Chapter 4 – A dualist theory of fictional objects
4. Fictitious objects
4.1 Indeterminacy
The first problem that must be addressed is indeterminacy. If the Paul Atreides about which I imagine is a possible person, which possible person is he? There seems to be no principled reason to choose one possible person accurately described by the fiction over the many, perhaps infinitely many, others. Implicit in this objection, famously made by Saul Kripke, is that fictional names like ‘Paul Atreides’ are rigid designators – terms that pick out the same object in all possible worlds. The name ‘Frank Herbert’ as we use it in the actual world picks out the man who is author of Dune in the actual world. However, we can use this rigid designator to discuss that very man as he is in other possible worlds, regardless of whether or not his name is ‘Frank Herbert’ in that world or if he’s the author of Dune in that world. Thus, it does not function like a role such as ‘the current president of the United States.’ In the actual world, that expression refers (unfortunately) to Donald Trump. In some other possible world, it refers to Hillary Clinton. It is a non-rigid
designator.
It's reasonable at first blush to suppose that all names come to have their referents in the same way and that they function as referring terms in the same way. After all, it may seem a bit strange to accept the causal theory of reference for some names but be a descriptivist about others. However, standard views of fiction, particularly fictive utterance views, do deny that fictional names function like ordinary proper names, and have principled reasons to maintain this small disunity. I think there are two ways to go about an analysis of fictional names, and which is better will depend on one’s prior commitments. If one is committed to the view that proper names are directly referring expressions, then we can
245 This may be redundant if one takes propositions to be linguistic entities, but I’ll remain neutral with respect to the ontological status and metaphysical nature of propositions.
proceed with a view that the names of fictitious objects do not function in the same way as everyday names do. The difference would not arbitrary or ad hoc but would have to do with how we come to know these names. If one is put off by the disunity of suggesting a completely different semantic analysis for fictional names but is not particularly committed to a direct reference view, we could instead conceive of names as predicates.
To put it very simply, to use a name is to predicate of an object that it has that name.246 This is controversial, because proper names are widely thought to have only their one referent as their semantic content. Though many people have the ‘same’ name, on direct reference and definite description accounts, the ‘same’ names that refer to different people are not really the same, but merely homophones. On a predicate view, they genuinely are the same name, because they predicate the same property to objects: the property of being called that name.247 Nevertheless, instances of names that appear to be the subject of an utterance, or in the ‘argument position,’ can still designate a limited number, including just one, of the things that bear the name. This restriction is done by context.248 In the case of fictional names, it is true of each possible Atreides candidate that he has the name ‘Paul Atreides’ in his world. However, not all possible people who bear the name ‘Paul Atreides’ are one of our candidates.249 The context that restricts the designation of the name is fixed by the same thing that fixes the context of our utterances: the practice. Specifically, the work and the ‘world’ of the work, determined by the practice, fix the context.
Even on such an accommodating view as predicativism, there’s still something unusual about our use of fictional names. Fictional names are count-nouns on this view, meaning that we can refer to multiple people with a single use of the name.250 However, even in those uses, we still pluralize the name. That is, we will talk about all the Alfreds we know or the two Clotildes at the party. However, we only seem to talk about the one Paul Atreides, despite the name and context picking out many possible people. Though this is 246 Kent Bach, “The Predicate View of Proper Names,” Philosophy Compass 10, no. 11 (Oct. 2015): 772. 247 Delia Graff Fara, “Names are Predicates,” The Philosophical Review 124, no. 1 (Jan. 2015): 65. 248 Ibid., 75.
249 I have already restricted the context to possible people, having established that no actual person could be our fictional Paul Atreides regardless of what his name is.
strange, I don’t think it’s unique to fictional names. Rather, I think it’s a feature of using names for possibilia in general. Any time we talk about possibilities, we necessarily talk about many, even infinitely many, possible individuals; this is because the descriptions we use to pick out possibilia are always indeterminate. Typically, we talk about the possible counterparts of actual people related by identity. Because I am in some way identical to all the possible blond-haired Ash counterparts, it makes some kind of sense to just talk about ‘the possible blond-haired Ash.’ Perhaps when we pick out possible counterparts by some other relation, such as the non-identity relation of being an Atreides candidate, the
mechanism is the same despite the counterparts being non-numerically identical. However, there may be a principled reason that our talk about fictional possibilities is anomalous. Consider first the name of the actual-world person Frank Herbert. Without committing to any particular view of reference, let’s simply describe the facts as we know them. Frank Herbert was born and at some point, dubbed by his parents ‘Frank Herbert,’ and this name was confirmed by our society in the typical ways, with official documents and perhaps a christening. From then on, that particular person was introduced and known by this name.251 Some people were introduced to him directly and came to have knowledge of the reference of ‘Frank Herbert’ by their own perceptions. Others, such as myself, came to know Frank Herbert via a description: the author of Dune and other novels. Though my knowledge of the man is limited, the intentional object of my belief that Frank Herbert wrote Dune is the same as those who met him directly.
Now let’s think about Frank Herbert’s possibilities. We know that there is a possible Frank Herbert who is not called ‘Frank Herbert’ in his world but is named instead ‘John Smith.’ We also know that there is a possible Frank Herbert who never wrote the novel Dune. We also know that there is a possible Frank Herbert who was a peanut farmer. We also know that there is a possible Frank Herbert named ‘John Smith’ who never wrote Dune and was a peanut farmer. How do we know all these things? Merely because they are conceivable:
251 This is only intended as a description of typical usage of the name, not any commitment to a view about what makes the name refer to a particular person.
this is how we get our knowledge about possibilities.252 There can be no knowledge of mere possibilities gained by directly perceiving the merely possible, as possible worlds are causally isolated from one another. Yet notice that in each of our propositions about possible Frank Herberts, ‘Frank Herbert’ refers to the very man regardless of whether he is named ‘Frank Herbert’ in his world. We use the name as a rigid designator, fixed by an actual-world person. We comfortably refer to his possible counterparts with that name, and there is no confusion caused by the proposition “Frank Herbert is named ‘John Smith’” when we understand that we are discussing other possible Frank Herberts.
Most fictitious objects are different.253 There is no one person about whom we imagine in accordance with Dune that we rigidly designate with the name ‘Paul Atreides.’ When we imagine about Paul Atreides, we are not imagining the possibilities of any one actual person, conjuring up propositions about their possible-world counterparts and only their counterparts (whether those counterparts enjoy an identity relation with the actual person or something like Lewis’ counterpart relation).254 The possibilities we imagine are instead what I’ll call free floating; they are possible states of affairs known to be possible only through their conceivability and that are unanchored to actual-world objects.
If it’s conceivable that the reader or I are named ‘Paul Atreides,’ live far in the future, and otherwise have all the properties and experiences about which we imagine in accordance with Dune, both the reader and I imagine, in some sense, about our own possibilities when we imagine about Paul Atreides. But this is only true in the loosest sense; we need not ever explicitly think that we, or any other actual person we know of, could have been Paul Atreides.
252 This is a controversial point but arguing for it is beyond the scope of this work.
253 I say ‘most’ because sometimes we imagine the possible counterparts of actual people who are rigidly designated by their name in works of fiction, such as the common example of Napoleon in War and Peace. 254 From now on, I will frequently refer to either identity or a ‘counterpart relation.’ By this I mean
particularly Lewis’ I-relation, not something else like Lewis’ counterparts by acquaintance, which will be mentioned shortly. Though the I-relation is not identity, it functions similarly to identity in that it makes modal claims true. The basic account argued for here ought to get along either with transworld identity or something like the I-relation, without having to elaborate too much about what exactly the I-relation is beyond the truth-maker for modal claims.
It's worth recalling the distinction discussed in Chapter 2 between object-based imagining and propositional imagining. Object-based imagining is the kind of imagining in Walton’s tree stump example. To imagine that a tree stump is a bear during an imaginative game is to imagine of an actual thing that it is something else. Propositional imagining need not involve imagining of any actual thing that it is something else. So, in keeping with the example, imagining while you sit at your desk that you are in a forest and that there is a bear before you is propositional imagining. There is no prop in your environment to materially represent the bear. But even much of propositional imagining is object-based in a different sense. To imagine that you are in a forest is to imagine something of an actually existing object: yourself. Represented in your imaginative model is the possible state of affairs that you are in a forest.
To imagine free-floating possibilities, on the other hand, involves the creation of a role that possible objects fulfill. To imagine my possible selves involves no such creation – the ‘role’ exists independently of any creative activities of mine as a prior, metaphysically existing relation (either an identity or counterpart relation). The roles created by
imaginings of free-floating possibilities are created via descriptions that possible people and objects fulfill. Thus, all the possible people who fulfill the role of my Dune model’s Paul Atreides are not identical to one another in the way that I’m identical to my possible manifestations. This view is a notational variant of Lewis’ counterparts by acquaintance and Currie’s account of transfictive uses of fictional names.255
We ought not to think that these are two distinct types of imagining, as with the distinction between object-based and propositional imagining, but rather imagining about two distinct types of things. For example, in War and Peace, ‘Napoleon’ rigidly designates the actual historical figure. Thus, in my War and Peace model, the content of my imaginings includes a collection of possible Napoleons: possible men who must bear an identity relation to the actual man Napoleon and who do and say all the things Napoleon does and says in War and Peace. ‘Napoleon’ in War and Peace models is a rigid designator and is 255 The major point of departure from Lewis’ view is that I’m not convinced it’s necessary that we limit the candidates for fictitious characters to those possible people residing in a world in which the fiction is told as known fact. More on this to come.
not a free-floating possibility. In the same model, I imagine a free-floating possible person Pierre Bezukhof. ‘Pierre Bezukhof’ is an entirely created role fulfilled by many, perhaps even infinitely many, possible people who do not necessarily bear an identity relation to one another. ‘Pierre Bezukhof’ functions more like ‘the president of the United States.’ Alternatively, the name is a true predicate of all the possible objects that are restricted by the context of War and Peace.