Gideon Rosen's Modal Fictionalism: Counterparts, Fictional Counterparts, and Fidelity ConstraintsRobert R. Wadholm, Missouri University, 2012

It was the best of times, it was also the best of times, it was the age of foolishness, it was

also the age of foolishness, it was the epoch of belief, it was also the epoch of belief… This

is a tale of two possible worlds, an exploration of two specific incarnations of theories about

"possible worlds" (one of which is unambiguously parasitic), and an exploration of how

counterpart theory might break fidelity constraints. And this is not fiction.

Modal Realism & Fictionalism

David Lewis' modal realism (elucidated in his A Philosopher's Paradise and other writings)

proposes that talk of possibilities, like "I could have eaten a hamburger for breakfast this

morning," are truly just talk of possible worlds that really exist in some concrete way. What

the proposition means is that I (actually, my counterpart in a possible world, not me) did

eat a hamburger for breakfast this morning in some possible world. And there are perhaps

infinite possible worlds, where every thing or event that was possible in our world actually

occurred in theirs. Perhaps I ate a unicorn hamburger for breakfast in one of them. In this

form of realism, modal language ("possibly","necessarily," etc.) is about other worlds, not

our own.

Second, Gideon Rosen's modal fictionalism (expounded upon in the clearly titled

Modal Fictionalism) points out that philosophers (and even normal people) are a bit

incredulous concerning the concrete existence of all of those possible worlds. They are too

messy ontologically, and more profoundly, they are just too hard to believe in. Rosen

suggests a seemingly more palatable (yet parasitic) mutation of Lewis' realism, by positing

that possible worlds, and modal language about them, is actually about stories/fictions

concerning possible worlds. Rosen gently wraps most of Lewis' realism inside of a book

Robert R. Wadholm, Modal Fictionalism

jacket, and declares the piece "fiction". As long as the worlds are merely in fictional stories,

we may be more likely to buy them. All the suggested benefits of realism, without the

greatest weakness--unbelievability. Sure to be a bestseller.

The Argument from Concern

In Modal Fictionalism, Rosen presents a powerful argument against both kinds of possible

world talk: what he calls the argument from concern, built from Kripke's objection to

Lewis' counterpart theory. The argument goes something like this:

- In realism, when I say "I could have eaten a hamburger for breakfast," I really mean

"My counterpart in a possible world did eat a hamburger for breakfast."

- This counterpart to me is never identical to me.

- Someone else ate the hamburger, not me.

- I couldn't care less if someone else ate a hamburger, no matter how much like me

they are. I'm still hungry.

Rosen admits that this objection applies to both realism and fictionalism. In realism, why

should I be concerned about my counterpart, and in fictionalism, why should I be

concerned about my fictional counterpart? Here, Rosen points out that fidelity constraints

may be broken or mangled: my speech (and caring) about myself in modal language seems

to be incompatible with the ideas that I am just talking (or caring about) my counterpart in

another world. Fidelity to my original modal beliefs is gone.

Rosen responds to Kripke's objection by pointing out that maybe if we accept

realism (or his more charming and elegant fictionalism) we will start to care about our

counterparts in other worlds and the fidelity will reemerge (whether in the form of realism

or fictionalism). Rosen admits that Kripke's objection is a strong one, though perhaps not

undefeatable. I will suggest what in my mind seems to be a stronger argument based on

the beginning of Kripke's objection.

2

Robert R. Wadholm, Modal Fictionalism

The Argument of Identity in Modal Language

The central thesis of counterpart theory seems to be that counterparts are not

identical to the objects or people in this world of which they are counterparts (as opposed

to the idea of trans-world individuals). I think that is the argument. When I say "I could

have," I don't mean "My counterpart did." I and my counterpart are not identical. If we

were identical, I would have a hamburger in my stomach. I would live in a possible world

other than this one. I would have higher cholesterol levels than I do right now. If we say

that my proposition "I could have" really means "my counterpart did," there is no fidelity to

the subject of the sentence, namely me. The only reason I don’t care about the breakfast of

my counterpart (following Kripke) is because it was not my breakfast, it was that of a

stranger. The not caring is important, but the identity of who I am not caring for is more so.

If I say “Socrates was bald,” I do not mean a stranger to Socrates was bald. This

non-modal proposition is about Socrates, not a stranger to Socrates. Similarly, if I say

(using modal language) “I could have eaten a hamburger for breakfast,” I do not mean “A

stranger to me ate a hamburger for breakfast.” If I exist, and at least one other thing

besides me exists as well, when I say “I am”, I do not mean “That other thing is.” I am me,

that thing is that thing. Both realism and fictionalism (if tied to counterpart theory) seem to

have snuck in someone else into my proposition about me (and left me out of it entirely!).

No wonder I don’t care: it wasn’t me that the proposition was about. Realism and

fictionalism have broken fundamental fidelity constraints: propositions about one thing

(me) are changed to mean propositions about a completely different thing (not-me). This is

not merely a “substantial revision” of modal belief, as Rosen suggests. It is a substantial

revision of the original proposition.

When I say “I could have eaten a hamburger for breakfast” is true, I say it is true

about me. If we change who the proposition is about, we must reevaluate whether the new

proposition is true (because it is a new proposition, not the same one). In realism and

fictionalism, we would change the proposition to “My counterpart (or my fictional

3

Robert R. Wadholm, Modal Fictionalism

counterpart, or fictionally my counterpart) ate a hamburger for breakfast.” This is a

different proposition. Because we cannot verify that my counterpart even exists (we have

no way of knowing about him for sure), we can’t say whether this is a true or false

statement. On the other hand, my fictional counterpart does exist (I just made him up), and

he did eat the fictional hamburger for breakfast (and it was fictionally delicious). However,

the two propositions “I could have” and “My (fictional) counterpart did” are fundamentally

different propositions, and other than similarities concerning details, don’t say anything

about the other (or depend on the truth of the other, or constitute the truth-maker for the

other). It is my belief that part of the utility of realism is undone by Lewis’ counterpart

theory, if not by other objections (like unbelievability). Fictionalism may still stand (and

truthfully, I think it is a beautiful idea), but if counterpart theory is applied, it says nothing

about me or propositions about me, and thus fails to address modal language about persons

or specified individuals (like I, Socrates, that horse, the Pacific Ocean, and my hamburger).

That kind of modal language seems to be out of bounds (as far as applying analyses go) as

long as counterpart theory is held.

Rosen (or Lewis) might respond to such a critique with a simple assertion:

counterparts, while not identical, are equivalent to individuals and are not merely

comparative. Equivalence may be construed to mean that two things or persons may be

interchangeable now (or in the past), but at some point may diverge. I could have been

substituted for my counterpart before breakfast (and that fateful hamburger) and no one

would be the wiser (even me or him). So my counterpart would not have the same identity

as me (we are not the same person), but would be equivalent to me in the past though not

in the present. So we could swap “I could have” in my proposition with “My counterpart

did” since the two are equivalent.

This assertion of equivalence, however, suffers the same problems as identity: the

person we are talking about is still not me. Even if x+2=5, and thus x=3, we cannot always

say in every equation that x=3 (x does not equal 3 in many equations—just this particular x

equals 3). We mean this x equals three (in this context). Similarly, if x=3 and y=3, we

4

Robert R. Wadholm, Modal Fictionalism

cannot always say that in every equation x=y (in some equations x=10 and y=2x). In

general, two individual things may be equivalent at some point, but they might not be

identical, and if we speak about specific things in specific contexts they are not necessarily

equivalent outside of those contexts. If we specify this thing or person, equivalence just will

not do. Specific things and individuals are not equivalent indexically. The realist’s

proposition about my counterpart and his hamburger is not equivalent to me and my

hamburger. I am in this world, and I am referencing myself and my own hamburger. A

possible hamburger just will not do for breakfast.

(And Bob and his hamburger lived happily ever after…)

5