Vacuous names and fictional entities: do names like 'Sherlock Holmes' or 'Vulcan' refer to anything, and if so, to what?

Saul Kripke, answered from the texts and cited to the page.

The position in brief

Two cases require sharp separation. Fictional names — like "Sherlock Holmes" or "Hamlet" — appear in fictional discourse where the author pretends to refer; outside that pretense they refer to fictional entities understood as abstract artifacts created by their authors at the time of composition.¹ There really is something to which "Sherlock Holmes" refers — not a flesh-and-blood detective, but an abstract object that did not exist before Conan Doyle wrote the stories. Vacuous names of putative non-fictional entities — "Vulcan" (the planet Le Verrier postulated between Mercury and the Sun to explain Mercury's perihelion precession), "phlogiston," "the philosopher's stone" — are genuinely empty: they fail to refer, even to abstract artifacts, because they were not introduced as fictional from the start.²³⁴⁵

Sources

1. What I'm saying about pretense could be applied, mutatis mutandis, to mistakes and other cases like that—error as opposed to pretense or fiction. (Russell does, perhaps, by his certainty condition, avoid the possibility of mistake.) And one should not regard it as strange that such errors can exist—one should regard it as natural and inevitable.

   3. Vacuous Names and Fictional Entities, pp. 61–62

2. They are not ghostly possible entities; they are abstract entities of a certain sort that exist in virtue of the activities of people.

   3. Vacuous Names and Fictional Entities, pp. 64–65

3. So the fi ctional character is not something which is being said not to exist. If you say that the fi ctional character didn't exist, you would be wrongly assimilating this case to the case of Moloch, where one can say truly, as I did before, there was no such god as Moloch, that Moloch didn't exist. 11.

   Reference and Existence, p. 148

4. Some of these statements are true and do express propositions when we are just reporting on what is said in the story, but, as I said, that is not the sort of usage in question here. (Similarly, sometimes one uses the statement just to say that the story says or implies that p, but that is a different sort of usage.) Nor am I talking about statements about fictional characters.

   3. Vacuous Names and Fictional Entities, p. 68

5. If this means that it is not truly a name of the figure in question, or that it fails to refer to him or her, I don't think this is right. When in War and Peace Tolstoy mentions Napoleon, and has him as a character in the work, he is talking about Napoleon (See Chapter 9.) 12 In the preface, the supposed editor of the manuscript says that the names are not real.

   3. Vacuous Names and Fictional Entities, pp. 58–59

Where the argument is made

• Reference and Existence: The John Locke Lectures (2013, delivered at Oxford 1973)
• "Vacuous Names and Fictional Entities" (Philosophical Troubles Ch 3, ms originally from the late 1960s, widely circulated long before publication)

How the argument was received

Developed in the 1973 Locke Lectures, published in 2013 as Reference and Existence. The position is a measured one: Kripke acknowledges throughout that the ontology of fictional entities is delicate and that he is offering "a view" rather than a worked-out theory. The abstract-artifact account has been taken up and refined by van Inwagen, Thomasson, and others; resisted by strict Millians like Salmon, who prefer to handle fictional discourse without ontological commitment to fictional entities. The vacuous- name side is less contested: most contemporary philosophers of language agree that "Vulcan" simply fails to refer.