Objects of Inquiry in Philosophy of Language and Linguistics

by Piotr Stalmaszczyk (Volume editor)
Edited Collection 230 Pages

Table Of Content

| 7 →


Eros Corazza

ILCLI, The University of the Basque Country, Donostia, Spain

Basque Foundation for Science, Bilbao, Spain

Carleton University, Ottawa, Canada


C.M.K. Djordjevic

University of Zurich, Switzerland


Bohumil Fořt

Masaryk University, Brno, Czech Republic


Christopher Genovesi

Carleton University, Ottawa, Canada


Daniela Glavaničová

Slovak Academy of Sciences, Bratislava, Slovakia


Richmond Kwesi

University of Ghana, Ghana


Ben Martin

University of Bergen, Norway


Stefano Predelli

University of Nottingham, UK


Marga Reimer

University of Arizona, USA

margareimer78@gmail.com ← 7 | 8 →

Göran Rossholm

Stockholm University, Sweden


Piotr Stalmaszczyk

University of Łódź, Poland


Alberto Voltolini

University of Turin, Italy


| 9 →

Piotr Stalmaszczyk

University of Łódź, Poland

On Objects of Inquiry in Philosophy of Language and Literature

The first two volumes published in the series Studies in Philosophy of Language and Linguistics concentrated on various topics connected with the semantics of proper names and reference.1 Individual studies discussed abstract names, proper names, definite descriptions, types of reference, identity, and several other important recent developments in the field of (analytic) philosophy of language.

In this volume, the objects of inquiry – the possible common objects of inquiry in philosophy of language and literature – once again include proper names, however, this time with special focus on fictional names; authors also investigate the ontological status of fictional names and fiction, metaphysics of fictional characters, truth in fiction, metaphor, representation, interpretation, sense and nonsense, and other issues which in interesting and inspiring ways interconnect research in philosophy of language and philosophy of literature.

In a recent study devoted to metalinguistic descriptivism, Manuel García-Carpintero has presented the Mill-Frege view, which has two constitutive theses. According to the first, the Millian one, “proper names contribute their referents to the contents of the primary speech acts they help to perform, and are thus rigid designators” (García-Carpintero 2017: 1); according to the second thesis, the Fregean one, “proper names have metalinguistic senses, known by competent speakers on the basis of their competence, which figure in ancillary presuppositions” (García-Carpintero 2017: 1). Further on, García-Carpintero also remarks that “while a language is a general tool to be used in many different circumstances, proper names are intended to serve communicative needs concerning concrete entities circumscribed within a more or less limited spatiotemporal range” (García-Carpintero 2017: 43). Chapters which follow investigate the functions and uses of fictional names, and assume different theoretical perspectives and backgrounds, with Millianism strongly defended in the opening chapter. ← 9 | 10 →

Stefano Predelli concentrates on fictional names and talk about fiction. Predelli defends a Millian semantics for fictional names, consistent with Parsimony (‘frugal metaphysics’, i.e., the desire not to include metaphysically contentious referents) and Intuition (the idea that, say, ‘according to Alice’s Adventures in Wonderland, Alice conversed with a caterpillar’ is true). His approach, inspired by Lewis’ (1978) storytelling model, is based on the No-Name Hypothesis, i.e., the idea that ‘names from fiction’ are not (referring or empty) proper names, and that they are to be interpreted as merely fictional expressions fictionally employed by a storyteller.2 After an analysis of the un-prefixed talk (‘Alice conversed with a caterpillar’) in terms of re-telling, the author proposes a paratactic analysis of prefixed sentences (‘according to Alice’s Adventures in Wonderland, Alice conversed with a caterpillar’) grounded on the distinction between the semantic properties of a sentence and the semantic effects achieved by its use. In three appendixes which follow this chapter, Predelli addresses logical issues having to do with translation, narration, and argumentation.

Fictional names and fiction are analysed from different perspectives in the subsequent chapters. Eros Corazza and Christopher Genovesi compare and contrast the metaphorical use of proper names for existing individuals and proper names for fictional entities, Daniela Glavaničová offers a hybrid account for the semantics of fictional names, Ben Martin presents the challenge of reading fiction with reality in mind, and Alberto Voltolini focuses on non-existent intentional objects (of hallucination).

Eros Corazza and Christopher Genovesi investigate how both proper names for existing individuals and proper names for fictional entities can be used metaphorically, even when they appear in argument position. According to the authors, whether a name used metaphorically is a referential or a non-referential term does not affect the proposed interpretation. This follows from the fact that when they are used metaphorically, proper names cease to be, properly speaking, proper names. i.e., they cease to be Millian tools of direct reference. When used in argument position in a simple sentence of the subject/predicate form, the tokened name used metaphorically picks up an individual analogous to the way Donnellan (1966) treats the referential uses of descriptions. That is, just as a description used referentially can pick up an individual independently of the property used to identify it, a proper name used metaphorically in argument position can pick up an individual independently of the latter carrying the name uttered. In such metaphorical uses, the name serves the function of attributing to the individual the speaker has in mind those properties that are relevant in the grasping of the ← 10 | 11 → speaker’s metaphorical meaning. Thus, a name so used loses its individuative (referential) power. Corazza and Genovesi claim that in such uses, reference is guided and fixed, like in Donnellan’s referential use of descriptions, by the individual the speaker has in mind and raises to salience in a communicative interaction. Yet, the account proposed does not undermine the Millian view that proper names are tools of direct reference nor the Fregean view that when a proper name appears in a predicative position it ceases to be a proper name. Finally, the authors accommodate both views and address the question as to why the speaker would want to utter a metaphor in this way, stressing the speaker’s primary intention, which is to convey an attitude toward the intended referent.

Daniela Glavaničová starts with observing that according to the most common view, proper names are expressions that directly refer to individuals named by them. Hence, if proper names are directly referential, and fictional names are genuine proper names, then fictional names should be directly referential too. But, while it is quite straightforward to find the referents for standard proper names of people, it is far from clear how it can be done in the case of fictional names. She remarks that this issue leads from the semantics of fictional names to the metaphysics of fictional characters, and subsequently compares two contemporary approaches to the existence of fictional names. On the one hand, realists suggest that there are such entities, but they are nonconcrete, nonactual or non-existent. On the other hand, antirealists avoid this assumption by suggesting that fictional discourse is not to be taken at face value. Both approaches face some serious troubles, and Glavaničová proposes in her contribution a hybrid account that combines features of realism with features of antirealism. In particular, the semantic distinction between de dicto and de re is employed, and the resulting view suggests de dicto (role) realism and de re antirealism.

As observed by Ben Martin, whereas the interactions with the actual world play a substantial role in the interactions with fictions, the truths of a fiction extend far beyond those stated in its text. To infer these further truths, it is necessary to rely upon a host of previously acquired background assumptions. Establishing what exactly these assumptions are, and how they help extend the truths of a fiction, however, is not a simple matter. Stacie Friend (2017) has recently argued that fictions fundamentally rely upon the actual world for their content, appealing to evidence from the cognitive sciences. According to her Reality Assumption, every proposition which is actually true is also fictionally true, unless otherwise excluded by the fiction. Martin’s chapter challenges Friend’s Reality Assumption, arguing both that the empirical evidence advanced in its favour offers inadequate support, and that the Assumption in its present form inadequately serves its intended function. ← 11 | 12 →

Alberto Voltolini investigates the possibility of referring to and quantifying over non-existent intentional objects of hallucination. He observes that if we refer to and truly quantify over non-existent intentional ordinary objects of hallucination and of thoughts in general, this is because they are full-fledged entities just as existent intentional objects of perception and of other mental states in general. In such a case, metaphysically speaking, both existent and non-existent intentionalia are concrete objects, i.e., objects that may exist; while ontologically speaking, it is false parsimony to allow for concrete existent objects and to reject concrete non-existent objects. Otherwise, either such intentional objects do figure in our overall ontological domain but as entities of phenomenologically unexpected metaphysical kinds, or even better, there really are no such intentionalia, but we just pretend that there are such things, even if our thoughts really fail to be about them and have instead merely proposition-like intentional contents. As a result, Voltolini’s proposal not only is partially ontologically committed to intentional objects, but also it re-evaluates fictionalist treatments of the matter, yet again just partially: just in some cases, it is merely fictionally that we refer to and quantify over such objects.

Investigation into the status of objects of inquiry in philosophy of language and literature is strictly connected with the notions of convention and invention, and with different aspects of interpretation. As pessimistically claimed by Davidson (1986: 446): “We must give up the idea of a clearly defined shared structure which language-users acquire and then apply to cases. And we should try again to say how convention in any important sense is involved in language”. Compare this with Thomas Pavel’s (not necessarily that pessimistic) claim that “Works of fiction more or less dramatically combine incompatible world-structures, play with the impossible, and incessantly speak about the unspeakable” (Pavel 1986: 62). The next chapters concentrate on different aspects of both representation and interpretation in philosophy of language and literature, and the authors frequently take issue with Davidson’s ideas. Richmond Kwesi discusses metaphor, truth, and representation, whereas Göran Rossholm concentrates on chance, explanation and interpretation. Marga Reimer sheds light on ‘sensible nonsense’, C.M.K. Djordjevic follows with an inquiry into the (possible) meaning of Joyce’s Finnegans Wake, and finally, in the closing essay, Bohumil Fořt focuses on characters in Kafka’s two novels, The Trial and The Castle, using some analytic tools from fictional worlds theory.

Richmond Kwesi examines whether metaphors can be used to make truth-claims; i.e., whether metaphors can be regarded as assertions that can be evaluated as true or false. Some researchers have argued for a negative answer to the ← 12 | 13 → above-mentioned questions. They have claimed, among others, that metaphorical utterances are non-descriptive uses of language, that truth is not the constitutive aim of metaphors, that metaphorical sentences do not have propositional contents, and that metaphorical utterances are neither assertions nor expressions of beliefs. Kwesi discusses a particular view, dubbed by him Metaphorical Expressivism, which exploits the relationship between truth, belief and assertion, and argues for the irrelevancy of truth to metaphors on the premise that metaphorical utterances do not count as assertions and that they do not count as the expression of beliefs. The denial of the truth-evaluability of metaphors on this view is a product of an unmotivated tendency to see truth and meaning in terms of the portrayal of facts and a commitment to two untenable principles: literalism and representationalism.

Göran Rossholm is concerned with chance, explanation and interpretation, and he presents two contrasting theses. According to the first one, literary works generated by chance cannot be given interpretations which reflect the causal background (for instance, authorial intentions) of the work. On the other hand, according to the second thesis, almost all literary interpretations are set in an explanative perspective. These two theses are not contradictory if it is assumed that literature generated by chance constitutes a very small part of the class of literary works. The second thesis does not entail the claim that interpretations explain anything, only that interpretations are fit to be integrated into causal explanations that mirror the contents of the interpretations. The key to this fitness is the concept of simplicity as used in the philosophy of science. However, Rossholm argues that in general, interpreters of literature try to apply their readings to circumstances outside the explanative perspective. Thus, even if literature made by chance is not considered, interpreting literature involves more than explanation.

Marga Reimer analyses different aspects of convention and invention. She investigates ‘sensible nonsense’ against the work of Donald Davidson, and especially his “A Nice Derangement of Epitaphs” (1986). In that paper Davidson famously appeals to malapropisms in inveighing against theories of language and linguistic communication that make essential appeal to the notion of convention. In doing so, he makes brief reference to Lewis Carroll’s nonsense poem Jabberwocky (1871) suggesting that its relatively “easy”, if partial, interpretability lends further credence to his concerns with such theories. Reimer argues that Carroll’s poem, and nonsense literature more generally, while highlighting the potential for inventive uses of language, also illustrate the unequivocal (if “merely” factual) dependence of linguistic communication, however inventive, on convention. She also suggests that it is more fruitful to view linguistic communication as involving the spectrum ← 13 | 14 → phenomena of convention and invention. Thus, according to this approach, all linguistic utterances can be analyzed in terms of the complementary notions of convention and invention, thereby accommodating the degreed nature of communicative success. Such an approach would have the methodological advantage of allowing for the theoretical possibility of utterances involving 0 degrees of convention – and likewise of utterances involving 0 degrees of invention. In this way, it is possible to accommodate the obvious and theoretically important differences between the kind of communicative success associated with the highly conventional writing of The Wall Street Journal and the kind of dismal communicative failure associated with the highly inventive writing of James Joyce’s Finnegans Wake.

Whereas Reimer refers to Finnegans Wake as a case of (putative) “dismal communicative failure”, C.M.K. Djordjevic inquiries into the book’s possible meaning. He reflects on texts often considered as “abnormal”, and argues that limiting the scope of linguistic and philosophical inquiry to non-problematic cases restricts and may distort language, meaning, and understanding in important ways. Hence, more energy should be devoted to “abnormal” cases and Djordjevic concentrates on a highly abnormal case, i.e., James Joyce’s Finnegans Wake. He considers several arguments from thinkers – ranging from literary critics to philosophers – about the (alleged) meaning of Finnegans Wake, and next puts forward three tests that can determine if some x should be considered meaningful. Djordjevic further argues that Finnegans Wake passes all three. He also demonstrates that neither a referential nor an intentionalist theory of semantic meaning can coherently and cogently explain how Finnegans Wake manages to mean anything, and finally he considers some interesting avenues for possible research that may be occluded by a focus on more standard cases of texts.

Bohumil Fořt examines two particular protagonists of Franz Kafka’s novels, The Trial and The Castle. Josef K. and K. are viewed primarily through their connections with the worlds they inhabit and with the actions in which they participate. In his analysis and classification, Fořt makes use of specific tools primarily borrowed from fictional worlds theory (influenced especially by Doležel 1998 and Pavel 1986), and also from other theoretical contexts, and he pays special attention to the issues of the typology of fictional worlds, the concept of narrative modalities, the concept of gaps, and motivation of the protagonists. He concludes his study with discussing the role of the specific qualities of the protagonists in the readers’ conceptualization of the novels in question.

It is hoped that studies collected in this volume clearly demonstrate the need for further investigation into possible common objects of inquiry in philosophy of ← 14 | 15 → language and literature, with fictional objects, the status of fiction, interpretation and nonsense being very promising fields of investigation.


I wish to thank all the contributors to this volume for their participation in the project, and for their eagerness to peer-review other texts. I am also grateful to the following external reviewers for their most helpful comments and suggestions: Kacper Bartczak (University of Łódź, Poland), Stacie Friend (Birkbeck College, University of London, UK), Wojciech Kalaga (University of Silesia, Poland), Krzysztof Kosecki (University of Łódź, Poland), Anna Kędra-Kardela (UMCS Lublin, Poland), Jakub Mácha (Masaryk University, Brno, Czech Republic), Anders Pettersson (Umeå University, Sweden), and Marián Zouhar (Comenius University, Bratislava, Slovakia).


Davidson, D. (1986). A Nice Derangement of Epitaphs. In: E. Lepore (ed.), Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson. Oxford: Basil Blackwell, 433–446.

Doležel, L. (1998). Heterocosmica: Fiction and Possible Worlds. Baltimore & London: John Hopkins University Press.

Donnellan, K. (1966). Reference and Definite Descriptions. The Philosophical Review 75: 281–304. Reprinted in K. Donnellan (2012). Essays on Reference, Language, and Mind. New York: Oxford University Press, 3–30.

Friend, S. (2017). The Real Foundation of Fictional Worlds. Australasian Journal of Philosophy 95(1): 29–42.

García-Carpintero, M. (2017). The Mill-Frege Theory of Proper Names. Mind, https://doi.org/10.1093/mind/fzx010. Accessed 10.01.2018.

Lewis, D. (1978). Truth in Fiction. American Philosophical Quarterly 15: 37–46.

Pavel, T. G. (1986). Fictional Worlds. Cambridge, MA: Harvard University Press.

Predelli, S. (2017). Proper Names. A Millian Account. Oxford: Oxford University Press.

Stalmaszczyk, P. (ed.) (2016). Philosophical and Linguistic Analyses of Reference (Studies in Philosophy of Language and Linguistics 2). Frankfurt am Main: Peter Lang.

Stalmaszczyk, P. and Fernández Moreno, L. (eds.) (2016). Philosophical Approaches to Proper Names (Studies in Philosophy of Language and Linguistics 1). Frankfurt am Main, etc.: Peter Lang.

1 See the contributions in Stalmaszczyk and Fernández Moreno (eds.) (2016), and Stalmaszczyk (ed.) (2016), respectively.

2 This hypothesis is comprehensively developed in Predelli (2017: Ch. 8).

| 17 →

Stefano Predelli

University of Nottingham, UK

Millians in Wonderland: Fictional Names and Talk About Fiction

Abstract: This essay defends a Millian semantics for so-called fictional names, consistent with Parsimony (the desire not to include metaphysically contentious referents) and Intuition (the idea that, say, ‘according to Alice’s Adventures in Wonderland, Alice conversed with a caterpillar’ is true). My approach is grounded on the No-Name Hypothesis, i.e., the idea that ‘names from fiction’ are not (referring or empty) proper names, and that they are to be interpreted as merely fictional expressions fictionally employed by a storyteller. After an analysis of our un-prefixed talk (‘Alice conversed with a caterpillar’) in terms of retelling, I propose a paratactic analysis of prefixed sentences (‘according to Alice’s Adventures in Wonderland, Alice conversed with a caterpillar’) grounded on the distinction between the semantic properties of a sentence and the semantic effects achieved by its use. Three appendixes address logical issues having to do with translation, narration, and argumentation.

Keywords: fiction, semantics, proper names, proposition, parsimony, Millianism, David Lewis

1 Millians in Wonderland: Fictional Names and Talk About Fiction

In this essay, I discuss sentences such as

(1) according to Alice’s Adventures in Wonderland, Alice conversed with a caterpillar


(2) according to Alice’s Adventures in Wonderland, Dinah conversed with a caterpillar.

My aim is to defend an analysis consistent with (i) a Millian theory of names (Millianism), (ii) a frugal metaphysics (Parsimony), and (iii) the intuitive desire that utterances of (1) and (2) be evaluated as, respectively, true and false (Intuition).

Intuition strikes me as a plausible pre-theoretic demand, with the possible exception of its theory-laden insistence on utterances, to be discussed and motivated as I proceed. Parsimony, roughly the invitation not to accept metaphysically disagreeable objects, is here given a straightforwardly semantic interpretation (among other things, so as to eschew an explanation about what sort of presumed entities I regard ← 17 | 18 → as off limits): no analysis satisfies Parsimony, so I assume, as long as it approaches the occurrences of ‘Alice’ and ‘Dinah’ in (1) and (2) as referring proper names.

Millianism is the central character in my study. It encompasses a family of familiar views of names, committed to the idea that the contribution of a name to the truth-conditions of sentences in which it occurs is exhausted by its referent (if any). A few minutiae in my understanding of Millianism are occasionally addressed later on, often (especially if not directly relevant for my topic) in the footnotes. At this preliminary stage, it suffices to note that nothing in what follows aims at providing direct evidence for any particular version of Millianism. Still, fictional talk is generally cited as a stumbling block for non-descriptivist theories of names – a phenomenon that allegedly saddles Millianism with the rejection of Intuition or Parsimony. My presentation of an Intuitive Parsimonious Millian treatment of (1) and (2) thus indirectly supports the Millian cause, in the sense of making it palatable to those sensitive to the desire that utterances of (1) and (2) be evaluated as true and false, without thereby burdening ‘Alice’ and ‘Dinah’ with undesirable referential commitments.1

In section two, I begin my analysis with the introduction of my guiding suggestion: the No-Name Hypothesis. This hypothesis pursues a Parsimonious stance alternative to a traditional parsimonious attitude, the notion that ‘Alice’ and ‘Dinah’ are empty names. Since Parsimony forbids referentially successful analyses of these expressions, it follows that, according to the No-Name Hypothesis, the occurrences of ‘Alice’ and ‘Dinah’ in (1) and (2) are not occurrences of (referring or empty) names, in some semantically relevant sense of ‘name’ yet to be explained.

In sections three and four I leave (1) and (2) on the backburner, and I develop a picture of fiction-making and so-called talk about fiction grounded on a Lewis-inspired storytelling model, modified according to the guidelines of the No-Name Hypothesis. The resulting conclusions, pertaining to cases such as

(3) Alice conversed with a caterpillar

(4) Dinah conversed with a caterpillar ← 18 | 19 →

are then put to use in section five, where I present an analysis of (1) and (2) consistent with Parsimony, Millianism, and Intuition: the Paratactic Analysis. In three appendixes, I pause of a few semantically interesting consequences of the Paratactic analysis, having to do with what I call the ‘narrative periphery’, with translation, and with issues of logical consequence.

2 The No-Name Hypothesis

What is the problem with ‘Alice’ in (1)? I take it to be obvious that nothing of significance has to do with the graphic (or phonetic) type ‘Alice’ – say, with the sheer concatenation of those five letters. No progress is made by moving from this typographic approach to a more demanding understanding – say, to a conception of ‘Alice’ as an expression governed by the syntactic regularities for simple noun phrases, and/or by the custom to employ it as a name for (typically) female individuals. For, surely, that name-type may be used so as to refer to author Alice Munro or pianist Alice Coltrane (or, for that matter, to Carroll’s acquaintance Alice Liddell), with no embarrassment for Millianism, and with no interesting repercussions for fictional discourse.2

It is relatively uncontroversial that, Millianism or no Millianism, the bearers of the referential burden are not name-types simpliciter, but items of a more fine-grained nature – uses of a name-type on a certain occasion, tokens of a name-type in a given context, disambiguations of a name-type on the basis of this or that pragmatic consideration, or something of that sort.3 In particular, according to a metasemantic picture dear to the Millians, names, in any semantically interesting understanding of ‘name’, are individuated according to their position within appropriate chains of events, eventually connecting them to this or that ‘baptism’ affair. As a result, my tokens of ‘Alice’ conform to the semantic pattern for, say, Munro’s name, without having anything to do with Coltrane or anyone else, by virtue of bearing a suitable relationship with an episode involving that woman – perhaps, in this case, a baptism in the literal sense of the term. ← 19 | 20 →

Let me call this view the Chain Picture of names.4 Admittedly, what my meagre hints manage to paint is, at best, only a preliminary sketch. But forget about the details in the Chain Picture’s content, and focus on its form, that of a theory about the individuation of the items with which semantic evaluation is concerned. Looking at the issue from the viewpoint of the initial link, David Kaplan puts the matter in terms of dubbings: “I introduce[d] the notion of a dubbing, for what I took to be the standard form of introduction of a proper name word. … Dubbings create words” (Kaplan 1977: 560–561).

As an analysis of some pre-theoretical sense of ‘word’ or ‘name’, Kaplan’s take on the Chain Picture may well be debatable, and his idea of a dubbing may not be to everyone’s taste.5 What interests me here, however, has to do with Kaplan’s metasemantic (or, as some put it, pre-semantic) emphasis: dubbings ‘create words’, or, at the very least, they determine ‘what words, if any, were spoken’ (Kaplan 1977: 562), in the sense that they identify the bearers of the referential burden.6 They identify, in particular, names, in a semantically suitable sense of the term: not promiscuous name-types, but constructs put to use as referential, truth-conditionally relevant devices.

The Chain Picture may well be consistent with referent-deprived names, that is, with so-called ‘empty’ names. Take your diligent engagement in a dubbing ceremony, with all the needed bells and whistles: let this be called ‘Alice’, you declare. But you hallucinate, and your demonstrative fails to secure a referent. Then, perhaps, the semantic analysis of our conversations will need to be equipped with the resources required by emptiness: on occasions where my use of ‘Alice’ is related to your dubbing, my utterances end up being false or truth-valueless.7

It may seem that (hallucinations aside) Parsimony requires a parallel emptiness-oriented explanation of our interaction with ‘Alice’ and ‘Dinah’, as they matter for discourse about Alice. For, so the story goes, any attempt at identifying a referent ends up with personae non gratae, or at least with controversial entities – merely possible girls, non-existent felines, abstract artefacts, or something of that sort. Leaving aside whether these presumed objects are indeed independently undesirable, I happily subscribe to this methodological stance: as far as semantics is ← 20 | 21 → concerned, so I assume, analyses of (1) and (2) in terms of referentially successful names are incompatible with Parsimony.8

I do, however, take issue with the further conclusion that this much inevitably paves the way for an approach in terms of empty names – from the Chain Picture’s viewpoint, in terms of referentially unsuccessful dubbings. Accordingly, I proceed by exploring the logical space between metaphysical opulence and semantic emptiness, a space occupied by the hypothesis that no dubbing whatsoever governs the occurrences of ‘Alice’ and ‘Dinah’ in (1) and (2). In a Kaplan-style metasemantic framework: the hypothesis that ‘Alice’ and ‘Dinah’, as they matter for (1) and (2), are not names, in the sense of not being among the items of concern for the truth-conditional analysis of our discourse.

Paraphrasing Russell: a semantic theory may be tested by its capacity for dealing with puzzles (Russell 1905: 484) – here, by its ability to sustain a Parsimonious Millian Intuitive approach to (1) and (2). Admittedly, the No-Name Hypothesis (NNH) introduced above does not provide an obviously promising starting point: banishing ‘Alice’ and ‘Dinah’ from the semantic record hardly paves a straightforward alley towards truth for utterances of (1), and falsehood for utterances of (2). The contrary seems to be the case: even generously granting that all goes as it should when it comes to ‘conversed’ or ‘a caterpillar’, the notion that the expressions occurring in subject positions are semantically uninterpretable name-types entails the apparently discouraging conclusion that the occurrences of (1) and (2) under discussion fail to encode ful-fledged propositions.

Still, as I explain in what follows, the NNH remains consistent with a more auspicious locus for the employment of those name-types as names: the fictional niche. In the next section, I thus explore the relationships between the NNH and fiction-making, with particular attention to the occurrences of ‘Alice’ and ‘Dinah’ in Carroll’s text. I do so against the background of a Lewis-inspired storytelling approach, importantly modified and adapted along lines consistent with the NNH.

3 Fictional Tellings

For reasons independent of Parsimonious Millianism, David Lewis found it helpful to think of fiction-making in terms of ‘a story told by a storyteller on ← 21 | 22 → a particular occasion’ (Lewis 1978: 39) – in Alice’s case, in terms of the story of a curious girl who falls down the rabbit hole, told by a storyteller by means of sentences such as

(5) Alice was beginning to get very tired of sitting by her sister (Carroll 1865: 37).9

According to Lewis,

the act of storytelling occurs [in worlds where the tale is told as known fact], just as it does here at our world; but there it is what here it falsely purports to be: truth-telling about matters whereof the teller has knowledge. (Lewis 1978: 40)

Of course, from the NNH’s viewpoint, Lewis’ ‘two-fold telling’ will not do: ‘here at our world’, in the absence of a dubbing governing the occurrence of the name-type in subject position, (5) fails to encode any full-fledged propositional content, and is a fortiori not a vehicle for recounting the curious girl’s tale. In some sense, as I am about to explain, the ‘other side’ of Lewis’ divide is nevertheless a source of more promising outcomes.

‘There’ (in Lewis’ possible world parlance, ‘in the worlds where the tale is told as known fact’) everything presumably works out all right. The expressions in (5), when occurring in the fictional teller’s mouth, follow their customary English profile, and what fictionally is being told is a tale of incipient boredom and sibling-relations (see Appendix 1 for a few details in this respect). Or so I can safely assume as far as NNH is concerned: at least in the absence of arguments to the contrary, all sides in the semantic debate on fictional names take on board the unproblematic nature of occurrences of, say, ‘tired’ or ‘her sister’, when they occur within fictional narratives to no lesser extent than in actual speech.

More interestingly, when occurring in the fictional teller’s mouth, ‘Alice’ appropriately connects with the fictional girl’s dubbing, so that what is fictionally being conveyed is that she was beginning to get very tired of sitting by her sister. Fictionally, then, a tale is being told by means of (5), that is, by means of a sentence containing a straightforwardly referring proper name, and unproblematically encoding a (singular) proposition (see Appendix 2 for certain independently important caveats).

Yet, from the NNH’s perspective, this is so only fictionally.10 And, from that perspective, merely fictional dubbings, names, and propositions are of no greater ← 22 | 23 → concern for semantics than merely fictional caterpillars are for zoology or gardening.11 What, according to the NNH, is in fact encoded in the teller’s fictional employment of (5) is not propositional content of a special type. It is not propositional content at all – no more a result of our interpretive procedures or an object of our understanding than fictional caterpillars are the target for pest control agendas.

Still, our exposure to (5) and, more generally, our acquaintance with Carroll’s text undoubtedly bring to light actual propositional content – in particular, content of a different type than what would have been engendered, had we read

(6) Dinah was beginning to get very tired of sitting by her sister

instead. Since, according to the NNH, the proposition in question is not semantically encoded in the storyteller’s speech, it must be prompted by regularities other than those appropriate for the truth-conditional evaluation of the teller’s sentences.12

It is in this respect that, even with the non-Lewisian twist imposed by the NNH, the storytelling model pays its first dividends. In particular, as I explain in the remainder of this section, although no propositional content is obtainable from the fictional sentences spoken by the teller, the storytelling approach makes allowances for a different sort of content, that imparted by the occurrence of those sentence-types as part of that fictional telling – a content ensuing not from their semantic affordances, but from their being told.

Set fictional talk momentarily aside, and take my utterance of, say,

(7) Aristotle was fond of dogs. ← 23 | 24 →

On the assumption of a suitable link between that occurrence of ‘Aristotle’ and a certain Greek intellectual, (7) encodes a (singular) proposition about that man. Yet, further information is being imparted by my utterance of that sentence – together, perhaps, with a few reasonable background assumptions, such as that I utter (7) as a sentence of English. For instance, as a result of being exposed to that utterance, you are in the position of concluding that, say, I sometimes speak, I sometimes utter sentence-types beginning with a vowel, and I token ‘Aristotle’ so as to refer to a certain individual.13 In particular, focusing on impartations related to the use of the name ‘Aristotle’, my utterance imparts that the referent of the name-type ‘Aristotle’, as tokened by me on that occasion, had a certain property – that is, it imparts the propositional content encoded in ‘the referent of “Aristotle”, as that name-type occurs on this occasion, was fond of dogs’, or, more succinctly,

(8) the referent of ‘Aristotle’ was fond of dogs.

By the same token, by virtue of being exposed to the fictional storyteller’s employment of (5), we are in the position of recovering the propositional content it fictionally imparts, to the effect that (among other things)

(9) the referent of ‘Alice’ was beginning to get very tired of sitting by her sister.

This, of course, is not the proposition encoded in (5): in general, because (5), not unlike (7), is clearly not truth-conditionally equivalent with any quotational sentence; and in particular because, once the NNH is taken on board, (5) does not semantically encode any (metalinguistic or non-metalinguistic) propositional content.14 Still, imparted content is content all right: in this case, a (singular, given standard views about quotation) proposition about a name-type, to the effect that its referent was entering a certain state of monotony.

A word of clarification is in order. In any parsimonious approach to matter of fictions, Alice’s storyteller is a fictional teller, in the sense of being merely fictionally a teller. In reality, no such thing is in place: the teller who recounts that curious girl’s adventures is a figment of Carroll’s imagination to no lesser extent than the caterpillar she encounters. As a result, his (or her?) telling, encoding, and, crucially, ← 24 | 25 → imparting, is merely fictional: nobody, in reality, imparts anything about a referent of ‘Alice’ and her ensuing boredom. Still, although the imparting is merely fictional, what is imparted is not: (9) straightforwardly encodes content that we can actually understand, entertain, and convey. There actually exists no content which that teller fictionally encodes by means of (5). But there actually exists a content C such that, fictionally, a teller imparts C by means of fictionally uttering (5). It is that content (and, of course, much more besides) which we actually absorb (and entertain, imagine, or what have you) when exposed to Carroll’s tale.15

As I explain in the next section, it is this sort of content that provides the subject-matter for our understanding of Carroll’s novel – that, as I put it in the next section, grounds the notion of a ‘storyworld’, and thereby provides a point of comparison with what is commonly known as ‘talk about fiction’.

4 Retellings

According to the storytelling model, Alice is to be approached in terms of ‘a story told by a storyteller’. The teller’s tale, as any other telling, will need to be assessed and evaluated as a (fictional) description of events, among other things on the basis of hypotheses about the teller’s (fictional) seriousness and trustworthiness. Since, as far as I can tell, nothing supports the hypothesis that Carroll’s teller is ironic or unreliable (at least when it comes to the curious girl’s weariness), his utterance of (5) may be taken at face value – it may, in a popular jargon, be accepted as a description of Alice’s storyworld.

According to the NNH’s take on the storytelling model, it is futile to investigate Alice’s storyworld by semantically interpreting the storyteller’s expressions: although, fictionally, the curious girl’s incidents are described in accordance with ← 25 | 26 → the content of the teller’s sentences, they are so only fictionally. ‘Here at our world’, Alice’s storyworld is rather one reflected by the teller’s imparted content: a set of circumstances in which the referent of ‘Alice’ was getting tired to sit by her sister, but the referent of ‘Dinah’ was not (side by side, presumably, with further information derivable by the usual processes of importation, such as that the referent of ‘Alice’ was human and that the referent of ‘Dinah’ was smaller than Jupiter).16

This content is, as I put it above, content all right, actually cognitively accessible and describable by means of expressions in our language. In this sense, Alice’s storyworld also provides an actual normative benchmark for our talk, that is, for what is commonly called ‘talk about fiction’: as the vernacular puts it, some of our comments get Alice’s story right, but some do not. Echoing section three, in what follows I propose an analysis of talk about Alice consistent with the NNH, and inspired by the storytelling model.17

Suppose that, in the right setting (say, during a conversation about Victorian literature, with the intention of discussing Alice, or something of that sort), I say

(3) Alice conversed with a caterpillar.18

According to the NNH, I, like Carroll, cannot use this sentence-type as a sentence, since ‘Alice’, the name, is no more at my disposal than it was at his. Still, not unlike Carroll, I can (in the right setting) make it fictional that someone employs ‘Alice’ as a name, and utters (3) so as to convey that the individual it picks out chatted with an insect. Just as Carroll made it fictional that someone told the story of the curious girl who falls down the rabbit hole, then, I can make it fictional that someone retells that story.19 ← 26 | 27 →

Fictional (and, for that matter, actual) tellings need not be retellings: the process of retelling your story, rather than making up my own, must presumably be characterized by some appropriate concern for what you did, with what the vernacular describes as ‘telling the same story’. I here accept without further ado what strikes me as a condition for retelling Alice’s story, that is, for Alice-retelling: the presence of some appropriate worldly relationship between my retelling and its original telling. (Presumably, this condition will end up having at least something to do with my intentions and with my acquaintance with Carroll’s text – here, I can afford not to dwell on this issue because of its uncontroversial independence from Parsimonious Millian premises).

Retellings engender at least two semantically interesting requirements, having to do with truth and co-reference – in the case of talk about fiction, fictional truth and co-reference. As for the first, Alice-retellings fictionally get things right or wrong: a reteller fictionally committed to (3) truly describes Alice’s storyworld, but s/he would not, were s/he to opt for

(4) Dinah conversed with a caterpillar

instead. Part of the reason why the (fictional) proposition encoded in (3) is (fictionally) true has to do with a presumption of (fictional) co-reference: my teller’s use of ‘Alice’ on Alice-retelling occasions conforms to that of Carroll’s storyteller. In the Chain Picture’s jargon: the (fictional) token of ‘Alice’ in the reteller’s mouth partakes in a reference-preserving chain eventually leading to the original teller’s tokens, and hence, in turn, to a (fictional) dubbing episode involving a certain curious girl (see Appendix 2 for a few additional details).

Here as before, merely fictional truths and reference-preserving relations are not a kind of truths and reference-preserving relations: they are not truths and reference-preserving relations at all. Still, here as before, what is not merely fictional is the proposition fictionally imparted by the retelling episode – in the cases of (3) and (4), the propositions that

(10) the referent of ‘Alice’ conversed with a caterpillar


(11) the referent of ‘Dinah’ conversed with a caterpillar. ← 27 | 28 →

When fictionally imparted by a retelling, then, these propositions may straightforwardly be assessed in terms of their conformity to the storyworld fictionally depicted by the original telling episode. The proposition in (10), in particular, turns out to be true with respect to Alice’s storyworld, whereas (11) does not. I say that (3), the sentence-type, accords with Alice, in the sense that an Alice-retelling in which it occurs imparts a proposition true with respect to Alice’s storyworld. (4), then does not accord with Alice, since the proposition in (11) inaccurately describes how things went with the curious girl’s cat.

Accordance is not truth: according to the NNH, ‘here at our world’ (3) does not encode any propositional content, and is a fortiori simply not truth-apt. What, given my adherence to Intuition, must turn out true or false are on the other hand utterances of expressions seemingly involving embedded occurrences of (3) and (4), such as, in particular, (1) and (2).20 I thus conclude my presentation of a Parsimonious Millian Intuitive stance with the explanation of how, given the aforementioned contrasting Alice-accordance results for (3) and (4), utterances of (1) are true, but utterances of (2) are not.

5 The Paratactic Analysis

The superficial feature that distinguishes (1) from (2), namely the occurrence of ‘Dinah’ instead of ‘Alice’, is, according to the NNH, semantically idle – at least in the sense that no distinct truth-conditions may be derived from the interpretation of those name-types. As a result, if anything in (1) and (2) is at all truth-conditionally approachable, it must be one and the same evaluable syntactic material.21 It follows that any approach to utterances of (1) and (2) consistent with both Intuition and ← 28 | 29 → the NNH must manage to obtain contrasting outcomes in the face of indistinguishable analysanda. In this respect, indexicality provides a promising model: utterances of the same expression may be endowed of different semantic properties, as long as some indexical component is evaluated with respect to different contexts. As in, say, the contrast between true and false utterances of ‘that is blue’ given different demonstrata.

An indexical component in (1) and (2) can be of help when it comes to Intuition only if it is sensitive to the discrepancy between these sentence-types – namely, to the very distinction which the NNH expelled from semantic evaluation. For this reason, the distinctive feature of (1) and (2), the occurrence of ‘Alice’ instead of ‘Dinah’, must be recorded at the contextual level, as a parameter that eventually makes a difference for truth-conditional evaluation. Everybody familiar with Davidson’s ‘paratactic’ analysis of quotation, indirect speech, and other phenomena, is now in the position of anticipating the target of this argumentative line.

According to the Paratactic Analysis, (1) is ‘of the form’

(12) that accords with Alice. Alice conversed with a caterpillar.22

In this two-fold layout, the truth-conditional burden lies entirely on the expression on the left-hand side, ‘that accords with Alice’. Truth is thus obtainable for (1) (with respect to a context c) iff the object demonstrated in c accords with Alice (substitute your favourite theory of demonstratives if you dislike any of the details here). Similarly, (2) is analysed as

(13) that accords with Alice. Dinah conversed with a caterpillar,

a construct which, for the aforementioned reasons, ends up being truth-conditionally indistinguishable from (12). As a result (unsurprisingly, given my commitment to the NNH), the Paratactic Analysis renders a verdict of equivalence for (1) and (2): the former is true with respect to a context c iff the latter is.23

It does however not follow that utterances of (1) and (2) share their truth-value, for the same reason why utterances of the uncontroversially equivalent ‘that is ← 29 | 30 → blue’ and ‘blue is what that is’ may differ in truth-value in different contexts. In particular, given that the material on the right-hand side in (12) and (13) provides the desired demonstrata, it follows that utterances of (1) end up being true, since ‘Alice conversed with a caterpillar’, i.e., (3), accords with that novel. Yet, in conformity with Intuition, utterances of (2) turn out false, since ‘Dinah conversed with a caterpillar’, i.e., (4), does not.

Relegating to three appendixes a few details of the Paratactic Analysis and of the background from which it stems, this wraps up my presentation of a Parsimonious Millian Intuitive analysis of (1) and (2). My approach straightforwardly follows Intuition, since it recognizes how, given how things go with Alice, utterances of (1) are true, but utterances of (2) are not. It also proceeds from squarely Parsimonious Millian premises – in particular, from the sort of premises imposed by the version of Parsimonious Millianism from section two: the No-Name Hypothesis.

My conclusions emerge against the background of a few additional hypotheses about fiction-making, storytelling, and retelling. Much in this respect deserves further attention: the conditions for storytelling, for instance, and for the possibly peculiar type of storytelling emerging from literary works; the background for retelling, and, in particular, for F-retelling, given a fiction F; the notions of a storyteller’s reliability, seriousness, and trustworthiness; and, in general, a variety of questions from the philosophy of literature, from pragmatics, and from wider approaches to communication. Yet, all of these details need to be assessed for independent reasons, having to do, in general, with the independently interesting ideas of storytelling and fiction. In particular, then, the outcomes of their analyses remain independent from Parsimonious Millian premises, and, a fortiori, from the NNH I proposed in section two.24

More directly semantically relevant are issues having to do with the extensions of my approach to other phenomena in the vicinity of fictional talk. Two notorious sources of befuddlement deserve at least to be flagged as tangential to my topic in this essay: negative existentials and so-called ‘meta-narrative’ instances, such as ‘Alice is a fictional character’ or ‘Alice has acquired a cultish following’.25 Echoing Lewis: “I shall have nothing to say here about the proper treatment of ← 30 | 31 → these sentences” (Lewis 1978: 38). Off the record, my reason for this indifference has to do with my conviction that the issues they bring to light are not problematic for Millianism alone – a conviction bolstered by the fact that, as far as I can tell, descriptivist manoeuvres are hardly in an obviously better position in those respects. Officially, I rest satisfied by noting that, if the Paratactic Analysis in on the right track, it ought to be welcomed as the dissolution of a source of befuddlement for Parsimonious Millians, independently of its inability to deal with other semantically troublesome aspects of fiction. This motivation would admittedly ring hollow if it could be argued that, though in the position of dealing with (1) and (2), my approach precludes any satisfactory treatment of negative existentials, meta-narrative speech, or anything in that neighbourhood. Yet, as far as this sort of argument goes, I am happy to leave the burden of proof on my opponents’ shoulders.

Appendix 1: Translations from the Periphery

According to Lewis, the storytelling approach is “most apparent when the fiction is told in the first person” (Lewis 1978: 40), as in Conan Doyle’s detective stories. It is a peculiarity of the Holmes’ stories that, fictionally, the storyteller inhabits the storyworld he describes: Watson, the narrator in A Study in Scarlet (Doyle 1887), roams the streets of London side by side with the other individuals whose tale he recounts. This is not true for all storytelling episodes, Alice being a case in point: in her journey through Wonderland, the curious girl encounters smoking caterpillars and mentally unstable hat-makers, but no storyteller intent in describing her adventures.

I say that Alice’s storyteller inhabits a narrative periphery disjoint from the storyworld s/he describes (I conceive of cases such as A Study in Scarlet as limiting overlapping instances, where the narrator also figures within the storyworld’s domain). The periphery-storyworld divide is motivated by considerations utterly independent of Parsimonious Millianism – think of name-deprived first-person tellings about prehistoric or post-apocalyptic landscapes, with no one in sight in the position of describing the events in question.26 Still, as I explain below, that ← 31 | 32 → divide also yields a few results of interest from the viewpoint of the NNH’s take on the storytelling model.

Given my focus on names, I have remained relatively casual when it came to, say, ‘sister’ or ‘caterpillar’ in (1), (3), or (5). Still, if Alice’s tale has any hope of being understood, the (fictional) occurrences of those expression-types had better be interpretable in the obvious fashion – in this case, as expressions from the author’s (and, I suppose, his primary audience’s) mother tongue.27 In a nutshell, then: Alice’s (and Scarlet’s) peripheries are inhabited by English-speaking tellers, who employ ‘sister’ or ‘caterpillar’ to speak of female siblings and insects, just as Carroll did and as we continue doing.

Incidentally, this is so also in cases of overlapping narratives, independently of the teller’s preferred tongue ‘within the storyworld’. So, Doyle might have written his tales in German, thereby fictionally engendering a German-speaking narrator who describes his own exploits as a (possibly monolingual) English-speaking sidekick.28 Our German-speaking acquaintances, in fact, routinely do so when retelling Scarlet or Alice by means of, say,

(14) Alice hat mit einer Raupe geschprochen.29

Since (14) happens to accord with Alice, it follows that our well-informed German acquaintances also subscribe to the German counterpart of (1), namely

(15) zufolge Alice, Alice hat mit einer Raupe geschprochen,

a sentence they rightly deem to be truly utterable.

Note that, according to the Paratactic Analysis, utterances of (1) and (15) end up with contents as semantically distinct as those for ‘that is blue’ accompanied by different demonstrations: utterances of (15), for instance, have to do with the sentence-type (14), but utterances of (1) do not. Still, in the approach from sections three and four, (14) accords with Alice iff (3) does – in either case, iff ← 32 | 33 → Alice-fictionally the referent of ‘Alice’ conversed with a caterpillar. As a result, the Paratactic Analysis, not unlike any reflexive-demonstrative analysis, remains incompatible with the idea that a sentence and its translation inevitably encode the same propositional content. But, like any reflexive-demonstrative analysis worth its salt, it reflects appropriate intuitions of truth-preservation, such as the notion that utterances of (15) are true iff utterances of its English translation, (1), are.

Appendix 2: Names in the Periphery

A variety of well-known prima facie puzzles concerning fiction and storytelling reverberate within the periphery-storytelling divide. For instance, as Walton observes, tellers sometimes “tell of events they could not possibly know about” (Walton 1990: 177), as in reports of private conversations or of unspoken thoughts. Walton lists the quandary of the knowledgeable teller among what he calls ‘silly questions’ – queries assumed to be unworthy of attention by anyone acquainted with the practice of storytelling.

From the semantic viewpoint, a parallel silly question emerges from the Chain Picture’s take on the storytelling model: overlapping instances aside, how can an inhabitant of the periphery bear any suitable relation to a dubbing events which, like the curious girl’s baptism, occur within the storyworld? This is a futile query: it is an ineliminable condition for our participation in storytelling practices that the periphery be understood as, to coin a phrase, referentially permeable. Though unable to kick the curious girl or to challenge her in a game of croquet, then, Alice’s teller is assumed to be in the position of partaking in the right kind of relation with that individual and her dubbing.

Similar considerations apply to cases where the teller’s responsibility extends to the very act of dubbing (in the Kaplan-inspired terminology from section two, to the introduction of a name). Consider my short Tale of the Unnamed Child, here reproduced in unabridged form:

According to the approach I pursued thus far, it is Tale-fictional that ‘Noname’ refers to a nameless offspring of hesitant parents. Still, Noname’ is not a name in Tale’s storyworld: if dubbings do indeed create words, ‘Noname’, the fictional child’s fictional name, is a denizen of the periphery, and of the periphery alone.

The case of Tale thus indicates the need for a refinement of the relatively simple-minded picture I took on board in the main body of this paper, mostly for reasons of pedagogical transparency. Recall that, in my view, a simple sentence such as ← 33 | 34 →

(16) Noname led a happy life

accords with the relevant scenario, in this case Tale, iff the proposition fictionally imparted by its occurrence is true in Tale’s storyworld. According to the rough account of imparted content in section four, that proposition is encoded in

(17) the referent of ‘Noname’ led a happy life.

The foregoing considerations about the periphery-storyworld divide force an important amendment to this account: for (16) is surely in accordance with Tale, but the proposition in (17), namely the claim that someone who served as the referent for ‘Noname’ led a happy life, is not true with respect to Tale’s storyworld.

The point here is not peculiar to fictional talk, and is the result of independent considerations about imparted content and the sort of ‘scoping-out’ effects it achieves. Consider an utterance of

(18) if no language ever developed, Aristotle would not have written about metaphysics.

This utterance imparts that the referent of ‘Aristotle’, as that name occurs in the speaker’s mouth, would not have produced his work on first philosophy in the circumstances described by the antecedent in (18). By the same token, the teller’s (fictional) employment of (16) imparts that the referent of ‘Noname’, as that name occurs in the periphery, is such that it led a happy life in Tale’s storyworld. This much, as desired, supports the intuitive conclusion that (16) accords with Tale.

Appendix 3: A Logic

According to Lewis, “truth in a given fiction is closed under implication” (Lewis 1978: 39), in the sense that if S1 … Sn-1, thus Sn is valid, so is ‘according to F S1 … according to F Sn-1, thus according to F Sn.

This presumably desirable outcome is trivially compatible with the Paratactic Analysis, since, according to it, the latter argument is valid on any choice of S1 … Sn.30 A compatibility of this sort is meagre consolation: intuitively valid arguments are recognized as valid only because all suitably ‘pre-fixed’ arguments turn out to involve equivalent premises and conclusions. Once again, though, this predicament reflects a familiar outcome for indexical scenarios – such as the notion that, say, ‘I am hungry’ entails ‘it is not the case that I am not hungry’, even though it may be the case that the former is uttered truly (by me), and the latter falsely ← 34 | 35 → (by you). Just as the incompatibility between our utterances is not reflected at the level of those sentences’ truth-at-a-context, so it must be the case that the intuitive properties of Lewis-style arguments are to be recorded from the viewpoint of the truth of sentences-at-a-context, i.e., of utterances.


(19) according to Alice, either Alice conversed with a caterpillar or the moon is made of cheese


(20) according to Alice, Alice conversed with a duck.

Intuitively, the ‘right sort’ of relationship holds between (1) and (19), but not between (1) and (20). According to the Paratactic Analysis, a relationship of this kind is not unveiled by considerations of entailment, in the classic sense of the term: (1) entails both (19) and (20), in the sense that both are true with respect to a context c if (1) is. Still, an utterance of (1) is true iff (3) accords with Alice, that is, iff the content imparted by (3) is true in Alice’s storyworld. Thus (at least as long as that storyworld is consistent), the content imparted by

(21) either Alice conversed with a caterpillar or the moon is made of cheese

also accords with Alice, so that an utterance of (19) must be true as well. Yet, from the fact that (3) accords with Alice, it does not follow that the content imparted by

(22) Alice conversed with a duck

is true in Alice’s storyworld, in conformity with the intuition that the step from (1) to (19) is on a better footing than that from (1) to (20).

As hinted above, it is a general fact about indexical instances that certain pre-theoretic intuitions of ‘cogency’ must be recorded at a level other than that appropriate for the truth-conditional evaluation of sentences. Being committed to a demonstrative component, the Paratactic Analysis is hardly in a worse shape in this respect, for the reasons sketched above. Note however, in conclusion, an interesting detail of interest from the viewpoint of fictional discourse in particular: the dependence of the relationships between utterances of (1), (19), or (20) on a hypothesis of storyworld-consistency. If, as some seem to believe, intelligible fiction-making may result in inconsistent storyworlds, even utterance-based account would (rightly, on that assumption) fail to yield the sort of outcomes Lewis desires. ← 35 | 36 →


Adams, F. and Dietrich, L. (2004). What’s In a(n Empty) Name? Pacific Philosophical Quarterly 85: 125–148.

Adams, F. and Stecker, R. (1994). Vacuous Singular Terms. Mind and Language 9: 387–401.

Adams, F., Fuller, G. and Stecker, R. (1997). The Semantics of Fictional Names. Pacific Philosophical Quarterly 78: 128–148.

Almog, J., Perry, J. and Wettstein, H. (eds.) (1989). Themes from Kaplan. Oxford: Oxford University Press.

Bach, K. (2002). Giorgione Was So-Called Because of His Name. Philosophical Perspectives 16: 73–103.

Bertolet, R. (1984). On a Fictional Ellipsis. Erkenntnis 21: 189–194.

Braun, D. (1993). Empty Names. Noûs 27: 449–469.

Braun, D. (2005). Empty Names, Fictional Names, Mythical Names. Noûs 39: 596–631.

Bromberger, S. (2011). What Are Words? Comments on Kaplan (1990), on Hawthorne and Lepore, and on the Issue. The Journal of Philosophy 108: 486–503.

Caplan, B. (2004). Creatures of Fiction, Myth, and Imagination. American Philosophical Quarterly 41: 331–337.

Cappelen, H. (1999). Intentions in Words. Noûs 33: 92–102.

Cappelen, H. and Lepore, E. (2007). Language Turned on Itself. Oxford: Oxford University Press.

Carroll, L. (1865). Alice’s Adventures in Wonderland. In: L. Carroll (ed.), Alice’s Adventures in Wonderland and Through the Looking Glass. Ware, Hertfordshire: Wordsworth Editions, 1992.

Currie, G. (1985). What is Fiction? Journal of Aesthetics and Art Criticism 43: 385–392.

Currie, G. (1986). Fictional Truth. Philosophical Studies 50: 195–212.

Davidson, D. (1969). On Saying That. Synthese 19: 130–146.

Davidson, D. (1979a). Quotation. Theory and Decision 11: 27–40.

Davidson, D. (1979b). Moods and Performances. In: A. Margalit (ed.), Meaning and Use. Dordrecht: Reidel Publishing Company, 9–20.

Donnellan, K. (1972). Proper Names and Identifying Descriptions. In: D. Davidson and G. Harman (eds.), Semantics of Natural Language. Dordrecht: Reidel, 356–379.

Donnellan, K. (1974). Speaking of Nothing. The Philosophical Review 83: 3–31. ← 36 | 37 →

Doyle, A. C. (1887). A Study in Scarlet. In: A. Doyle (ed.), The Complete Works of Sherlock Holmes. London: Simon and Schuster, 2012.

Evans, G. (1973). The Causal Theory of Names. Proceedings of the Aristotelian Society supp. vol. 47: 187–225.

Everett, A. (2000). Referentialism and Empty Names. In: A. Everett and T. Hofweber (eds.): 37–60.

Everett, A. (2003). Empty Names and ‘Gappy’ Propositions. Philosophical Studies 116: 1–36.

Everett, A. (2005). Against Fictional Realism. The Journal of Philosophy 102: 624–649.

Everett, A. and Hofweber, T. (eds.) (2000). Empty Names, Fiction and the Puzzles of Non-Existence. Stanford: CSLI Publications.

Friend, S. (2000). Real People in Unreal Contexts. In: A. Everett and T. Hofweber (eds.), 183–203.

Geurts, B. (1997). Good News about the Description Theory of Names. Journal of Semantics 14: 319–348.

Goodman, J. (2003). Where is Sherlock Holmes? Southern Journal of Philosophy 41: 183–198.

Hawthorne, J. and Lepore, E. (2011). On Words. The Journal of Philosophy 108: 447–485.

Kaplan, D. (1977). Demonstratives. Ms. Reprinted in: J. Almog et al. (eds.), 481–563.

Kaplan, D. (1989). Afterthoughts. In: J. Almog et al. (eds.), 565–614.

Kaplan, D. (1990). Words. Proceedings of the Aristotelian Society supp. vol. 64: 93–121.

Kaplan, D. (2011). Words on Words. The Journal of Philosophy 108: 504–529.

Kripke, S. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press.

Kroon, F. (1994a). A Problem About Make-Believe. Philosophical Studies 75: 201–229.

Kroon, F. (1994b). Make-Believe and Fictional Reference. Journal of Aesthetics and Art Criticism 52: 207–214.

Kroon, F. and Voltolini, A. (2011). Fiction. In: E. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 Edition), https://plato.stanford.edu/archives/win2016/entries/fiction/.

Lewis, D. (1978). Truth in Fiction. American Philosophical Quarterly 15: 37–46.

Pelczar, M. and Rainsbury, J. (1998). The Indexical Character of Names. Synthese 114: 293–317. ← 37 | 38 →

Perry, J. (2001). Reference and Reflexivity. Stanford: CSLI Publications.

Phillips, J. (1999). Truth and Inference in Fiction. Philosophical Studies 94: 273–293.

Predelli, S. (2001). Names and Character. Philosophical Studies 103.2: 145–163.

Predelli, S. (2005). Contexts: Meaning, Truth, and the Use of Language. Oxford: Oxford University Press.

Predelli, S. (2008). Modal Monsters and Talk About Fiction. Journal of Philosophical Logic 37: 277–297.

Predelli, S. (2013). Meaning Without Truth. Oxford: Oxford University Press.

Recanati, F. (1993). Direct Reference. From Language to Thought. Oxford: Blackwell Publishers.

Reimer, M. (2001a). A ‘Meinongian’ Solution to a Millian Problem. American Philosophical Quarterly 38: 233–248.

Reimer, M. (2001b). The Problem of Empty Names. Australasian Journal of Philosophy 79: 491–506.

Reimer, M. (2005). The Ellipsis Account of Fiction Talk. In: R. Elugardo and R. Stainton (eds.), Ellipsis and Nonsentential Speech. Dordrecht: Kluwer, 203–216.

Russell, B. (1905). On Denoting. Mind 14: 479–493.

Sainsbury, M. (2005). Reference Without Referents. Oxford: Oxford University Press.

Salmon, N. (1998). Nonexistence. Noûs 32: 277–319.

Salmon, N. (2002). Demonstrating and Necessity. The Philosophical Review 111: 497–537.

Schnieder, B. and von Solodkoff, T. (2008). In Defence of Fictional Realism. The Philosophical Quarterly 59: 138–149.

Searle, J. (1975). The Logical Status of Fictional Discourse. New Literary History 6: 319–332.

Taylor, K. (2000). Emptiness Without Compromise. In: A. Everett and T. Hofweber (eds.), 17–36.

Thomasson, A. (1999). Fiction and Metaphysics. Cambridge: Cambridge University Press.

van Inwagen, P. (1977). Creatures of Fiction. American Philosophical Quarterly 14: 299–308.

Walton, K. (1990). Mimesis as Make-Believe. Cambridge, MA: Harvard University Press. ← 38 | 39 →

Walton, K. (2000). Existence as Metaphor?. In: A. Everett and T. Hofweber (eds.), 69–94.

Yagisawa, T. (2001). Against Creationism in Fiction. Philosophical Perspectives 15: 153–172.

Zemach, E. (1998). Tom Sawyer and the Beige Unicorn. British Journal of Aesthetics 38: 167–179.

1 For a discussion of Millian rejections of Intuition (typically appealing to the distinction between semantically encoded and pragmatically conveyed content) see Taylor (2000), Braun (1993), Adams and Stecker (1994), Adams et al. (1997), Reimer (2001a, 2001b), Caplan (2004), Everett (2003), and Adams and Dietrich (2004). For a summary of non-Parsimonious approaches, see Kroon and Voltolini (2011); for a popular Millian non-Parsimonious strategy, the idea that fictional names refer to actual abstracta, see van Inwagen (1977,) Salmon (1998, 2002), Thomasson (1999); see also Everett (2005) and Schnieder and von Solodkoff (2008).

2 I assume that Carroll’s novel, though perhaps inspired by Ms. Liddell, is not about that woman (modify my examples if you disagree). Regarding ‘ordinary’ names in fiction, as in the case of ‘London’ in Doyle’s stories, see Walton (1990: chapter three), Kroon (1994a, 1994b), and Friend (2000).

3 I deliberately leave aside approaches to names as indexical (see Pelczar and Rainsbury 1998 and Recanati 1993; for criticisms see Predelli 2001): Millianism, in my understanding, is the view that names are endowed of a constant character (in the sense of Kaplan 1977), devoted to directly designating their referents (if any exist).

4 See Kripke (1980), Donnellan (1972), Evans (1973), and the ensuing literature.

5 Concerning the metaphysical status of words see Kaplan (1990), Cappelen (1999), Bromberger (2011), Hawthorne and Lepore (2011), and Kaplan (2011).

6 On the ideas of ‘pre-semantic’ and ‘metasemantic’ see Kaplan (1989), Perry (2001), and Predelli (2005).

7 For a pioneering study of emptiness and referentially unsuccessful dubbings see Donnellan (1974); see also Sainsbury (2005).

8 According to a popular Parsimonious Millian strategy, in particular, empty names engender ‘gappy propositions’. The truth-value for (atomic) gappy propositions is contentious. According to Braun (1993, 2005) they are inevitably false; they are neither true nor false according to Adams and Stecker (1994), Salmon (1998), Taylor (2000), and Reimer (2001a, 2001b). See also Caplan (2002) and Everett (2003).

9 Lewis’ position is straightforwardly non-Millian (“as we use it, [‘Holmes’] may… have a highly non-rigid sense …” [Lewis 1978: 41]), his take on storytelling being motivated by a characteristically descriptivist problem, that of ‘accidental reference’.

10 According to a widespread view, the author of a work of fiction ‘does not actually perform illocutionary acts but only pretends to’ (Searle 1975: 327; see also Currie 1985). I agree: not because a propositional content is being put forth, without thereby being the content of an assertion (question, command, etc.), but because no propositional content is at all appropriate for the example under discussion. See also footnote 18.

11 Kendall Walton “accept[s] the view that, since there is no such thing as [Alice], [(3) and (4)] do not express a proposition” (Walton 2000: 76). Still, at least on an empty-name approach to Parsimony, there may be ‘no such thing as Alice’, while (3) and (4) still do express a (gappy) proposition, in the sense of footnote 8. According to the NNH, then, Walton’s remark requires quotation-marks: I, for one, accept the view that, since there is no such thing as ‘Alice’, the name, the sentence-types (3) and (4) do not express a proposition.

12 Note that the NNH’s notion that (5) and (6) do not encode propositions is to be distinguished from the claim that they encode gappy propositions, in the sense of footnote eight. Since the same gappy proposition would presumably be encoded in either sentence, the empty-name theorist must appeal to non particularly well-developed epicycles, such as the notion that what is at stake in one case, but not n the other, is a presentation of that proposition in an ‘Alice-ish’ way (see Braun 2005: 602).

13 For discussion of imparted content see Perry (2001) and Predelli (2013). The content imparted by the utterance of an expression ought not to be confused with the content pragmatically conveyed by the use of a sentence on a given occasion – the latter, at least according to a traditional Gricean line, in turn derivable partly on the basis of encoded content.

14 For metalinguistic (and straightforwardly non-Millian) theories of names, see for instance Bach (2002) and Geurts (1997).

15 ‘And much more besides’: of course, here as in actuality, imparted content is opulent. My actual utterance of (7) imparts not only the content encoded in (8), but also, say, that someone is speaking, that Aristotle has the property contributed by ‘was fond of dogs’, and the like. By the same token, what is fictionally imparted by Carroll’s teller includes affairs such as that the referent of ‘Alice’ has the property encoded in ‘beginning to get very tired’, that someone speaks of siblings and boredom, and … much more besides. Here as in actuality, whether this sort of content plays any significant role hinges on general issues pertaining to the economy of communication. Much inevitably remains as mere ‘background noise’ – including, in many actual interactions involving an utterance of (7), the content put forth in (8). The case of (9) is of particular interest for a two-fold reason: firstly, due to the NNH’s interest in fictional name, and, relatedly, due to its insistence that, in these cases, no actual singular proposition ‘about Alice’ is available to the reader.

16 Concerning ‘imported’ truths of this sort, see Lewis (1978), Walton (1990: chapter four), Currie (1986), and Phillips (1999).

17 ‘Talk about fiction’ is sometimes characterized as ‘narrative’ or, more appropriately from my viewpoint, as ‘conniving’ (Everett 2000: 38). See the brief remarks in section five about the contrast with meta-narrative and factual scenarios.

18 Here as before, further details about ‘the right setting’ are left to the attention of neighbouring disciplines in this case, I suppose, a theory of conversational salience or some other type of pragmatic inquiry.

19 The locution ‘talk about fiction’ perniciously invites the “assumption that the speaker is making a genuine assertion about a fictional world … from a perspective outside of it …” (Walton 1990:392); see in particular Searle (1975), according to which, when ‘taken as a piece of discourse about fiction, [4] is true because it accurately reports’ the history of the fictional character Alice (Searle 1975: 329). Leaving Walton’s focus on assertion aside, the retelling model concurs with his ‘pretense construal’, at least with respect to the suggestion that “the appreciator [is] pretending to describe the real world rather than actually describing a fictional one” (Walton 1990: 392). Intriguingly, Walton also briefly mentions an extension of this idea to cases of mere interjections, as when Charles exclaims ‘yikes!’ when watching a horror movie (Walton 1990: 393).

20 According to a widely debated hypothesis, at least some utterances of (3) and (4) are to be understood as ‘elliptical’ for utterances of (1) and (2). If that is the case, they ought to be treated according to my analysis of (1) and (2) in section five – in particular, in that case, as true or false. See among many Lewis (1978), Bertolet (1984), and Reimer (2005).

21 For this reason, the customary analysis of (1) in terms of a sentential operator (‘according to Alice’) and a simpler sentence will not do for the NNH. This is hardly surprising, since that analysis generates an unresolvable clash between Intuition and any version of Parsimonious Millianism: according to the treatment of ‘Alice’ and ‘Dinah’ as empty, because the alleged operanda end up being semantically indistinguishable; according to the NNH, because the allegedly embedded material, being ‘semantically inert’, does not provide any argument whatsoever. Note that the clash ensues independently of the question whether the alleged operator is an intensional operator (as in Lewis 1978) or an operator on character (as in Predelli 2008).

22 Here (as, I suppose, in Davidson’s case), the notion of ‘being of the form’ is to be understood as neutral with respect to the syntactic analysis of (1), and in particular with respect to the structure for its Logical Form. See Davidson (1969, 1979a, 1979b).

23 ‘Equivalence’ in the sense of the classic double-indexed logic for indexical languages, that is, as extensional indistinguishablity with respect to (any circumstance of evaluation and) any content of interpretation. For parallel considerations regarding Davidson’s view of pure quotation, and for a detailed discussion of the ‘demonstrative’ component, see Cappelen and Lepore (2007) and Predelli (2013).

24 A further interesting query pertains to the similitudes and differences between fictional talk and myths, false theories, and imagination; see Braun (2005), Caplan (2004), Walton (1990, section 2.8.), Walton (2000).

25 See for instance Everett (2005), Goodman (2003), Reimer (2001a), Salmon (1998), Schnieder and von Solodkoff (2008), Searle (1975), Thomasson (1999), Yagisawa (2001), and Zemach (1998).

26 Oddly, Lewis insists that ‘… there is a pragmatic paradox akin to contradiction in a third-person narrative that ends “… and so none were left to tell the tale”’ (Lewis 1978: 40). As far as my intuitions are concerned, it is precisely the absence of paradoxality that fuels the need for the periphery-storyworld distinction (see also Walton 2000: section 6.5).

27 Or, better, the author’s language of choice: the narrator in A Clockwork Orange, for instance, employs expressions interpretable according to Nadsat, an argot invented by its (English-speaking) author.

28 It is an independent issue whether, in this latter case, he would have authored Alice, the very novel he did in fact write. But it is a consequence of the approach proposed above that Carroll’s teller would fictionally have described the same storyworld he, in fact, fictionally describes.

29 That a name-type at least phonetically distinct from that fictionally employed by Carroll’s narrator occurs in (13) is hardly cause of concern for the Chain Picture, at least if is at all in the position of explaining the link between, say my tokens of ‘Aristotle’ and a certain baptism involving the name-type ‘Aristoteles’.

30 At least in a fairly standard sense of validity in the logic of demonstratives: an argument is valid iff, for all models M and context c such that the premises are true in c and M, the conclusion is as well.

| 41 →

Eros Corazza

ILCLI The University of the Basque Country, Donostia, Spain
Basque Foundation for Science, Bilbao, Spain
Carleton University, Ottawa, Canada

Christopher Genovesi

Carleton University, Ottawa, Canada

On the Metaphoric Uses of (Fictional) Proper Names

Abstract: In this paper we will discuss how both proper names for existing individuals and proper names for fictional entities can be used metaphorically, even when they appear in argument position. Thus, whether a name used metaphorically is a referential or a non-referential term does not affect the interpretation we propose. For, when used metaphorically, proper names cease to be, properly speaking, proper names. That is to say, they cease to be Millian tools of direct reference. When used in argument position in a simple sentence of the subject/predicate form, we claim that the tokened name used metaphorically picks up an individual analogous to the way Donnellan treats the referential uses of descriptions. That is, just as a description used referentially can pick up an individual independently of the property used to identify it, a proper name used metaphorically in argument position can pick up an individual independently of the latter carrying the name uttered. In such metaphorical uses the name serves the function of attributing to the individual the speaker has in mind those properties that are relevant in the grasping of the speaker’s metaphorical meaning. Thus, a name so used loses its individuative (referential) power. We will claim that in such uses, reference is guided and fixed, like in Donnellan’s referential use of descriptions, by the individual the speaker has in mind and raises to salience in a communicative interaction. Yet, the account we propose does not undermine the Millian view that proper names are tools of direct reference nor the Fregean view that when a proper name appears in a predicative position it ceases to be a proper name. Our final aim is to accommodate both views and address the question as to why the speaker would want to utter a metaphor in this way.

Keywords: Frege, metaphors, predicates, proper names, property transfer, reference transfer ← 41 | 42 →

0 Introduction

Suppose we say of our friend Bill, who has just returned home from a long and arduous journey,

(1) Odysseus returned home

with the intent to communicate the acknowledgement that Bill has had a long and toiling journey, and is a brave and persistent man like the mythical Odysseus (or something along these lines). We refer to Bill as something he is not. For Bill is not Odysseus. We know that ‘Odysseus’ refers to a fictional character from the Homeric epic, The Odyssey. Despite this discrepancy in reference, it is fairly standard (i.e., statistically common) to use proper names metaphorically. A speaker understands and interprets utterances of this sort with relative ease. How is it that a tokening of ‘Odysseus’ can be used (metaphorically) to refer to someone else? That is, to someone who is not the bearer of the name ‘Odysseus’? If, instead, one says:

(2) Donald Trump returned to Washington

intending to talk about the man, who as of the 20th of January, 2017 became the president of the USA, the name ‘Donald Trump’ is used referentially. As such, the tokening of this name refers to the individual, Donald Trump. Therefore, with the tokening of ‘Donald Trump’ one refers to Donald Trump and what one says is true iff Donald Trump returned to Washington, false otherwise. The question we pursue presently concerns how utterances like (1) are understood and whether this understanding can be guided by our knowledge of more standard cases such as (2).1

We begin by highlighting some obvious differences between metaphorical uses of names and more standard cases. First, metaphorical uses of proper names seem to rely heavily on contextual input. Without such consideration, an utterance of (1) – where the name ‘Odysseus’ appears in argument position – operates in the referential way, selecting Odysseus, and thereby predicating of him the property of having returned home. If so, the truth-conditions could be captured, roughly, as follows (avoiding, for simplicity, issues of time pertaining to the tense of the utterance):

(3) An utterance u of “Odysseus returned home” is true iff Odysseus returned home. ← 42 | 43 →

While truth-conditions along these lines are suited for utterances like (2), they do not capture what the speaker of (1) intended to communicate. The speaker intends to communicate something about Bill returning home, not about Odysseus returning home. Moreover, the speaker makes a further speech act by using the name ‘Odysseus’ metaphorically to express something about Bill and his homeward journey. Thus, the tokened name ‘Odysseus’ in (1) cannot refer to Odysseus, the fictional character. The challenge is, therefore, to figure out how a fictional name, or any other name used metaphorically for that matter, can be used to refer to someone who is not the bearer of the name.2 Furthermore, on the Fregean interpretation, fictional names are empty terms. So, we are also left to wonder (i) how the tokening of an empty name comes to refer to Bill, and (ii) says something meaningful about him.

To start, we compare our account with more usual cases. There is a great deal of literature discussing the view that proper names are predicates (see for instance, among others, Burge 1973, Castañeda 1989, Fara Graff 2015). We do not follow this path and maintain that proper names are singular terms and tools of direct reference. Yet we recognize that in some cases proper names can be used in a metaphorical way, like ‘Maria Callas’ in the following utterance

(4) Jane Smith is not Maria Callas

with the intended meaning that as a singer Jane Smith is not as good as Maria Callas. In this example the name ‘Maria Callas’ is not (primarily) used to refer to Maria Callas. In such an utterance, ‘Maria Callas’ is used metaphorically. After all, the speaker does not intend to communicate a trivially true statement of the form a is not b. Rather, on the intended interpretation, ‘Maria Callas’ works in a predicative way picking out some stereotypical properties (e.g., being a good singer, etc.) of the name’s referent, in our case Maria Callas, that are relevant to the context of discourse. In such a case the proper name ceases to be a (logical) proper name and such an utterance is not of the form a is not b, but of the form Fa. Following Frege, we endorse the view that a proper name (Eigenname) cannot be used as a predicate unless it ceases to be a proper name. As Geach puts it:

A proper name is never used predicatively – unless it ceases to be a proper name, as in ‘He is not a Napoleon of finance’ or (Frege’s example) ‘Trieste is no Vienna’; in such cases the word alludes to certain attributes of the object customarily designated by the proper name. (Geach 1962: 42; italics added) ← 43 | 44 →

We maintain that the Frege/Geach view is essentially right, and believe that it can illuminate the phenomenon under our present investigation.

So, putting this altogether, the issues we tackle in this paper will be threefold. How do we analyse utterances such as (1) if: (i) proper names cannot appear in predicative position without ceasing to be proper names, (ii) names can be used to pick up a referent even though the latter does not carry that name, and (iii) fictional names are empty terms?

Bearing Frege’s insights in mind we proceed to give an account of cases like (1) that rest on understanding the contribution of the name as one of property transfer. In such cases, we argue that the name ceases to be a proper name. In metaphorical contexts (such as [1]), the tokening of a fictional name is associated with sets of properties from its fictional discourse. Rather, the name functions as a predicate and selects relevant properties that are projected onto the speaker’s intended referent, such as our example of Bill.

1 Frege on Fictional Names and Predicative Expressions

Before we get into those nonstandard cases identified at the beginning of the paper, it will help to understand how Frege treats normal cases. Consider the following utterance:

(5) Bill returned home

Let us assume for the moment that Bill did in fact return home (again, ignoring issues of time). For Frege, the subject of this utterance refers to an object, while the predicate to a concept. The tokened name ‘Bill’ refers to the object Bill, while the predicate ‘returned home’ refers to a concept. If the object (the referent) is subsumed under the concept the utterance is true, false otherwise. Thus, granting our supposition, (5) is true. This conception seems to mesh well with our intuitive notion of the way we take and understand standard cases (such as (2) and (5) above). Here, the predicate describes that some action was performed by the referent of the subject of the sentence – that Bill returned home. So far, so good.

To highlight the importance of Frege’s view that names cannot appear in predicative position without ceasing to be proper names, it is worth taking into account what Frege says about the relationship between objects and concepts and, in particular, between predicates qua incomplete expressions and proper names qua complete ones. In “Function and Concept” (1891) and “On Concept and Object” (1892a) Frege advances the view that a functional expression is incomplete. As such, it needs to be completed by an argument. This is often recognized as one of Frege’s main logical insights: ← 44 | 45 →

When we come to interpret a functional expression, say Fx, we have to assign some domain of objects to the variable x. That is, we have to determine the range of the variable. Furthermore, we have to assign to the function-letter F a function from and onto that domain. If the function-letter F is coupled with a constant, then to obtain Fa we have to assign to the constant a an object in the relevant domain. In translating this discourse into the subject/predicate parlance we would have to assign a domain to the incomplete predicate and an object to the singular term occupying the argument place.3 Let us consider the utterance of a simple subject/predicate sentence such as:

(6) Bill Smith is brave

Intuitively, the speaker uses the name ‘Bill Smith’ to designate Bill Smith and attributes to him the property of being brave. What the speaker says is true if Bill Smith instantiates that property. But Bill Smith, the object picked out by the tokened name ‘Bill Smith’, is not defined by him possessing the property of being brave or any other property for that matter. For example, (6) could be false, yet ‘Bill Smith’ would designate Bill Smith independently of him being brave or having any of his other physical (or psychological) characteristics. What matters, for (6) to be true (or false) is that the name ‘Bill Smith’ designates an object that satisfies (or not) what the predicate ‘is brave’ stands for. In Frege’s parlance, that is to say whether the object Bill Smith is subsumed or falls under the concept designated by the predicate ‘is brave’. Following Frege’s insights, (6) must be analysed in terms of a function and its value range. Hence, if we withdraw the name ‘Bill Smith’ from (6) we are left with an open sentence:

(7) x is brave ← 45 | 46 →

The predicate, qua functional expression, must take an argument to form a sentence (a whole). In the case of a simple sentence of the form Fa, it must take a singular term. In order for a monadic predicate to form a sentence of the form Fa it must take a proper name (in Frege’s sense) in its argument position.4

What is pertinent for us is to grasp how a simple sentence of the subject/predicate form, Fa, functions when a no longer denotes an object. To simplify, consider the truth-conditions of (5):

(8) An utterance u of “Bill returned home” is true iff Bill returned home

If we analyse (1) in the same way as (4), we would have:

(9) An utterance u of “Odysseus returned home” is true iff Odysseus returned home

For Frege two things are missing. On the one hand, Odysseus does not exist, so there is nothing to refer to. In Frege’s terminology, the name ‘Odysseus’ is an empty term. Recall that sentences refer to whatever the reference of the subject gets mapped to by the function referred to by the predicate. Therefore, the entire sentence lacks a reference, a truth value. In short, since the subject lacks reference, the predicate qua functional expression returns no value. As it is well known, for Frege (see Frege 1892b), names refer via a sense (or mode of presentation). The referent is the object that satisfies the sense. The sense expressed by the name enters as a constituent of the thought expressed by the utterance. In the case of empty names, although an utterance expresses a thought, it lacks truth value, insofar as we have no object subsumed by the concept referred to by the predicate.

Furthermore, when a proper name is embedded in an oratio obliqua construction, it is never an empty term. In such contexts, words change their mode of reference. For the referent of an embedded name is its customary sense. Consider:

(10) Penelope believes that Odysseus returned home

This ascription is true iff Penelope believes that Odysseus returned home, independently of whether ‘Odysseus’ is an empty term. The name within the subordinate clause comes to denote the sense that it normally expresses in an extensional context, and the entire subordinate clause refers to the thought that Odysseus returned home.

In short: for Frege, fictional names are meaningful (they express senses), but lack reference. As such, an utterance containing an empty name lacks a truth-value (with the exception of oblique contexts where the names denote the ordinary sense and the that-clause refers to its ordinary thought rather than a truth value). ← 46 | 47 → This was all Frege had to say on the matter of fictional names.5 As a result, he was silent as to how cases such as (1) referred not to their standard referent, but are somehow used to pick out another referent. After all, Frege was not in the business of figuring out non-literal uses of language. In what follows, we hope to demonstrate that despite the fact that fictional proper names lack reference, they are still part of a fictional universe, and are equipped with properties and attributes that are used to make a substantial contribution within the fictional discourse. This point will be important to note especially when we speak about the metaphoric use of fictional names as a matter of property transfer.

2 Proper Names: Predicative and Metaphoric Uses

The fact that fictional characters can be the bearers of properties, characteristics and attributes (or using Frege’s vernacular, that the names of fictional characters express a sense, as Frege would claim) is an important point of the story we want to tell.6 For we believe that the metaphoric use of a proper name serves to pick out contextually relevant properties of the referent (when the name used metaphorically refers to an existent entity) or the fictional character (when the ← 47 | 48 → name used metaphorically is a fictional name). Which properties are relevant is of course a matter of the discourse context--and so, in our example the relevant properties associated to Odysseus.7 We can see how this works when comparing it with attributive uses of proper names in predicative position. Consider the following utterance:

(11) Bill Smith is not Lionel Messi

‘Lionel Messi’ is used predicatively (and it picks up a particular attribute of Messi, e.g., a good soccer player). It is not expressing the trivial true proposition of the form a is not b. This is the same case with (4) outlined above (“Mary Smith is not Maria Callas”). The story we would like to tell about the metaphoric use of proper names in argument position takes into consideration these observations from cases such as (2) and (5). Reconsider the original utterance (1) that we repeat here:

(1) Odysseus returned home

In this case as well, though the name appears in argument position, it is not a proper name, but a means through which we refer to our friend Bill, and predicate something of him while doing it. When a name – even in argument position – is used attributively/predicatively etc. it does not work as a singular term, or what we may refer to as a Millian tool of direct reference. In our example the speaker is not saying that Odysseus returned home; he is talking about Bill.8 We believe that by using the name metaphorically, we force a predicative use of the name (and thereby the name ceases to function as a proper name). It follows that we are not dealing with a case of deferred reference, whereby we must first pick out a unique individual to get the interpretation going. Rather than selecting a unique individual (e.g., Odysseus) the name delivers relevant stereotypical properties – which happens whether the name is empty or not.9 The question we face is then: How the tokening of ‘Odysseus’ in our example, can pick up Bill as the object of discourse and our utterance be of the subject/predicate form, Fa. If so, ← 48 | 49 → the referent of the name does not depend on the existence of the referent of the name appearing in argument position, i.e., Odysseus. We uttered (1) as a means to communicate that our friend Bill had returned home. Hence, reference cannot be fixed by the tokening of ‘Odysseus’. At the same time, we made a further speech act to indicate certain attitudes we believe obtain of our friend Bill. For example, his bravery, persistence, and perhaps that he is worthy of praise.

Yet, it is one thing to take Frege at his word, and another thing altogether to demonstrate that his claims apply to the metaphorical use of proper names. In the next section, we attempt to show why we must appeal to property transfer as opposed to the view of deferred reference.

3 Property Transfer and Reference Transfer

In order to get at what we have in mind with property transfer, we will contrast it with cases of deferred reference (à la Nunberg 1993, 1996, 2004).10 A standard case of deferred reference could be captured in the following example. Imagine a situation where a customer is handing the key of her car to the valet, she says to him:

(12) This is parked out back

Nunberg tells us that (12) is a case of deferred indexical reference. In (12), the speaker’s demonstration of holding out her key along with the utterance informs the valet of the car’s location. At the same time, the utterance of ‘this’ corresponds to her key, and functions referentially to her car. She makes the car salient by relying on the hearer to make the (metonymic) association between key and car. Here, the key is the index of ‘this’ and the speaker’s car is the referent.11 To make the point clearer, we shall consider a slightly different example:

(13) I am parked out back.

Nunberg and Bezuidenhout both suggest that an utterance like (13) is not a matter of deferred reference. That is to say, the indexical ‘I’ does not function referentially to select the speaker’s car. Nunberg emphasizes this by relying on a series of tests to distinguish between cases of reference and property transfer. We are told that in case of genuine deferred reference, the referent drives the referring ← 49 | 50 → expression. So, in a context where the speaker has two cars parked in the back lot, she could not say:

(14) * We are parked out back.

(Following the linguists’ convention we signal infelicitous and ungrammatical sentences by prefixing them with ‘*’). But, if the context was such that the speaker had two or more cars parked out back and one single key for all of them, then she could say:

(15) These are parked out back.

The second test that Nunberg employs to examine whether we are dealing with deferred reference is a conjunction test. In cases of deferred reference, it is possible to conjoin new information to the referent in a way that is not as clear as when the interpretation of the utterance relies on the transfer of relevant properties. Example (16) has a felicity that (17) lacks:

(16) This is parked out back and in need of cleaning.

(17) * I am parked out back and in need of cleaning.

What these two examples show is that what is picked up from the referring expression in (16) is the referent (demonstrated by the key). By contrast, we are told that the infelicity in (17) is due to the fact that the indexical does not function in a deferred referential way to refer to the speaker’s car. We submit that with these considerations in mind, we can show that the metaphoric use of a proper name does not function as a case of deferred reference. Rather, what happens in examples (1), (4), and (11) is that the proper names are an efficient way to select some subset of properties associated with their referents in order to make some statement about the intended referent.12 We turn to our original example (1). We may only conjoin information to the sentence if the information is about Bill; but not about Odysseus:

(18) Odysseus returned home and wanted to wash his clothes.

(19) * Odysseus returned home and kissed Penelope.

As (18) shows us, property transfer is restricted to the proper name. The predicate ‘returned home’ is not included in the property transfer, nor is the second ← 50 | 51 → conjunct.13 For example, it would be impermissible to say of Bill having returned from his journey:

(20) * Odysseus slayed the cyclopes/returned to Ithaca.14

By now we hope that it seems clear that we are trying to preserve Frege’s insight that proper names in argument position when used metaphorically cease to be proper names. As such, they do not provide a referent. Instead they work like a predicate and contribute in the transferring of some relevant properties to the intended referent. In this section, we provided evidence in favour of treating the metaphorical use of proper names as a means of property transfer. In the following section, we try and formalize the observations we made concerning property transfer.

4 Metaphorical Uses of Proper Names in Argument Position

Our initial question was: How one can refer to Bill by uttering ‘Odysseus’? We have here a case similar to the referential uses of descriptions that mischaracterize the referent (see Donnellan 1966). As one can succeed in the referring to Bill using the description ‘The man with the Martini’ even if Bill is drinking water, one can refer to Bill using the name ‘Odysseus’ even if Bill does not carry that name. In other words, reference can be fixed without the referent having to instantiate the property attributed. So, Bill can be referred to using ‘Odysseus’ despite the fact that he does not satisfy the property of being called/labelled/named ‘Odysseus’. Here, like in the cases discussed by Donnellan, reference is fixed by the having in mind.15 The speaker succeeds to refer to Bill using ‘The man with the Martini’ or ‘Odysseus’ because she has Bill in mind. Actually, for the audience to understand ← 51 | 52 → the use of ‘Odysseus’ as a metaphorical use she ought to grasp that the speaker does not uses the name in a referential way, i.e., to refer to the bearer of ‘Odysseus’. In understanding that the name is used metaphorically the audience ought to understand that the name is used predicatively (and, thus, ceases to be a proper name in Frege’s sense) as a means to transfer contextually salient properties from the actual name’s bearer to the intended referent. Furthermore, the audience needs to grasp who the speaker has in mind. In our example we intended to talk about Bill. The audience can then process the metaphor and infer that Bill had a long and tumultuous journey home. Thus, the two utterances:

(21) Bill returned home

(22) Odysseus returned home

with the speaker having Bill in mind and intending to refer to him get regimented differently. In (21) ‘Bill’ works as a proper name, as a tool of direct reference picking up Bill as the object of discourse. In (22) ‘Odysseus’ does not and cannot work as a proper name. It works as a predicate picking up some relevant properties of Odysseus that get transferred to Bill, the referent.

Yet we want to maintain that this utterance is of the subject predicate form, Fa, and thus of the same form as (21). To deal with this phenomenon we propose to adopt Kaplan’s treatment of Donnellan’s referential use of descriptions. Kaplan (1978), like Donnellan, takes a description used referentially to be a singular term and suggests that it works like a complex demonstrative. Kaplan (1978: 23) introduces “a new word, ‘Dthat’ for the demonstrative use of ‘that’” and argues that the referential use of the description can be regimented as: ‘Dthat[the F]’. Kaplan claims that ‘Dthat’ can be used to make any description both a singular term and a rigid designator, i.e., a term referring to the same individual in all possible worlds in which the latter exists. Following Kaplan’s suggestion, a sentence of the form “The F is G”, with the description ‘the F’ used referentially, can be represented as:

(23) Dthat[the F] is G

A word of clarification may be appropriate. In his latest study, Kaplan (1989) recognizes that when he first introduced ‘Dthat’, he confused two distinct uses, namely ‘Dthat’ as an operator with ‘Dthat’ as a demonstrative surrogate. In the former case, ‘Dthat’ is better understood as a rigidifier and, as such, the dthat-term ← 52 | 53 → becomes a rigid designator without being a directly referential term.16 In the latter case, however, the dthat-term is a singular, directly referential, term contributing the referent into the proposition expressed. In order to capture the referential use of a definite description, ‘Dthat’ should be understood as a demonstrative surrogate.

The operator interpretation is not what I originally intended. The word ‘Dthat’ was intended to be a surrogate for a true demonstrative, and the description which completes it was intended to be a surrogate for the completing demonstration. On this interpretation ‘Dthat’ is a syntactically complete singular term that requires no syntactical completion by an operand. (A ‘pointing’, being extralinguistic, could hardly be a part of syntax.) The description completes the character of the associated occurrence of ‘Dthat’, but makes no contribution to content. … ‘Dthat’ is no more an operator than is ‘I’, though neither has a referent unless semantically ‘completed’ by a context in the one case and a demonstration in the other. (Kaplan 1989: 581)

To summarize, the referential use of a description goes proxy with the use of a complex demonstrative, i.e., an expression of the form ‘that/this F’.17

We think that a similar story can be told about the metaphorical uses of proper names. Thus, our initial example, “Odysseus returned home” can be regimented as:

(24) Dthat[Odysseus] returned home18 ← 53 | 54 →

with the surrogate demonstrative picking up, in the context of the conversation Bill, the individual the speaker has in mind. This utterance is true if Bill returned home. Yet in uttering ‘Odysseus’ as a metaphor the speaker also attributes to Bill some stereotypical features commonly associated with the fictional character Odysseus. We understand the projection of some of these features along the same lines as Nunberg’s account of property transfer. We refined this to mean that in cases of metaphorical use of proper names in argument position, reference is picked up by the implicit demonstrative surrogate. As an implicit index, ‘Dthat’ and the name appearing within the parenthesis helps in transferring some relevant properties associated with the name to the referent picked up by ‘Dthat’.

One could ask how is it that the name ‘Odysseus’ within the parenthesis contributes in selecting relevant properties that then get transferred to the referent of the ‘Dthat’ term, Bill. Our view is simple, for the speaker and the audience to understand an utterance like (1), as regimented as (23), they must consider the name ‘Odysseus’ to be used metaphorically. It is then a matter of the context which relevant properties, or stereotypes, are raised to salience and then transferred to Bill, the referent of the ‘Dthat-term’ which happen to be the individual the speaker has in mind and intends to talk about. In other words, it is a matter of the discourse situation that partly determines how a name used metaphorically contributes in bringing forward some properties usually associated to the bearer of the name. In a given context, the use of ‘Odysseus’ will transfer to the individual just those properties of the fictional character that the speaker intends to talk about in a metaphorical way. It is worth keeping in mind, that a name so used, like the descriptive content associated with a complex demonstrative of the form ‘that F’, does not play any individuative role. The individuating role is played (or, perhaps, represented) by the demonstrative ‘that’ or the implicit ‘Dthat’. The ‘F’ associated to the complex demonstrative ‘that F’ and ‘Odysseus’ associated to ‘Dthat (Odysseus)’ are mere pragmatic tools that help the audience in grasping whom the speaker has in mind and, in the case of the metaphorical use of ‘Odysseus’, what features the speaker intends to metaphorically attribute to the referent.19 ← 54 | 55 →

Our investigation was propelled by Frege’s view of proper names. We attempted to expand his view to cases we classified as metaphorical uses of proper names in argument position. We claim that when a name in argument position is used metaphorically, it ceases to be, properly speaking, a proper name. It loses the feature to refer to its object (read: referent). It becomes a predicate and, like any predicate, it should refer to a concept. The concept is a bundle of stereotypical properties based on the individual it is usually used to refer to. In keeping with Donnellan, we have shown that these properties are attributed to the individual the speaker has in mind and intends to talk about. In short, in an example like (1) we have two kinds of predication at work: The first is given by the metaphorical use of the name ‘Odysseus’ and the second by the predicate ‘returned home’. Therefore, in a certain respect, the speaker somewhat ends up saying20 two things: that Bill returned home and that Bill had a hard journey, has been brave, etc.

However, this raises an interesting question: in using a proper name metaphorically in argument position, such as our original example (1), does the speaker express two distinct propositions that can be independently evaluated, or simply the proposition that Bill came home while also pragmatically conveying some stereotypical properties usually associated with the mythical Odysseus to Bill. We leave this question open for further investigation.

5 Reflection and Conclusion

Why would one want to utter “Odysseus returned home” referring to Bill, rather than “He/that guy returned home”? In the latter case, the speaker’s intention is primarily a matter of identifying the referent, and picking out some external state of affairs – that the referent the speaker has in mind returned home. By contrast, by uttering the former, the speaker’s primary intention is to convey ← 55 | 56 → an attitude toward their intended referent. That is to say, it is not enough just to identify the intended referent, rather we are invited – perhaps even ‘bullied’ – into adopting, or at least entertaining, the speaker’s attitude toward the intended referent.21 This intention is a reflexive one: the speaker prompts the audience to conjure up a relevant subset of properties they take the speaker to believe obtain between Bill and Odysseus.22 In uttering (1), the speaker’s intentions are to get their audience to understand that Bill had an arduous journey home, and that Bill was steadily persistent in the course of action despite the difficulties, etc. We are able to come to this interpretation by transferring those relevant properties about Odysseus to Bill.

In this paper we defended the Frege/Geach view that when names appear in predicative position they cease to be proper names and often works in a metaphorical way. We also claim that a name can be used metaphorically when it appears in argument position in a simple sentence of the subject/predicate form. We believe the same thoughtful considerations hold between our view of the metaphorical use of proper names in argument position and Donnellan’s treatment of the referential use of definite descriptions, which can be represented using Kaplan’s ‘Dthat’ understood as a demonstrative surrogate. Furthermore, we argued that when so used a name contributes in transferring some relevant properties usually associated with the name bearer to the referent the speaker intends to talk about. ← 56 | 57 →


For comments and suggestions on the penultimate version of this paper we would like to thank two referees for the volume. While working on this paper Eros Corazza has been partly supported by a grant from the Spanish Minister: FFI2015-63719-P (MINECO/FEDER) and the Basque Government (IT1032-16).


Bezuidenhout, A. (2001). Metaphor and What is Said: A Defense of a Direct Expression View of Metaphor. Midwest Studies in Philosophy 25: 156–186.

Bezuidenhout, A. (2008). Metaphor Singular Reference: The Role of Enriched Composition in Reference Resolution. The Baltic International Yearbook of Cognition, Logic, and Communication 3: 1–22.

Burge, T. (1973). Reference and Proper Names. Journal of Philosophy 70 (14): 425–439.

Camp, E. (2008). Showing, Telling, and Seeing: Metaphor and “Poetic” Language. The Baltic International Yearbook of Cognition, Logic, and Communication 3: 1–22.

Castañeda, H.-N. (1989). Thinking, Language, and Experience. Minneapolis, MN: University of Minnesota Press.

Corazza, E. (2002). Description-Names. Journal of Philosophical Logic 31 (4): 313–325.

Corazza, E. (2003). Complex Demonstratives qua Singular Terms. Erkenntnis 59 (2): 263–283.

Corazza, E. (2009). Fiction without Ficta. Dialectica 63 (1): 67–74.

Corazza, E. (2017). Names, Identity, and Predication. Philosophical Studies. DOI 10.1007/s11098-017-0975-5.

Davidson, D. (1978). What Metaphors Mean. Critical Inquiry 5 (1): 31–47.

Dever, J. (2001). Complex Demonstratives. Linguistics and Philosophy 24: 271–330.

Donnellan, K. (1966). Reference and Definite Descriptions. The Philosophical Review 75: 281–304. Reprinted in K. Donnellan (2012). Essays on Reference, Language, and Mind. New York: Oxford University Press, 3–30.

Donnellan, K. (1970). Proper Names and Identifying Descriptions. Synthese 21: 335–358. Reprinted in K. Donnellan (2012). Essays on Reference, Language, and Mind. New York: Oxford University Press: 49–79.

Dummett, M. (1973/81). Frege: Philosophy of Language (second edition). Oxford: Duckworth. ← 57 | 58 →

Dummett, M. (1981). The Interpretation of Frege’s Philosophy. Oxford: Duckworth.

Frege, G. (1891). Function and Concept. In: P. Geach and M. Back (eds.) (1952). Translations from the Philosophical Writings of Gottlob Frege. Oxford: Blackwell, 21–41.

Frege, G. (1892a). On Concept and Object. In: P. Geach and M. Back (eds.) (1952). Translations from the Philosophical Writings of Gottlob Frege. Oxford: Blackwell, 42–55.

Biographical notes

Piotr Stalmaszczyk (Volume editor)

Piotr Stalmaszczyk is a professor of English Language and Linguistics at the University of Lodz (Poland), where he holds the Chair of English and General Linguistics. His research is concerned predominantly with linguistic methodology and the philosophy of language and linguistics.


Title: Objects of Inquiry in Philosophy of Language and Linguistics