our stance regarding fictional entities has
central relevance for issues of ontological commitment and quantification: If the
Meinongian is right, we can quantify over entities that don’t exist, and existence must
be distinguished from quantification. If the minimalist is right, then the measure of ontological commitment is not whether or not we quantify over the relevant entities—
for if we accept that there are authors who use fictional names pretensefully in writing
works of fiction, we are already tacitly committed to fictional characters regardless of
whether they explicitly quantify over them."
here's a link to her whole paper:
http://www.class.uh.edu/phil/garson/Thomasson - Fictional Entities.pdf