Programme

Kripke famously thought that pretend-reference only gets us so far: it has a role in story-telling, maybe in the way consumers of stories engage with stories, but not, for example, when it comes to parafictional and metafictional statements ('e.g., [In the Holmes stories] Holmes is a brilliant detective', 'Hamlet is a fictional prince of Denmark). Nor does it play a role in negative existentials ('e.g., Bezukhov doesn't exist'). In this talk I argue that relaxing this restriction has virtues even Kripke might have acknowledged — if only up to a point. Although few would recommend dropping the restriction in the case of metafictional uses, more are prepared to accept a pretend-referential account of parafictional uses of fictional names. Far less widely accepted, however, is the claim that the use of real names in telling stories ('London', 'Napoleon', etc) also involves pretend-reference. In this talk I discuss such a treatment of 'real' names and the way it naturally gives rise to the idea of surrogate fictional objects (an idea that even Kripke accepted at one point), and I suggest that such a treatment may hold some general lessons for the analysis of negative existentials and metafictional statements.

Numerous philosophers—champions of abstract fictional characters, for instance, and neo-Meinongians of various stripes—have appealed to intentional objects in relation to fictions: to individuals that stand in appropriately distinguished relationships to the properties that, according to the relevant stories, certain beings possess, while not themselves actually instantiating those features. This strategy promises to allow us to treat philosophically puzzling uses of language, in making statements about fictions, as continuous with their employment in ordinary contexts, while preventing the logical chaos of the former—their inconsistency, say, and their incompleteness—from escaping into actuality. The talk will argue that there are important limits to the extent to which intentional objects can be used in providing a semantically uniform and logically well-behaved account of uses of linguistic devices within and without fictions. It will argue, in particular, that intentional objects do not help us better to understand some common phenomena that are perplexing in ways that parallel more familiar puzzles generated by suitable uses of singular terms: viz. various phenomena involving uses of quantifiers, and other domain-sensitive expressions.metafictional statements.

I will revisit the phenomenon of authors' error with respect to introducing (and subsequently using) names for fictional characters. I aim to improve upon my previous argument in favor of artifactualism about fictional characters — a form of realism according to which fictional characters are nonconcrete human-created objects, that is, nonconcrete artifacts.  In prior work, I envisioned a counterfactual scenario in which Tolstoy, while writing War and Peace, mistakenly construes Andrei Bolkonsky as a historical figure just as Napoleon, also featured in the novel. If one is a realist about fictional characters (and about posits of failed hypotheses like Vulcan, a view criticized by Jeff Goodman, to whom I responded), then, given counterfactual Tolstoy's error about Bolkonsky, counterfactual-Tolstoy plausibly created a fictional character Bolkonsky and did so inadvertently. My broader conclusion was: when considered in a broader context, fictional characters emerge as but one of various kinds of social and cultural constructions—for example, words, songs, poems—for which a metaphysics of nonconcrete artifacts is our best available realist theory, and for which inadvertent creation phenomena like that in the counterfactual-Tolstoy scenario are unmysterious and even expected given Saul Kripke's general arguments about name-users' potential error.

In the present talk, I will first explore how, along the lines of proposals about name-using practices by Mark Sainsbury and Dolf Rami, we can generalize my earlier arguments in a way that remains ontologically neutral about posits like Vulcan as well as about fictional characters and other social and cultural constructions, thereby leaving open whether we opt for a realist (specifically an artifactualist) or an irrealist alternative for these. While writing War and Peace, Counterfactual-Tolstoy, with his initial uses of the name 'Andrei Bolkonsky', has inadvertently introduced a new name-using practice. So did Le Verrier in formulating his hypothesis about Vulcan. Second, I aim to show that this move to ontological neutrality affords sharper focus on locating and systematizing some metaphysical and semantic motivations and their interplay. This motivation package affords (i) reasons to opt for artifactualism over irrealism for fictional characters and for a range of other social and cultural constructions like words, songs; (ii) reasons against artifactualism and in favor of irrealism about posits of failed hypotheses like Vulcan — agreeing with points Alberto Voltolini and others have made about names like 'Vulcan'. I'll also briefly explore connections to numerous workshop participants' contribution to the exceptionalism/nonexceptionalism debate about names like 'Napoleon', 'Austerlitz' imported into historical works of fiction like War and Peace.

Fictional characters don't exist, but that doesn't stop us from talking or thinking about them. Does this mean that one can refer to something that does not exist? Not necessarily: we can maintain that reference implies existence, and account for fictional discourse by arguing that the act of reference is as fictional as the entity it refers to. Metafictional discourse, however, lends itself to another kind of analysis, one that is also compatible with the idea that reference implies existence. According to this analysis, fictional characters are not only fictitious entities that are pretended to exist: they are also, as cultural creations, things that really exist, and to which one can legitimately refer.

Everyone has the intuition that, as regards the properties that are predicated of a fictional character (fictum) within a story, such properties are not possessed in the same sense as properties that are predicated of ordinary individuals. Yet this intuition can be accomodated in different ways.

Many different people (e.g. artefactualists like Salmon 1998 and Thomasson 1999, (im)possibilists like Priest 2016) appeal to story-properties – being F in story S – or to converse-intentional properties of the kind being represented as F; anyway, representation-dependent properties (Crane 2013). One may see any such property as a species of the corresponding genus being a fictional F, viz. being F in fiction, being fictionally F. Meinongians instead appeal either to special properties – socalled nuclear properties, to be distinguished from extra-nuclear ones (Parsons 1980, Routley 1990) or to a special form of predication of one and the same property, internal predication, to be distinguished from another form of predication, external predication (Zalta 1983, Castañeda 1989). The appeal to the second distinction is better than the appeal to the first distinction (Voltolini 2006).

In this talk, I want to claim that 1) Being fictionally F and being internally F are different features; 2) The latter is better than the former for it is explanatory richer.

In this paper, I aim provide a formal semantic analysis of fictional modifiers like “fictitious” and “fictional”. I will argue that unlike “fictitious”, “fictional” is a type-ambiguous expression. With respect to the first type it is a modifying adjective like “fictitious”; and relative to the second type it is a polysemous intersective adjective. Furthermore, I will show that the pair “mythical”/mythological” behave similar like the pair “ficticious/fictional”. The expression “imaginary” and “hallucinated” are modifying adjectives without intersective counterparts in English. I will provide intuitive tests and inferential patterns to distinguish the different kinds of predicates. Finally, I will show how the modifying meaning transform existence-entailing predicates into non-existence-entailing predicates.

Metafictional statements are statements which display an external perspective on a fictional character (object, event, place). Paradigmatic examples are so-called creationist locutions like "Mary Shelley created Frankenstein's monster" which is typically taken to be true in the real world, while false in the fiction (indeed, in the fiction, Frankenstein's monster was created by Victor Frankenstein). Many philosophers (following Kripke 1973 and Van Inwagen 1977) take metafictional statements to show that fictional characters exist and fictional names typically refer to them in some contexts of use. Others (following Evans 1982 and Walton 1990) rather construe metafictional statements as sophisticated fictional statements, holding on to the idea that fictional characters do not exist and that fictional names never refer.

The linguistic data, grounding this debate between realism and antirealism, is in fact a lot messier than theoretical philosophers usually acknowledge. I think this is time to look at the data with a neutral frame of mind, and try to make subtler distinctions within metafictional statements. I will propose a systematic taxonomy for metafictional statements, based on both philosophical and non-philosophical talk. Such taxonomy is interesting for its own sake, for it aims at clarifying what "external" means in "external perspective". It will also shed interesting new light on the realism vs. antirealism debate, by showing where each side falls short of a proper analysis of the data.

In this paper I suggest that focusing on the ontological commitments of metafictional discourse about fictional characters is misleading. Such discourse is relatively rare outside the philosophy context. Most discourse is 'mixed', so that fictional characters are construed as human beings and abstract objects in the very same sentence: 'Dorothea Brooke and Isabel Archer are fictional characters who choose terrible husbands'. Even some 'purely' metafictional discourse seems mixed, depending on how it is read: 'Dorothea Brooke is a fictional character created by George Eliot; she doesn't exist'. Following Recanati (2018) we can think of such statements as co-predications, attributing different categories of property to 'the same thing' in the same way as 'lunch' may be used to designate the event and the food. In this paper I consider how best to account for such cases from a non-realist perspective.